Prms Guidelines Combined-Final

PRMS
Guidelines
Guidelines for Application of the

Petroleum Resources Management System
REVISED July 2022
Sponsored by:
World Petroleum Council
CM
Guidelines for Application of the
Petroleum Resources Management System
Sponsored by:
Society of Petroleum Engineers (SPE)

World Petroleum Council (WPC)
American Association of Petroleum Geologists (AAPG)
Society of Petroleum Evaluation Engineers (SPEE)
Society of Exploration Geophysicists (SEG)
Society of Petrophysicists and Well Log Analysts (SPWLA)
European Association of Geoscientists & Engineers (EAGE)
Version 1.0
ISBN 978-1-61399-983-7
© Copyright 2022 Society of Petroleum Engineers
All rights reserved. No portion of this report may be reproduced in any form or by any means, including electronic
storage and retrieval systems, except by explicit, prior written permission of the publisher except for brief
passages excerpted for review and critical purposes.
Foreword
Charles Vanorsdale
This document, Guidelines for Application of the PRMS, or “AG,” represents an update to the 2011
version of the same-named document. It is a companion volume to the Petroleum Resources
Management System, or “PRMS,” which was updated in 2018. The AG is not intended to replace
any principles within the PRMS but rather to provide further clarity on the PRMS guidelines by
way of textual detail and example situations. In the event of any conflict between specific
principles in the AG and the PRMS, the PRMS principles will take precedence.
Note that sections within the two documents are referenced differently herein. When the AG, in
any of its individual chapters, refers to a section with the symbol “§,” this will indicate a specific
passage within the PRMS. Any other reference to a section, such as “Section” or “Sec.,” indicates
a specific passage within the current chapter of the AG.
The individual chapters of the AG contain numerous example applications of PRMS guidelines.
As with all resource assessment processes, the examples are based on the particular situations and
data as spelled out in the chapter, and the results from the examples are not intended to be
universally true for all situations and/or data.
Finally, the AG is based on and aligned with the PRMS, and it is not intended to be utilized in
conjunction with any other resources guidelines or regulatory reporting requirements.

Contents
1.0 INTRODUCTION ....................................................................................................... 1
Charles Vanorsdale, Ron Harrell
1.1 Rationale for the Applications Guidelines Revision .............................................................1
1.2 History of Petroleum Reserves and Resources Definitions ..................................................2
1.3 References .................................................................................................................................3
2.0 PETROLEUM RESOURCES DEFINITIONS, CLASSIFICATION,

AND CATEGORIZATION GUIDELINES .................................................................. 4
Rich DuCharme, Dan Olds, and Xavier Troussaut
2.1 Introduction .............................................................................................................................4
2.1.1 Project-Based Classification System ............................................................................5
2.2 Defining a Project ...................................................................................................................6
2.3 Project Classification ..............................................................................................................9
2.4 Range of Uncertainty Categorization..................................................................................10
2.5 Methods for Estimating the Range of Uncertainty in Recoverable Quantities ...............12
2.5.1 Deterministic “Scenario” Method ..............................................................................12
2.5.2 Deterministic “Incremental” Method ........................................................................12
2.5.3 Geostatistical Method ..................................................................................................13
2.5.4 Probabilistic Method ...................................................................................................13
2.5.5 Integrated Method .......................................................................................................13
2.6 Chance of Commerciality ......................................................................................................13
2.7 Project Maturity Sub-Classes ...............................................................................................15
2.8 Reserves Status .......................................................................................................................18
2.9 Economic Status .....................................................................................................................20
2.10 Estimated Ultimate Recovery and Technically Recoverable Resources .........................21
2.11 Example Application of Commercial Maturity Sub-Classes ...........................................23
2.12 Acknowledgments ................................................................................................................25
2.13 Reference ..............................................................................................................................25
3.0 SEISMIC APPLICATIONS ...................................................................................... 26

Chuandong “Richard” Xu, Dan Maguire, Andrew Royle, David Johnston,
Eric Von Lunen, and Jean-Pierre Blangy
3.1 Introduction and Overview ...................................................................................................26
3.2 Shared Earth Modeling .........................................................................................................26
3.2.1 Static Model—Geologic Framework Architecture ....................................................27
3.2.2 Dynamic Model—Reservoir Simulation .....................................................................27
3.3 Volumetrics, Trap Geometry, and Gross Rock Volume ....................................................28
3.3.1 Defining Seismic Data ...................................................................................................28
3.3.2 Uncertainty in GRV ......................................................................................................29
3.4 Amplitude Variation With Offset and Direct Hydrocarbon Indicator Analyses ............30
3.4.1 Introduction ...................................................................................................................30
3.4.2 Confidence Measures ....................................................................................................32
3.4.3 Evaluation of Resources Using Seismic and Well Data .............................................33
3.4.4 Recommended Guidelines for Using Seismic DHIs ...................................................34
3.4.5 DHI Examples ...............................................................................................................36
3.5 Seismic Inversion ...................................................................................................................38
3.5.1 Concepts and Definitions..............................................................................................38
3.5.2 Well-to-Seismic Correlation via Synthetic Seismograms ..........................................40
3.5.3 Rock Physics Modeling/Analysis .................................................................................41
3.5.4 Seismic Impedance Inversion.......................................................................................41
3.5.5 Stochastic Inversion ......................................................................................................44
3.5.6 Uncertainty Assessment................................................................................................46
3.6 Seismic Surveillance...............................................................................................................48
3.6.1 Seismic Surveillance Example A..................................................................................50
3.6.2 Seismic Surveillance Example B ..................................................................................52
3.6.3 Seismic Surveillance Example C..................................................................................53
3.6.4 Seismic Surveillance Example D..................................................................................54
3.7 Conclusion ..............................................................................................................................54
3.8 Acknowledgments ..................................................................................................................55
3.9 References ...............................................................................................................................55
4.0 ASSESSMENT OF PETROLEUM RESOURCES USING

DETERMINISTIC PROCEDURES .......................................................................... 57
Danilo Bandiziol, Dominique Salacz, Jes Christensen, Joel Turnbull, and Oluyemisi Jeje
4.1 Introduction ............................................................................................................................57

4.2 Project Life Cycle...................................................................................................................58
4.3 Analogy-Based Assessment ...................................................................................................60
4.3.1 Purpose of Analogs .......................................................................................................61
4.3.2 Methodology for the Application of Analogy .............................................................61
4.3.3 Application of Analogy to an Oil Example .................................................................62
4.4 Volumetric Analysis ...............................................................................................................64
4.4.1 Estimating Volumetric Parameters .............................................................................65
4.4.2 Application of Volumetrics to Oil Example................................................................68
4.5 Performance-Based Methods ................................................................................................69
4.5.1 Material Balance ...........................................................................................................70
4.5.2 Reserve Simulation .......................................................................................................78
4.5.3 Decline Curve Analysis .................................................................................................78
4.6 Summary and Conclusions ...................................................................................................91
4.7 Acknowledgments ..................................................................................................................92
4.8 References ...............................................................................................................................92
Appendix A: Special Considerations for Material Balance Analyses ..................... 95

4A.1 Diagnostic Plots ...................................................................................................................95
4A.2 Low-Permeability Reservoirs.............................................................................................97
4A.3 Volatile Oils .........................................................................................................................98
4A.4 Material Balance Pressure Data Quality ..........................................................................98
4A.5 Abandonment Pressure and Compression .......................................................................98
4A.6 Nonlinear p/z Relationships ................................................................................................99
4A.7 Overpressured Reservoirs ................................................................................................100
4A.8 Waterdrive Reservoirs......................................................................................................101
4A.9 Retrograde Condensate Reservoirs .................................................................................101
4A.9.1 Considerations Related to Liquid Comments ....................................................102
5.0 PETROPHYSICS .................................................................................................. 104

Luis Quintero, Javier Miranda, Joshua Oletu, Cecilia Flores, George Dames,
and Philip Gibbons
5.1 Introduction ..........................................................................................................................104
5.2 Volumetric Estimation of PIIP and Estimated Ultimate Recovery.................................106
5.3 Petrophysical Evaluation ....................................................................................................106
5.4 What is “Net Pay”? ..............................................................................................................106
5.4.1 Pay Cutoffs ..................................................................................................................107
5.5 Core Analysis........................................................................................................................110
5.5.1 How Representative Is the Core Sample? ................................................................110
5.5.2 Capillary Pressure ......................................................................................................111
5.5.3 Core Wettability ..........................................................................................................112
5.5.4 Rock Typing ................................................................................................................114
5.5.5 Reservoir Continuity ..................................................................................................114
5.5.6 Quantifying Residual Oil Saturation ........................................................................116
5.6 Examples ...............................................................................................................................117
5.6.1 Example A....................................................................................................................117
5.6.2 Example B ....................................................................................................................118
5.6.3 Example C....................................................................................................................118
5.6.4 Example D....................................................................................................................119
5.6.5 Example E ....................................................................................................................120
5.7 Conclusion ............................................................................................................................121
5.9 References .............................................................................................................................122
6.0 RESERVOIR SIMULATION .................................................................................. 124
Miles Palke, Avi Chakravarty, Ali Albinali, and Charles Vanorsdale
6.1 Introduction ..........................................................................................................................124
6.2 What is Reservoir Simulation .............................................................................................124
6.3 Use of Simulation in Resource Estimation.........................................................................125
6.3.1 Immature Reservoirs ..................................................................................................126
6.3.2 Mature Reservoirs ......................................................................................................127
6.4 Fundamentals of Simulation Quality Assessment.............................................................128
6.4.1 Static Model Construction .........................................................................................128
6.4.2 Dynamic Data Integration..........................................................................................132
6.4.3 History Matching ........................................................................................................133
6.4.4 Validity of the Predictions ..........................................................................................137
6.5 Proved (1P) Reserves Case ..................................................................................................141
6.5.1 Adjusting 2P Results for 1P Reporting .....................................................................142
6.5.2 Further Comment on Adjusting 2P Results for 1P Reporting ...............................143
6.6 Other Output from Simulation ...........................................................................................144
6.8 References .............................................................................................................................144
7.0 PROBABILISTIC RESOURCES ESTIMATION .................................................... 146

Carolina Coll, David Elliott, Enrique Morales, Karl Stephen, and Richard Wheaton
7.1 Introduction ..........................................................................................................................146
7.2 Resources Uncertainty..........................................................................................................148
7.2.1 Uncertainty in Input Parameters ..............................................................................149
7.2.2 Project Uncertainties ..................................................................................................150
7.3 Deterministic Methods.........................................................................................................151
7.3.1 Deterministic Analysis ................................................................................................151
7.3.2 Sensitivity Analysis and the Multiple-Realizations Method ...................................152
7.4 Hybrid Methods ...................................................................................................................154
7.4.1 Probability Tree Analysis ...........................................................................................154
7.4.2 Multi-Scenario Method ..............................................................................................157
7.5 Probabilistic Methods ..........................................................................................................159
7.5.1 Monte Carlo Methods .................................................................................................163
7.5.2 Experimental Design Methods and Response Surface Models ...............................164
7.5.3 Global Optimization Methods ..................................................................................166
7.6 Probabilistic Workflows for Recoverable Resources Estimates ......................................167
7.6.1 Preproduction Phase ...................................................................................................167
7.6.2 Production Phase ........................................................................................................168
7.7 Consistency between Probabilistic and Deterministic Methods ......................................170
7.8 Commercial Considerations................................................................................................172
7.9 Final Considerations ............................................................................................................173
7.10 References ...........................................................................................................................174
7.11 Glossary and Definitions ...................................................................................................178
8.0 AGGREGATION OF RESERVES AND RESOURCES ........................................ 182

William J. Haskett and Tyler Schlosser
8.1 Introduction ..........................................................................................................................182
8.1.1 Defining Reservoir Relationships ..............................................................................184
8.2 Aggregating Over Reserves Levels .....................................................................................185
8.2.1 Reservoir Performance ...............................................................................................185
8.2.2 Issues with Arithmetic Summation ...........................................................................185
8.2.3 Correlations Between Estimates ................................................................................190
8.2.4 Dependency ..................................................................................................................194
8.2.5 Higher-Level Aggregation ..........................................................................................199
8.3 Adding Proved Reserves......................................................................................................200
8.3.1 Pitfalls of Using Arithmetic Addition of Proved Reservoirs ...................................200
8.3.2 Arithmetic or Correlation Inclusive Summation .....................................................200
8.3.3 Probabilistic Aggregation and the Portfolio Effect..................................................202
8.3.4 Simple Probabilistic, Stochastic Simulation, and Correlation Matrices ...............203
8.4 Regional Aggregation ..........................................................................................................207
8.5 Scenario Methods .................................................................................................................210
8.5.1 Tree-Based Example of Correlation Between Reservoir Elements ........................210
8.6 Summary—Some Guidelines ..............................................................................................213
8.8 References .............................................................................................................................214
9.0 EVALUATION OF PETROLEUM RESERVES AND RESOURCES .................... 216

Charles Vanorsdale
9.1 Introduction ..........................................................................................................................216
9.2 Net Cash-Flow Evaluation ..................................................................................................217
9.2.1 Step 1—Test That the Project Is Economic..............................................................217
9.2.2 Step 2—Determine the Project Life ..........................................................................218
9.3 Terminology of Cash-Flow-Based Evaluation ..................................................................219
9.3.1 Net Cash Flow (NCF)..................................................................................................219
9.3.2 Abandonment Decommissioning, and Restoration (ADR) .....................................220
9.3.3 Economic Limit (EL) ..................................................................................................220
9.3.4 Economic Conditions ..................................................................................................221
9.3.5 Discount Rate ..............................................................................................................222
9.3.6 Other Key Terms ........................................................................................................223
9.4 Required Input for a Cash-Flow-Based Evaluation .........................................................224
9.4.1 Net Entitlement Sales Production Forecast ..............................................................224
9.4.2 Product Prices .............................................................................................................224
9.4.3 Project Capital Costs ..................................................................................................225
9.4.4 Operating Costs...........................................................................................................226
9.4.5 Ownership and Royalties ...........................................................................................227
9.4.6 Taxes ............................................................................................................................227
9.4.7 Legal/Contract/Fiscal Terms .....................................................................................227
9.5 Generating a Cash-Flow-Based Evaluation ......................................................................228
9.6 Analyzing a Cash-Flow-Based Evaluation ........................................................................231
9.6.1 Investment (Hurdle) Criteria .....................................................................................231
9.6.2 Sunk Costs ...................................................................................................................232
9.7 “Economic” Compared to “Commercial” .........................................................................232
9.8 Example ................................................................................................................................233
9.9 Probabilistic Evaluation ......................................................................................................237
9.10 Environmental, Social, and Governance Issues ..............................................................237
9.10.1 ESG Within the PRMS Context ............................................................................237
9.10.2 ESG Impact on Economic Evaluation ..................................................................238
9.11 Conclusion ..........................................................................................................................238
9.12 Acknowledgments ..............................................................................................................239
9.13 References ...........................................................................................................................239
9.14 Bibliography .......................................................................................................................239
10.0 UNCONVENTIONAL RESOURCES ESTIMATION ............................................ 240

Dilhan Ilk, Roberto Aguilera, Creties Jenkins, Christopher R. Clarkson,
John Etherington, and Charles Vanorsdale
10.1 Introduction ........................................................................................................................240
10.1.1 Assessment and Classification Issues ....................................................................241
10.2 Tight Gas and Oil Formations ..........................................................................................243
10.2.1 Introduction .............................................................................................................243
10.2.2 Reservoir and Hydrocarbon Characteristics .......................................................245
10.2.3 Assessment Methods ...............................................................................................247
10.2.4 Drilling, Completion, and Stimulation Issues ......................................................249
10.2.5 Processing and Marketing......................................................................................250
10.2.6 Commerciality Issues ..............................................................................................250
10.2.7 Classification and Reporting Issues ......................................................................250
10.3 Shale Gas and Oil ...............................................................................................................251
10.3.1 Introduction .............................................................................................................251
10.3.2 Reservoir Characteristics .......................................................................................252
10.3.3 Drilling and Completions .......................................................................................253
10.4 Evaluation Methodologies for Tight/Shale Oil and Gas .................................................255
10.4.1 Introduction .............................................................................................................255
10.4.2 Overview ..................................................................................................................255
10.4.3 Production Diagnostics ...........................................................................................260
10.4.4 Model-Based Analysis.............................................................................................262
10.4.5 Well Performance Analysis and Forecasting Example .......................................263
10.4.6 Estimating Recoverable Quantities in Unconventional Resources ....................270
10.5 Coalbed Methane ...............................................................................................................273
10.5.1 Introduction .............................................................................................................273
10.5.2 CBM Reservoir Characteristics.............................................................................273
10.5.3 Drilling and Completions .......................................................................................279
10.5.4 Evaluation Methods for Reserves/Resource Estimation .....................................279
10.5.5 Additional Exploration and Development Considerations .................................288
10.5.7 Classification and Reporting Issues ......................................................................289
10.6 Other Unconventional Oil .................................................................................................290
10.6.1 Introduction .............................................................................................................290
10.6.2 Reservoir Characteristics, Risk, and Uncertainty................................................291
10.7 References ...........................................................................................................................294
11.0 PRODUCTION MEASUREMENT AND OPERATIONAL ISSUES ..................... 303

Mohammed Al Alshaikh
11.1 Introduction ........................................................................................................................303
11.2 Background ........................................................................................................................303
11.3 Reference Point ..................................................................................................................305
11.4 Consumed Operations (CIO) ............................................................................................306
11.5 Associated Non-Hydrocarbon Components ....................................................................307
11.6 Natural Gas Reinjection ....................................................................................................308
11.7 Underground Natural Gas Storage ..................................................................................308
11.8 Production Balancing ........................................................................................................309
11.8.1 Production Imbalances (Overlift/Underlift) .........................................................309
11.8.2 Gas Balancing ..........................................................................................................309
11.9 Shared Processing Facilities ..............................................................................................309
11.10 Hydrocarbon Equivalence Issues ...................................................................................310
11.10.1 Gas Conversion to Oil Equivalent .....................................................................310
11.10.2 Liquid Conversion to Oil Equivalent ................................................................311
11.11 Example ............................................................................................................................312
11.12 Acknowledgments ............................................................................................................313
11.13 References .........................................................................................................................313
12.0 RESOURCES ENTITLEMENT AND RECOGNITION......................................... 314
Monica Clapauch Motta, Elliott Young and Regnald A. (Reggie) Boles
12.1 Foreword.............................................................................................................................314
12.2 Introduction ........................................................................................................................314
12.3 Regulations, Standards, and Definitions..........................................................................315
12.4 Reserves and Resources Recognition Under the PRMS .................................................316
12.4.1 Reserves and Resources Entitlement Elements....................................................316
12.4.2 Working Interest and Economic Interest .............................................................317
12.4.3 Mineral Property Conveyances .............................................................................319
12.5 Agreements and Contracts ................................................................................................319
12.5.1 Concessions, Mineral Leases, and Permits ...........................................................320
12.5.2 Production-Sharing Contracts...............................................................................321
12.5.3 Revenue-Sharing Contracts ...................................................................................325
12.5.4 Risked-Service Contracts .......................................................................................326
12.5.5 Pure-Service Contracts...........................................................................................327
12.5.6 Loan Agreements ....................................................................................................328
12.5.7 Production Loans, Forward Sales, and Similar Arrangements .........................328
12.5.8 Carried Interests .....................................................................................................329
12.5.9 Entitlement with Interest Before Payout and Interest After Payout Example ...330
12.5.10 Purchase Contracts ...............................................................................................330
12.5.11 Volume-Denominated Production Payments .....................................................331
12.5.12 Other Contracts and Agreements........................................................................331
12.6 Other Concerns of Resources Entitlement and Recognition .........................................332
12.6.1 Taxes and Reserves/Resources...............................................................................332
12.6.2 Royalties, Overriding Royalty, and Government Fixed Volume
Entitlement and Reserves/Resources ....................................................................332
12.6.3 Unitization Agreements ..........................................................................................337
12.6.4 Contract Extensions or Renewals ..........................................................................344
12.6.5 Appropriate Date on Which to Recognize Entitlement for Reserves
Estimates in Specific Cases ....................................................................................345
12.7 Example Case .....................................................................................................................346
12.7.1 Reserves in a PSC....................................................................................................346
12.8 Summary.............................................................................................................................352
12.9 Acknowledgments ..............................................................................................................353
12.10 References .........................................................................................................................353
12.11 Bibliography .....................................................................................................................354
GLOSSARY ................................................................................................................ 355

Chapter 1
Introduction
Charles Vanorsdale
1.1 Rationale for the Applications Guidelines Revision

The Guidelines for Application of the PRMS (or Application Guidelines, AG), as the name
suggests, is the companion document to the Petroleum Resources Management System (PRMS),
with the intended purpose of providing the PRMS user with a more detailed understanding of the
principles involved therein for consistent practice in petroleum reserves and resources evaluation.
Industry feedback through professional conferences, workshops, and public comment since the
2007 publication of the PRMS reinforced the need to issue further clarification and amplification
of the PRMS and its guidelines. Consequently, to address these concerns, the original AG
document was published in November 2011 (Guidelines for Application of the Petroleum
Resources Management System 2011).
In 2014, an effort began to update the PRMS (2007) with the formation of several small teams
within the Society of Petroleum Engineers (SPE) Oil and Gas Reserves Committee. The Oil and
Gas Reserves Committee called upon representatives from the five sponsoring professional
organizations that had created the 2007 PRMS (SPE, World Petroleum Council, American
Association of Petroleum Geologists, Society of Petroleum Evaluation Engineers, and Society of
Exploration Geophysicists) with the addition of two further organizations (Society of
Petrophysicists and Well Log Analysts, and the European Association of Geoscientists and
Engineers) in the endeavor to collaboratively issue the updated PRMS. The clarification process
culminated in the release of PRMS (2018) in June of that year, retaining the founding
principles embodied in PRMS 2007. With that revision, an update to the 2011 AG was necessary,
and it was planned to be co-released by the seven sponsoring organizations.
The 2011 AG itself was a significant update of the prior Guidelines for the Evaluation of
Petroleum Reserves and Resources released in 2001. The 2011 AG was likewise a four-year
undertaking that drew upon the resources of the five original sponsors and resulted in the inclusion
of two additional major topic chapters, “Assessment of Petroleum Reserves Using Deterministic
Procedures” and “Unconventional Resources.” Example applications were inserted in several
chapters for the first time.
With that same professional collaborative vision, steered by the Oil and Gas Reserves
Committee, the seven organizations devoted considerable resources toward this new AG. As with
the PRMS, the AG update involved worldwide representation and a broad spectrum of industry
participants from major producers, national oil companies, service companies, the banking and
financial community, petroleum consulting firms, and universities. Tens of thousands of hours
were dedicated by highly qualified individuals volunteering their time toward these efforts.
The ever-changing petroleum industry has necessitated a reevaluation of material in the AG
2011. The AG 2011 was the starting point for this work and has been augmented considerably by
subject matter experts incorporating industry feedback and the inclusion of several illustrative
examples in each chapter. These examples were created for the express purpose of conveying
principles and concepts that users of the previous AG noted were not clearly articulated by the text
1
alone. Further, the chapter on “Unconventional Resources” has been revised to reflect the advances
in interpretation and assessment methods since 2011. Two new chapters have been added in this
AG, namely:
• Petrophysics (Chapter 5) and
• Reservoir Simulation (Chapter 6).
These chapters have been added due to the continued integration of multidisciplinary technical
activities associated with petroleum resources evaluation. PRMS 2018 introduced, for the first
time, a definition of “net pay,” a term used frequently in the industry, but one that actually carries
different connotations between different disciplines. The term obviously impacts petrophysics, as
well as reservoir simulation and several other chapters.
In fact, a mission of this updated AG was to view the content not as separate, standalone
chapters but rather from a holistic standpoint. The editorial teams for each chapter were responsible
not only for the review of their assigned chapter, but also the review of all chapters, in an attempt
to ensure a common and consistent language for the concepts and principles throughout. A concern
noted by several authors of this AG has been that the broader perspective of reserves and resources
assessment has been supplanted by reliance on the opinion of subject matter experts, without
independent scrutiny on the part of the evaluator. Ultimately, the evaluator is accountable not only
for the end product (typically a reserves/resources assessment report), but also for the
multidisciplinary input used to generate that report.
A list of Reference Terms used in resources evaluations is included at the end of the AG
document. The list does not replace the PRMS Glossary but refers the user to the chapters and
sections where the terms are used in the AG.
Finally, as with the PRMS 2018, material updates to this AG revision will be uploaded using
version control on the SPE website.
Ron Harrell
1.2 History of Petroleum Reserves and Resources Definitions

The initial efforts at establishing oil reserves definitions in the US were led by the American
Petroleum Institute. At the beginning of World War I, the US government formed the National
Petroleum War Service Committee to ensure adequate oil supplies for the war effort. At the close
of World War I, the National Petroleum War Service Committee was reborn as the American
Petroleum Institute. In 1937, American Petroleum Institute created definitions for Proved oil
reserves that they followed in their annual estimates of US oil reserves. Little attention was paid
to natural gas reserves until after 1946, when the American Gas Association created similar
definitions for Proved gas reserves.
SPE’s initial involvement in establishing petroleum reserves definitions began in 1962
following a plea from US banks and other investors for a consistent set of reserves definitions
that could be both understood and relied upon by the industry in financial transactions where
petroleum reserves served as collateral. Individual lenders and oil producers had their own “in-
house” definitions, but these varied widely in content and purpose. In 1962, the SPE Board of
Directors appointed a 12-person committee of well-recognized and respected individuals. They
were known as the Special Committee on Definitions of Proved Reserves for Property
Evaluation. The group was composed of two oil producers, one pipeline company, one university
professor, two banks, two insurance companies (lenders), and four petroleum consultants.
2
This group collaborated over a period of three years, debating the exact wording and terms of
their assignment before submitting their single-page work product to the SPE board in 1965. The
SPE board adopted the committee’s recommendation by a vote of seven in favor, three dissenting,
and two abstaining. The American Petroleum Institute observer was supportive; the American Gas
Association observer opposed the result.
In 1981, SPE released updated Proved oil and gas definitions that contained only minor
revisions of the initial 1965 version.
The 1987 SPE petroleum reserves definitions were the result of an effort initiated by the
Society of Petroleum Evaluation Engineers, but ultimately developed and sponsored by SPE.
These definitions, issued for the first time by a large professional organization, included
recognition of the unproved categories of Probable and Possible Reserves. Much discussion
centered around the use of probabilistic assessment techniques as a supplement or alternative to
more-traditional deterministic methods. Following the receipt of comments from members
worldwide, and in particular members from North America, the SPE board did not approve the
inclusion of any discussion about probabilistic methods of reserves evaluation in the 1987
definitions. As a consequence, these definitions failed to garner widespread international
acceptance and adoption.
The 1997 SPE/World Petroleum Council reserves definitions grew out of a cooperative
agreement between the World Petroleum Council and SPE and appropriately embraced the
recognition of probabilistic assessment methods. The American Association of Petroleum
Geologists became a sponsor of and an integral contributor to the 2000 SPE/World Petroleum
Council/American Association of Petroleum Geologists definitions and provided invaluable
contributions toward the implementation of resources definitions. In 2007, the Society of
Petroleum Evaluation Engineers became the fourth sponsoring society, followed by the Society of
Exploration Geophysicists joining as an adopting sponsor shortly thereafter.
This recitation is not intended to omit or minimize the creative influence of numerous other
individuals, organizations, or countries who have made valuable contributions over time to the
derivation of petroleum resources definitions out of an initial mining perspective. Users of this
information are encouraged, however, to recognize that the foregoing history of relevant industry-
adopted reserves and resources definitions and evaluation practices has been entirely created by
competent industry professionals tirelessly volunteering their time over many years—several
decades for some—to create the controlling standards and guidance in reserves and resources
management practices worldwide.
Further, the PRMS/AG sponsors recognize that the reserves and resources definitions must
remain relevant and up-to-date and will remain diligent in working toward periodic updates
and improvements.
1.3 References
Guidelines for Application of the Petroleum Resources Management System. 2011. Richardson,
Texas, USA: Society of Petroleum Engineers, and London, UK: World Petroleum Council.
Petroleum Resources Management System. 2007. Richardson, Texas, USA: Society of
Petroleum Engineers.
Petroleum Resources Management System, Version 1.01. 2018. Richardson, Texas, USA:
Society of Petroleum Engineers.
3
Chapter 2
Petroleum Resources Definitions,

Classification, and Categorization
Guidelines
Rich DuCharme (Chair)
Dan Olds and Xavier Troussaut
2.1 Introduction
The Petroleum Resources Management System (2018), termed PRMS herein, is a fully integrated
system that provides the basis for classification and categorization of all petroleum reserves and
resources. Although the system encompasses the entire in-place petroleum resource and characterizes
projects at various levels of technical and commercial maturity, its widest application has been for
estimating commercially recoverable quantities using a globally recognized system. Because no
petroleum quantities can be commercially recovered without the installation of (or access to) the
appropriate production, processing, and transportation facilities, application of the PRMS focuses on
the development project that has been (or will be) implemented to recover petroleum from one or more
accumulations. Further, the PRMS provides an explicit distinction between the chance of
commerciality of that project, which defines its maturity (Classification), and the range of uncertainty
in the petroleum quantities forecast to be potentially recovered and marketed in the future from that
project (Categorization). This two-axis system is illustrated in Fig. 2.1.
Fig. 2.1—Resources classification framework.
4
Each project is classified according to its maturity or status using three main classes, with the
option to subdivide further using sub-classes. The three classes are Reserves, Contingent
Resources, and Prospective Resources, which reflect the remaining estimated recoverable
quantities at the time of the evaluation. These estimates of remaining recoverable quantities
exclude previously produced volumes as well as unrecoverable quantities that may become
recoverable in the future as new technology is developed or economic conditions improve.
Separately, the range of uncertainty in the estimated recoverable and marketable quantities from
that specific project is categorized based on the principle of identifying three estimates (Low, Best,
and High) of the potential outcome.
For projects that satisfy the requirements for commerciality (PRMS § 2.1.2), Reserves may
be assigned to the project, and the three estimates of the recoverable commercial quantities would
be designated as 1P (Proved: equivalent to low estimate or at least P90 if probabilistic methods are
used), 2P (Proved plus Probable: equivalent to best estimate or at least P50 if probabilistic methods
are used), and 3P (Proved plus Probable plus Possible: equivalent to high estimate or at least P10
if probabilistic methods are used) Reserves. The equivalent categories for sub-commercial projects
with Contingent Resources are 1C, 2C, and 3C, while equivalent categories for Prospective
Resources are 1U, 2U, and 3U.
The PRMS also accommodates the ability to categorize and report Reserve quantities
incrementally as Proved (P1), Probable (P2), and Possible (P3), rather than using the realizable
scenarios of 1P, 2P, and 3P. Likewise, the incremental quantities of Contingent Resources can be
reported as C1, C2, and C3 rather than using the scenarios of 1C, 2C, and 3C. Because of the
undiscovered nature of Prospective Resources and the likely application of probabilistic
approaches, recognition of incremental quantities for Prospective Resources implies more
precision than is warranted and, therefore, they are not differentiated incrementally in the PRMS.
2.1.1 Project-Based Classification System. The PRMS is a project-based system, where a project
is “a defined activity or set of activities that provides the link between the petroleum accumulation’s
resources sub-class and the decision-making process, including budget allocation. A project may, for
example, constitute the development of a single reservoir or field, an incremental development in a
larger producing field, or the integrated development of a group of several fields and associated
facilities (e.g., compression) with a common ownership. In general, an individual project will
represent a specific maturity level (sub-class) at which a decision is made on whether or not to
proceed (i.e., spend money), suspend, or remove. There should be an associated range of estimated
recoverable resources for that project” (PRMS 2018, Appendix A, 47).
There are actually three elements that constitute an evaluation of net recoverable resources.
The Project definition is one element. In addition, we have the Reservoir (which contains the
petroleum accumulation) and the Property (which defines the unique associated contractual rights
and obligations, including the fiscal terms, of the lease or license area) elements. As discussed in
more detail in PRMS § 1.2, these elements are related as shown in Fig. 2.2.
5
Fig. 2.2—Key elements of a resources evaluation.
The scope of this chapter relates to the Project and the associated definitions, classification, and
categorization of the recoverable petroleum resources. Aspects concerning the related cash flow and
commerciality considerations of the Project are detailed in Chapter 9—Evaluation of Petroleum
Reserves and Resources of this document. The technical assessments of the in-place volumes and
their recoverable quantities are discussed from a deterministic approach in Chapter 4—Assessment
of Petroleum Resources using Deterministic Procedures and from a probabilistic approach in
Chapter 7—Probabilistic Resources Estimation. Ownership and entitlement are the focus of
Chapter 12—Resources Entitlement and Recognition.
2.2 Defining a Project

Each project is considered as an investment opportunity, and a decision must be made to proceed,
reassess, postpone, or exit the project. An entity’s decisions reflect the selection or rejection of
investment opportunities from a portfolio based on consideration of the funds available, the cost
of the specific project, and the expected outcome (in terms of value) of that investment. The critical
point in the PRMS is the linkage between the decision to proceed with a project and the estimated
future recoverable quantities associated with that project. It is worth noting that the rigor of the
assessment used to estimate the recoverable quantities (Low, Best, High) for the project will tend
to align with the maturity of the project (i.e., more mature projects will require more detailed
development plans and more rigorous analysis) (see PRMS § 1.2.0.9).
Defining the term “project” unambiguously can be difficult because its nature will vary with the
level of maturity. For example, a mature project may be defined in great detail by a comprehensive
development plan, which may be prepared and submitted to the partners, host government, or
relevant regulatory authority for approval to proceed with development, or a development plan may
contain multiple projects at different maturity levels. The development plan may be revised over
time as updated results are obtained or as less mature projects in the development plan increase their
commercial maturity and likelihood of occurrence. These development plans may include full details
of all the planned development wells and their locations, specifications for surface processing and
export facilities, discussion of environmental considerations, staffing requirements, market
assessment, estimated capital, operating and site rehabilitation costs, etc. In contrast, the drilling of
an exploration prospect represents a less mature project that could become a commercial
development if the well is successful. The assessment of the economic viability of the exploration
project will still require a view of the likely development scheme, although the development plan
may be outlined in broad conceptual terms based on analogs.
In all cases, the decision to proceed with a project not only requires an estimate of future
product prices, but also an assessment of future costs, based on an evaluation of the development
scope (e.g., wells, facilities) and operation, to support the economic evaluation of that investment.
In this context, the development facilities include all the necessary production, processing, and
6
transportation facilities to enable delivery of petroleum from the accumulation(s) to a product sales
point (or to an internal transfer point between upstream operations and midstream/downstream
operations); for more detail, please see Chapter 11—Production Measurement & Operational
Issues of this document. This development scope and cost define the project because the planned
investment of the capital costs is the basis for the economic evaluation of the investment and,
hence, the decision to proceed (or not) with the project. Evaluation of the estimated recoverable
sales quantities and the range of uncertainty in that estimate will also be key inputs to the economic
evaluation, and these are based on a defined development project.
A project may involve the development of a single petroleum accumulation (e.g., a reservoir
or field), or a group of accumulations, or there may be more than one project implemented on a
single accumulation. The following are some examples of projects:
• When a detailed development plan is prepared for partner and/or government approval,
then the plan itself defines the project. If the plan included optional wells that had received
capital commitment with project and government approval, these would not constitute a
separate project, but would form part of the assessment of the range of uncertainty in
potentially recoverable quantities from the project.
• When a development project is defined to produce oil from an accumulation that also
contains a significant gas cap, and the gas cap development is not an integral part of the
oil development, then a separate gas development project may also be defined for a future
project decision, even if there is currently no gas market (i.e., Contingent Resources).
• When a development plan is based on primary recovery only, and a secondary recovery
process is envisioned but will be subject to a separate capital commitment decision and/or
approval process at the appropriate time, then it is considered as two separate projects.
• When decision making is entirely on an individual, well-by-well basis, then each well
constitutes a separate project.
• When late-life installation of gas-compression facilities is included in the original
approved development plan, then it is part of a single gas development project. When
compression was not part of the approved development plan, yet it is technically feasible
and will require economic justification and a capital commitment decision and/or
approval before installation, then the installation of gas-compression facilities represents
a separate project (PRMS § 2.3.2).
• In the assessment of an undrilled prospect, an economic evaluation will be made to
underpin the decision to drill. This evaluation includes consideration of a conceptual
development plan in order to estimate costs and recoverable quantities (Prospective
Resources) assuming a positive outcome from the exploration well. The project is defined
by the exploration well and a conceptual development plan that will have less detail than
more mature opportunities (e.g., often relying on analog developments and reasonable
assumptions).
• In some cases, an investment decision may involve multiple projects of exploration,
appraisal, and/or development activities. Because the PRMS subdivides resource
quantities on the basis of three classes that reflect the distinction between these activities
(i.e., Reserves, Contingent Resources, and Prospective Resources), it is appropriate to
consider that the investment decision is based on implementing a group of projects,
whereby each project is uniquely placed in one of the three classes (by identifying
separate projects, the evaluator avoids the problem of split classification).
7
• When a developed field’s reserves are constrained by a license or contract expiry, then
those quantities that may be produced beyond the expiration date of the current contract
will be attributed to a separate project associated with renegotiated contract extension and
fiscal terms and extended field life. This project is typically assigned to Contingent
Resources with a reduced chance of commercialization, unless there is reasonable
expectation that an extension, renewal, or new contract will be granted, which, in this
case, may allow for these quantities to be assigned to Reserves (PRMS § 3.3.3.2). A
reasonable expectation of extension and fiscal terms may be based on existing contract
extension terms, negotiations, or historical treatment of similar agreements. If the
evaluator is unable to support reasonable expectations, then the recoverable quantities
associated with the potential extension should be retained in Contingent Resources and
may only be assigned to Reserves when reasonable expectation of proceeding with the
extension has been achieved. Note that this separate project associated with pursuing an
extension will be assigned to a single sub-class and, therefore, is not considered split
classification (PRMS § 2.2.0.4).
Projects may change in character over time and can aggregate or subdivide. For example, an
exploration project may initially be defined on the basis of drilling a well; yet, if a discovery is
made, a subsequent project decision to develop the accumulation will be considered as a separate
standalone project. However, if the discovery is smaller than expected and perhaps is unable to
support an export pipeline on its own, then the project might be placed in the appropriate
contingent sub-class (see Section 2.7) and delayed until another discovery is made nearby, and the
two discoveries could then be developed as a single project that is able to justify the cost of the
pipeline. The investment decision following the second successful exploration well is then based
on proceeding with the development of the two accumulations simultaneously using shared
facilities (the pipeline), and the combined development plan then constitutes the project. Again,
the key is that the project is defined by the basis on which the investment decision is made.
Similarly, a discovered accumulation, following successful exploration, may initially be
assessed as a single development opportunity, but upon further evaluation, it may then be split into
two or more (i.e., exploratory/appraisal and development) distinct projects. For example, the level
of uncertainty (e.g., in reservoir performance) may be such that it is considered more prudent to
implement a pilot project first. The initial concept of a single field development project then
becomes two separate projects: the pilot project and the subsequent development of the remainder
of the field, with the latter project contingent on the successful outcome of the first project.
A key strength of using a project-based system such as PRMS is that it encourages the
consideration of all possible technically feasible opportunities to maximize project outcomes, even
though some projects may not be economically viable when initially evaluated. These projects are
still part of the portfolio, and identifying and classifying them ensures that they remain visible as
potential opportunities for the future.
The quantities that are classified as Unrecoverable should be limited to those that are currently
not technically recoverable by a defined project based on established technology or technology
under development (see Section 2.3). A portion of these Unrecoverable quantities may become
recoverable in the future if new technology is developed, commercial circumstances change, or
additional data is acquired.
8
2.3 Project Classification
Under the PRMS, projects must be classified individually so that the estimated recoverable quantities
associated with the project can be correctly assigned to one of the three main classes (see Fig. 2.1).
The distinction between the three classes is based on the definitions of discovery and commerciality,
as documented in PRMS § 2.1.1 and § 2.1.2, respectively. The evaluation of a discovery begins at
the level of the well-penetrated accumulation (including adjacent reservoirs interpreted with
reasonable confidence to be in communication with the well-penetrated reservoir) and determines
whether there is confidence in the presence of significant hydrocarbons. A discovery should be
significant enough to merit assessment as a potential commercial development in order to estimate
discovered petroleum initially in place. It is the existence of “significant hydrocarbons” that justifies
the application of a project (detailed or conceptual) to further differentiate the recoverable from the
unrecoverable resources. (Note: “Unrecoverable” is a subset within the petroleum initially in place
and a part of the resource base, but it is not part of Prospective or Contingent Resources; see
Fig. 2.1.) The project is then evaluated to determine its maturity (commercial or sub-commercial)
and classification of the project’s recoverable resource quantity.
The definition of “discovery” requires evidence (testing, sampling, seismic and/or logging
data) from at least one well penetration in the accumulation (or group of penetrated accumulations)
to have demonstrated a “significant quantity of potentially recoverable hydrocarbons” (PRMS §
2.1.1.1). In the absence of a flow test or sampling, the discovery determination requires confidence
in the presence of hydrocarbons and evidence of producibility, which may be supported by suitable
producing analogs. In this context, “significant quantity” implies that there is evidence of a
sufficient quantity of petroleum to justify estimating the in-place quantity demonstrated by the
well(s) and evaluating the potential for commercial recovery.
Estimated recoverable quantities from a discovery are classified as Contingent Resources until
such time that a defined project can be shown to have satisfied all the criteria necessary to
reclassify some or all of the quantities as Reserves (i.e., commerciality is achieved). In cases where
the discovery is, for example, adjacent to existing infrastructure with sufficient excess capacity,
and a commercially viable development project is immediately evident (e.g., by connecting the
discovery well into the available infrastructure) with firm intent by the entity to proceed with
development, then the estimated recoverable quantities may be classified as Reserves. More
commonly, the estimated recoverable quantities for a discovery will be classified as Contingent
Resources while further appraisal and/or evaluation is carried out.
Discovered accumulations may contain in-place quantities that are not considered viable to
recover using established technology or technology under development; such quantities may be
classified as Discovered Unrecoverable (i.e., no Contingent Resources). Future technological
advances or improvements in commercial circumstances may move those Unrecoverable
quantities to Contingent Resources or Reserves (PRMS § 2.1.1.2).
The criteria for commerciality (and hence assigning Reserves to a project) are set out in the
PRMS § 2.1.2.1 and should be considered carefully. While estimates of Reserve quantities may
change with time, including during the period before production startup, it should be a rare event
for a project that had been assigned to the Reserves class to be reclassified to Contingent
Resources. Such a reclassification should occur only as the consequence of an unforeseeable
event that is beyond the control of the entity, such as an unexpected political, legal, or market
(including pricing) change that, for example, causes development activities to be delayed beyond
a reasonable time frame, which is typically considered to be five years, unless a longer time
frame is justifiable (PRMS § 2.1.2.3). Even so, if there are any identifiable areas of concern
9
regarding receipt of all the necessary approvals/contracts for a new development, it is
recommended that the project remain in the Contingent Resources class until such time that the
specific concern has been addressed.
For projects not yet deemed commercial, Contingent Resources may be assigned if the
recoverable quantity is dependent on either “established technology” or “technology under
development.” The following guidelines should be used to distinguish these Contingent Resources
from those significant quantities that should be classified as Discovered Unrecoverable:
• The technology has been demonstrated to be commercially viable in analogous reservoirs;
in this case, the Discovered Recoverable quantities may be classified as Contingent
Resources.
• The technology has been demonstrated to be commercially viable in other reservoirs that
are not analogous, and a pilot project will be necessary to demonstrate commerciality for
this reservoir.
o If a pilot project is conducted and deemed technically successful, then Discovered
Recoverable quantities from the full project may be classified as Contingent
Resources.
o If a pilot project is conducted and deemed technically unsuccessful, then all
quantities should be classified as Discovered Unrecoverable.
• The technology has not been demonstrated to be commercially viable but is currently
under active development, and there is sufficient direct evidence (e.g., from a test project
in an analogous reservoir) to indicate that it may reasonably be expected to be available
for commercial application. In this case, Discovered Recoverable quantities from the full
project may be classified as Contingent Resources.
• The technology has not been demonstrated to be viable and is not currently under active
development; in this case, all quantities should be classified as Discovered Unrecoverable.
2.4 Range of Uncertainty Categorization

The “range of uncertainty” (see Fig. 2.1 horizontal axis) reflects a range of estimated quantities
potentially recoverable from an accumulation (or group of accumulations) by a specific, defined
project. Because all potentially recoverable quantities are estimates that are based on assumptions
regarding future reservoir performance (among other things), there will always be some
uncertainty in the estimate of the recoverable quantity resulting from the implementation of a
specific project. There will be uncertainty in both the estimated in-place quantities and in the
recovery efficiency, and there may also be project-specific commercial uncertainties. Where
performance-based estimates are used (e.g., based on decline curve analysis), there will be some
uncertainty. However, for very mature projects or constrained market access, the level of technical
uncertainty may be relatively minor; in such cases, the best estimate scenario may be justifiable to
also use for the Low and High estimate scenarios (PRMS § 4.1.4.3).
In the PRMS, once the project has been classified as Prospective Resources, Contingent
Resources, or Reserves based on the project’s level of commercial maturity (see Fig. 2.1 vertical
axis), the range of uncertainty then determines the categorization of the estimated recoverable
quantities, and it is characterized by three specific outcomes reflecting Low, Best, and High (or
P90, P50, and P10 if probabilistic methods are used) estimates from the project. For example, if
the project satisfies all the criteria for Reserves, and the Low, Best, and High estimates are
economically viable, then these estimates will be designated as Proved (1P), Proved plus Probable
(2P), and Proved plus Probable plus Possible (3P), respectively.
10
The three estimates (Low, Best, and High) may be based on deterministic or probabilistic
methods, as discussed in Section 2.5, and they will reflect the range of uncertainty of project
outcomes. While estimates may be made using deterministic or probabilistic methods (or, for that
matter, using multiscenario methods), the underlying principles must be the same if comparable
results are to be achieved. It is useful, therefore, to keep in mind certain characteristics of the
probabilistic method when applying a deterministic approach:
• The range of uncertainty relates to the uncertainty in the estimate of recoverable quantities
for a specific project. The full range of uncertainty extends from a minimum estimated
recoverable quantity for the project through all potential outcomes up to a maximum
recoverable quantity. Because the minimum and maximum outcomes are the extreme
cases, it is considered more practical to use Low and High estimates as a reasonable
representation of the range of uncertainty in the estimate of recoverable quantities. Where
probabilistic methods are used, the P90 and P10 outcomes are typically selected for the
Low and High estimates, respectively, although the official requirement is that the
selected case should have at least a 90% probability that the quantities actually recovered
will equal or exceed the Low estimate and at least a 10% probability that the quantities
actually recovered will equal or exceed the High estimate (see PRMS § 2.2.1.2).
• In the probabilistic method, probabilities actually correspond to ranges of outcomes,
rather than to a specific scenario. The P90 estimate, for example, corresponds to the
situation wherein there is an estimated 90% probability that the quantities actually
recovered will lie somewhere between the P90 and the P0 (maximum) outcomes.
Obviously, there is a corresponding 10% probability that the quantities recovered will lie
between the P90 and the P100 (minimum) outcomes, assuming of course that the
evaluation of the full range of uncertainty is valid. In a deterministic context, “a high
degree of confidence that the quantities will be recovered” (PRMS § 2.2.2.8) does not
mean that there is a high probability that the exact quantity designated as Proved will be
the actual quantity recovered; it means there is a high degree of confidence that the actual
quantity recovered will equal or exceed the Proved amount.
• In this uncertainty-based approach, a deterministic estimate is a single discrete scenario
that should lie within the range that would be generated by a probabilistic analysis. The
range of uncertainty reflects our inability to estimate the actual recoverable quantities for
a project exactly, and the Low, Best, and High estimates are simply single discrete
scenarios that are representative of the range of uncertainty in estimating the remaining
recoverable quantities.
As noted earlier, for very mature producing projects or projects with ample supply but
market or license off-take constraints (e.g., the high confidence recoverable quantity exceeds the
off-take constraint), it may be concluded that there is such a small range of uncertainty in
estimated remaining recoverable quantities that 1P, 2P, and 3P Reserves can be assumed to be
equal, and the incremental P2 (Probable) and P3 (Possible) quantities are approximately equal
to zero. Often, this approach is used when a producing well has sufficient long-term production
history such that a forecast based on decline curve analysis is considered to be subject to
relatively little uncertainty (see PRMS § 4.1.4.3). In reality, of course, the range of uncertainty
is never zero (especially when considered in the context of remaining quantities), and any
assumption that the uncertainty is not material to the estimate should be carefully considered,
and the basis for the assumption should be documented.
11
Typically, there always will be a range of uncertainty and, therefore, Low, Best, and High
estimates of recoverable quantities (or a full probabilistic distribution) that characterize the range,
whether for Reserves, Contingent Resources, or Prospective Resources. However, there are
specific circumstances that can lead to having 2P and 3P Reserves, but zero 1P Reserves. For
example, an undeveloped project may satisfy the criteria to be classified as Reserves based on the
Best estimate, but the Low estimate is not economic and therefore fails to qualify as 1P Reserves.
In this circumstance, the entity may record 2P and 3P Reserves for the development project, but
no 1P Reserves (PRMS § 3.1.2.8 and the first sentence of § 2.1.3.7.4). Moreover, in this example,
the Low estimate cannot be reported as a different classification (e.g., the entity cannot report 1C,
2P, and 3P), since this would result in split classification, which is not allowed (PRMS § 2.2.0.4).
2.5 Methods for Estimating the Range of Uncertainty in Recoverable Quantities

There are several different methods commonly used to estimate the range of uncertainty in
recoverable quantities for a project. While the objective of the exercise is to estimate at least three
outcomes (Low, Best, and High estimates of recoverable quantities) that reflect the range of
uncertainty, it is important to recognize that the underlying philosophy must be the same,
regardless of the approach used. In this context “deterministic” methods rely on a single set of
discrete parameters (gross rock volume, average porosity, etc.) that represent a physically
realizable and realistic combination in order to derive a single, specific estimate of recoverable
quantities (e.g., a combination of parameters represents a specific scenario).
Evaluators may choose to apply more than one method to a specific project, especially for
more complex developments. For example, three deterministic scenarios may be selected after
reviewing a Monte Carlo analysis of the same project. The following terminology is recommended
for the primary methods in current use. These methods are discussed in more detail in subsequent
chapters of this document (see Chapter 4—Assessment of Petroleum Resources Using
Deterministic Procedures, and Chapter 7—Probabilistic Reserves Estimation).
2.5.1 Deterministic “Scenario” Method. In this method, three discrete scenarios are developed
that reflect the Low, Best, and High estimates of recoverable quantities. These scenarios must
reflect realistic combinations of parameters with particular care required to ensure that a reasonable
range of values is used for each reservoir property, and one set of input parameters is selected that
best represents the corresponding confidence category for each estimated recoverable quantity. It
is generally not appropriate to combine the Low estimate for each input parameter to determine a
Low case outcome, as this would not represent a realistic Low case scenario (it would be closer to
the absolute minimum possible outcome).
2.5.2 Deterministic “Incremental” Method. This method is often used in mature, spatially
extensive onshore environments. Typically, this approach defines discrete areal (and sometimes
vertical) segments of the accumulation as High, Best, and Low confidence based on considerations
of well spacing and/or geological knowledge (i.e., the different degrees of confidence are governed
by distance from known data) in determining estimates of recoverable quantities under the defined
development plan. For example, Proved Developed Reserves are assigned within the immediate
drilled area, and Proved Undeveloped Reserves are assigned to adjacent, high-confidence drilling
locations based on continuity of the delineated productive reservoir. Probable and Possible
Reserves are assigned to less confident locations, often at a further distance from Proved well
locations; beyond these locations, there may exist Contingent Resources, if they are part of the
12
same discovered accumulation. These additional quantities (e.g., Probable Reserves) are estimated
discretely as opposed to defining a Proved plus Probable Reserves scenario. In such cases,
particular care is required to define the project correctly (e.g., distinguishing between wells that
are planned and committed as opposed to those that are still contingent, whether individual wells
or area development patterns comprise the project) and to ensure that all uncertainties, including
recovery efficiency, are appropriately addressed.
2.5.3 Geostatistical Method. This approach may be used to more reliably describe the spatial
distribution of geoscience and reservoir engineering data, thereby better describing the variability
and uncertainties within an accumulation. By preserving spatial distribution of data, which can be
represented in a static reservoir model and incorporated in a reservoir simulation, these techniques
can provide improved estimates of the range of recoverable quantities. For example, incorporating
seismic analyses can improve the understanding, mapping, and modeling of the static geological
framework and the geostatistical variability of reservoir properties beyond limited well control,
thereby enhancing the resource estimation.
2.5.4 Probabilistic Method. This method examines the probabilistic distribution of the input
parameters and the resource or reserves subsequently estimated. It is commonly implemented
using Monte Carlo simulation or stochastic modelling. The evaluator defines the probability
distributions of the input parameters and the relationships (correlations) between them, and the
model derives an output distribution based on a combination of those input assumptions. Each
iteration of the model is a single, discrete deterministic scenario; yet, in this case, the simulation
model, rather than the evaluator, determines the combination of parameters for each iteration, and
it runs many different possible combinations (often several thousand) in order to develop a full
probability distribution of the range of possible outcomes, from which three representative
outcomes are selected (i.e., P90, P50, and P10).
2.5.5 Integrated Method. Resource assessments can integrate several methods, including
deterministic, geostatistical, and probabilistic, in combination to better define uncertainty and
ensure that the results of the methods are reasonable. An example is the multiscenario method,
which is a combination of the deterministic (scenario) method and the probabilistic method. In this
case, significant numbers of discrete deterministic scenarios are developed by the evaluator
(perhaps 100 or more), and probabilities are assigned to each possible discrete input assumption.
For instance, three depth conversion models may be considered possible, and each one is assigned
a probability based on the evaluator’s assessment of the relative likelihood of each of the models.
Each scenario leads to a single deterministic outcome, and the probabilities for each of the input
parameters are combined to give a probability for that scenario/outcome. Given a sufficient
number of scenarios, it is possible to develop a full probability distribution, from which three
specific deterministic scenarios that are closest to P90, P50, and P10 may be selected.
2.6 Chance of Commerciality

The “chance of commerciality” (see Fig. 2.1 vertical axis) reflects the chance that a project will be
committed for development and reach commercial producing status. As a project moves to a higher
level of commercial maturity, there is typically an increasing chance that the accumulation will be
commercially developed and the project’s recoverable quantities will be moved to Reserves.
13
An evaluator has the option to express commercial risk qualitatively by allocation to classes
and sub-classes (see Section 2.7 on Project Maturity sub-classes) and/or quantitatively as the
chance of commerciality (Pc), which, for undiscovered resources, is defined as the product of two
risk components (PRMS § 2.1.3.2):
• Chance of geologic discovery (Pg): The chance that the potential accumulation will result
in the discovery of a significant quantity of petroleum, and
• Chance of development (Pd): Once discovered, the chance that the known accumulation
will be commercially developed.
For a Prospective Resource, Pc is the product of Pg and Pd (Pc = Pg × Pd), reflecting the risk of
discovery as well as the risk of commercial development. Yet, once a discovery is made (i.e., Pg =
100%), then the chance of commerciality becomes equivalent to the chance of development (Pc =
Pd) for Reserves and Contingent Resources, because these classifications are attributable only to
discovered accumulations. Furthermore, for a project to be classified as Reserves, there should be
a very high probability that it will proceed to commercial development (i.e., very little, if any,
commercial risk). Consequently, commercial risk is generally ignored in the estimation and
reporting of Reserves, although some Reserves may still carry minor commercial risk if final
approvals or contracts (items outside the entity’s control) have not yet been obtained, but they are
deemed, at the time of the Reserve classification, to be reasonably certain to obtain (see Section
2.7 discussion of Justified for Development). However, for projects with Contingent or Prospective
Resources, the commercial risk may be quite significant and should be carefully considered.
Consider first the approach for estimating Prospective Resources. The chance of geologic
discovery (Pg) is assessed based on the probability that all the components necessary for an
accumulation to form (hydrocarbon source, reservoir, trap, migration, etc.) are present. Separately,
an evaluation of the potential size of the discovery is undertaken. Often, this is performed
probabilistically and leads to a “full distribution” of the range of uncertainty in potentially
recoverable quantities, given that a discovery is made. This “full distribution” of potential
outcomes provides the basis for the Low, Best, and High estimates for the full range of recoverable
quantities for Prospective Resources, and it is consistent with the use of the full range of
recoverable quantities in assessing Reserves and Contingent Resources as described in Section 2.4
and below. Because this range may include outcomes that are below the economic threshold for a
commercially viable project, the probability of being above the economic threshold is used to
define the chance of development (Pd), and hence the chance of commerciality (Pc) is obtained by
multiplying Pd by the chance of geologic discovery (Pg).
An alternative approach for estimating Prospective Resources is to include only the
commercially successful portion of the full distribution (i.e., the “success case” portion rather than
the “full distribution”). In this case, the range would exclude outcomes that are below the economic
threshold for a commercially viable project; therefore, the probability of achieving this “success
case” portion of the range with a discovery well is lower, which must be incorporated into the
chance of development (Pd) when using the “success case” as the Prospective Resource range (i.e.,
“full distribution” Pd > “success case” Pd).
In both approaches for determining Prospective Resource ranges (“full distribution” vs.
“success case”), the risked mean Prospective Resource will be the same. That is, achieving the
higher “success case” distribution is riskier than achieving the “full distribution,” which also
includes the uneconomic outcomes; therefore, the risked mean of the two cases will be equal (i.e.,
the higher risk/lower probability applied to the “success case” mean will be equal to the product
of the lower risk/higher probability applied to the “full distribution” mean). The key for the
14
evaluator is to clearly document the basis and assumptions used to estimate the Prospective
Resources and assess the underlying risks related to commercialization (i.e., chance of geologic
discovery, chance of development).
Once a discovery has been made, and a range of technically recoverable quantities has been
assessed for a known accumulation from the “full distribution” of outcomes, these estimated
recoverable quantities (Low, Best, and High) will be assigned as Contingent Resources if there are
any contingencies that currently preclude the project from being classified as commercial (i.e.,
prevent recognition as Reserves). If the contingency is purely nontechnical (such as lack of
environmental approval, for example), the uncertainty in the estimated recoverable quantities
generally will not be impacted by the removal of the contingency. The Contingent Resource
quantities (1C, 2C, and 3C) should then theoretically move directly to 1P, 2P, and 3P Reserves
(PRMS § 2.2.2.6) once this nontechnical contingency is removed, provided of course that the same
economic criteria are applied, all other criteria for assigning Reserves have been satisfied, and the
planned recovery project has not changed in any way. In this example, the chance of commerciality
is the probability that the necessary permits will be obtained.
However, another possible contingency precluding a development decision could be that the
estimated 1C quantities are considered to be too small to commit to the project, even though the
2C level is commercially viable. It is not uncommon, for example, for an entity first to test that the
2C estimate satisfies all their corporate hurdles and then, as a project robustness test, to require
that the Low (1C) outcome at least reaches a break-even economic threshold. If the project fails
this latter test, and development remains contingent on satisfying this break-even test, further data
acquisition (potentially appraisal drilling) would be required to reduce the range of uncertainty. In
such a case, the chance of commerciality is the probability that the appraisal efforts will increase
the Low (1C) estimate above the break-even level, which is not the same as the probability
(assessed before the additional appraisal) that the actual recoverable quantity will exceed the
break-even level.
When reporting Contingent Resource estimates, the commercial risk associated with such
projects can vary widely, with some having a high chance of proceeding to development, while
others might have a much lower chance of being developed. If Contingent Resources are reported
externally, the commercial risk can be communicated to users (e.g., stakeholders) by various
means, including: describing the specific contingencies associated with individual projects;
reporting a quantitative chance of commerciality for each project; and/or assigning each project to
one of the Project Maturity sub-classes (see Section 2.7).
Aggregation of quantities that are subject to commercial risk raises further complications and
is discussed in more detail in Chapter 8—Aggregation of Reserves and Resources in this document.
2.7 Project Maturity Sub-Classes

Under the PRMS, identified recoverable petroleum-initially-in-place must always be assigned to
one of the three classes: Reserves, Contingent Resources, or Prospective Resources. Further
subdivision is optional, and three options are provided in PRMS that can be used together or
separately to identify particular characteristics of the project and its associated recoverable
quantities. These options are Project Maturity sub-classes as described below, Reserves Status (see
Section 2.8), and Economic Status (see Section 2.9).
As illustrated in Fig. 2.3, development projects (and their associated recoverable quantities)
may be sub-classified according to project maturity levels and the associated actions (e.g., business
decisions, data acquisition, evaluation) required to move a project toward commercial production.
15
This approach supports management of portfolios of opportunities at various stages of exploration,
appraisal, and development and may be supplemented by associated quantitative estimates of
chance of commerciality, as discussed in Section 2.6. The boundaries between different levels of
project maturity may align with internal (entity-specific) project “decision gates,” thus providing
a direct link between the decision-making process within an entity and characterization of its
portfolio through resource classification. This link can also act to facilitate the consistent
assignment of appropriate quantified risk factors for the chance of commerciality.
Fig. 2.3—Sub-classes based on project maturity.
Evaluators may adopt alternative sub-classes and project maturity modifiers to align with their
own decision-making process, but the concept of increasing chance of commerciality should be a
key enabler in applying the overall classification system and supporting portfolio management.
Note that, in quantitative terms, the “chance of commerciality” axis shown in Figs. 2.1 and 2.3 is
not intended to represent a linear scale, nor is it necessarily wholly sequential in the sense that a
Contingent Resource project that is classified as “Development Not Viable” could have a lower
chance of commerciality than a low-risk Prospect, for example. In general, however, quantitative
estimates of the chance of commerciality will increase as a project moves “up the ladder” from an
exploration concept to a field that is producing.
If the sub-classes in Fig. 2.3 are adopted, the following general guidelines should be
considered in addition to those documented in Table 1 of the PRMS:
• On Production refers to a development project that is producing or is capable of
producing and selling petroleum to the market at the effective date of the evaluation.
Implementation of the project may not be 100% complete at that date, and so some of the
reserves may still be Undeveloped (see Section 2.8). In some cases, the project may have
advanced past the Approved for Development phase and is capable of production, but
production has not yet commenced due to short-term circumstances outside of the project.
In these cases, the project is still considered On Production. Likewise, in situations where
16
production is temporarily interrupted (i.e., shut-in) and requires minor costs to restart, the
project would still be considered On Production. Alternatively, if the interruption requires
significant additional capital investment or the time frame to resume production is not
certain to occur, then the project no longer has Proved Developed Reserves and should
be reclassified to the appropriate sub-class.
• Approved for Development requires technical maturity of the project and all elements
of commerciality to be in place (i.e., approvals are obtained, capital funds are committed).
Construction and installation of project facilities should be under way or due to start
imminently. Only a completely unforeseeable change in circumstances that is beyond the
control of the developers would be an acceptable reason for failure of the project to be
developed within a reasonable time frame.
• Justified for Development essentially covers the period when the operator and its
partners have agreed that the defined project is commercially viable, and they have
decided to proceed with development on the basis of an agreed development plan (i.e.,
there is a “firm intent”), but before the project has reached the Approved for Development
sub-class, when all necessary approvals are in place (e.g., regulatory approval of the
development plan) and a “final investment decision” has been made by the partners to
commit the necessary capital funds. “Firm intent,” in this context, is supported by
evidence of alignment to proceed by all participating entities, with no known
contingencies that would preclude development from proceeding at the appropriate time
and within a reasonable time frame. In the PRMS, the recommended benchmark is that
development would be expected to be initiated within five years of assignment as
Reserves to this sub-class, unless a longer time frame is justified; for example, a remote
development location, late-life compression installation, plateau extension, or
infrastructure capacity constraint drives a longer time frame. In all cases, the justification
for longer duration should be critically reviewed and documented (PRMS § 2.1.2.3 and
§ 2.1.3.6.4). It is worth noting that most projects progress directly from Development
Pending into Approved for Development once all approvals and funds are secured.
• Development Pending is limited to those projects that are actively subject to project-
specific activities, such as appraisal drilling or detailed evaluation, that are designed to
confirm commerciality and/or to determine the optimum development scenario. In
addition, it may include projects that have nontechnical contingencies, provided these
contingencies are currently being actively pursued by the developers and are expected to
be resolved positively within a reasonable time frame. Such projects would be expected
to have a high probability of becoming a commercial development (i.e., a high chance of
commerciality).
• Development On Hold is typified when a development plan has been identified, and the
project is considered to have at least a reasonable chance of commerciality, but there are
contingencies that need to be resolved before the project can move toward development.
The contingencies may be either internal or external (examples: lack of funding,
uncertainty of obtaining necessary permits). The primary difference between
Development Pending and Development On Hold is that in the former case, the remaining
contingencies are being addressed (e.g., data collection, negotiations) and are reasonably
expected to be resolved within a reasonable time frame, whereas in the latter case,
resolution of the primary contingencies may be seen less favorably and be subject to a
significant time delay (e.g., technology advancement, market development, regulatory
17
policy progression). Any change in circumstances, such that there is no longer a probable
chance that a critical contingency can be removed in the foreseeable future, could lead to
a reclassification of the project to a Development Not Viable sub-class.
• Development Unclarified is for discovered accumulations where additional data and
analysis are required to define a development plan, and appraisal activities are ongoing
to clarify the potential for development (e.g., a recent discovery). Contingencies have yet
to be fully defined, and the chance of commerciality may be difficult to assess with any
confidence.
• Development Not Viable is assigned when a technically viable project has been assessed
as being of insufficient potential to warrant development, acquiring additional data, or
any further efforts to remove contingencies. However, the estimated recoverable
quantities are recorded so that the potential development opportunity will be recognized
in the event of a major change in technology and/or commercial conditions. Projects in
this sub-class would be expected to have a low chance of commerciality.
It is important to note that while the aim is always to move projects to the Reserves class with
higher levels of maturity, and eventually to production, a change in circumstances (disappointing
well results, change in fiscal regime, etc.) can lead to projects being reclassified to a lower sub-
class. Further, projects may not necessarily pass sequentially through each sub-class as they move
up or down the y-axis of commerciality.
One area of possible confusion is the distinction between Development Not Viable and
Discovered Unrecoverable. A key goal of portfolio management should be to identify potential
incremental development options for a reservoir, so it is strongly recommended that all technically
feasible projects (i.e., based on established technology or technology under development) that
could be applied to a reservoir are identified even though some may not be economically viable at
the time. Such an approach highlights the extent to which identified incremental development
projects would achieve a level of recovery efficiency that is at least comparable to analogous
reservoirs. Looking at it from the other direction, if analogous reservoirs are achieving levels of
recovery efficiency significantly better than the reservoir under consideration, it is possible that
there are development options that have been overlooked.
Quantities should be classified as Discovered Unrecoverable only if no technically feasible
projects have been identified that could lead to the recovery of any of these quantities (PRMS §
2.1.1.2). A portion of Unrecoverable quantities may become recoverable in the future as
commercial circumstances change, technology is developed, or additional data are acquired; the
remaining portion may never be recovered due to physical/chemical constraints represented by
subsurface interaction of fluids and reservoir rocks.
Section 2.11 herein presents example applications of the Project Maturity sub-classes.
2.8 Reserves Status

Estimated recoverable quantities associated with projects that fully satisfy the requirements for
Reserves may be subdivided according to their operational and funding status.
Under the PRMS, subdivision by Reserves Status is optional and includes the following status
levels: Developed Producing, Developed Non-Producing, and Undeveloped. These subdivisions
may be applied to all categories of Reserves (i.e., Proved, Probable, and Possible).
Reserves Status has long been used as a subdivision of Reserves in certain environments, and
it is obligatory under some reporting regulations to subdivide Proved Reserves into Proved
Developed and Proved Undeveloped. In many other areas, subdivision by Reserves Status is not
18
required by relevant reporting regulations and is not widely used by evaluators. Unless mandated
by regulation, it is up to the evaluator to determine the usefulness of these, or any of the other,
subdivisions in any particular situation.
Subdivision by Reserves Status or by Project Maturity sub-class is optional, and, because they
are to some degree independent of each other, both can be applied together. Such an approach
requires some care, because it is possible to confuse the fact that Project Maturity sub-classes are
linked to the status of the project as a whole, whereas Reserves Status considers the level of
implementation of the project, essentially on a well-by-well basis. Therefore, unless each well
constitutes a separate project, Reserves Status is a subdivision of Reserves within a project.
The relationship between the two optional classification approaches may be best understood
by considering all the possible combinations, as illustrated below. Table 2.1 shows that a project
that is On Production could have Reserves in all three Reserves Status subdivisions, but it is the
only sub-class with Developed Producing status, whereas all project Reserves must be
Undeveloped if the project is classified as Justified for Development.
Reserves Status
Project Maturity Sub-Class Developed Producing Developed Non-Producing
Undeveloped Reserves
Reserves Reserves
On Production   
Approved for Development

  
Justified for Development
  
Table 2.1—Project Maturity sub-class relative to Reserves Status sub-class.
Applying Reserves Status in the absence of Project Maturity sub-classes can lead to the
mixing of different types of Undeveloped Reserves and may hide the fact that they are subject
to different levels of project maturity. By using Project Maturity sub-classes, a clear distinction
can be made among:
• Those Reserves from “On Production” projects that remain Undeveloped because,
although production has begun or is capable of producing from completed wells, some of
the project wells have yet to be drilled at the date of the evaluation;
• Those Reserves from Approved for Development projects that remain Undeveloped
because implementation of the approved, committed, and budgeted development project
is ongoing, and drilling of the production wells, for example, is still in progress at the
date of the evaluation; and
• Those Reserves from Justified for Development projects that remain Undeveloped while
awaiting final approvals or contracts to be obtained before proceeding with development
on the basis of an agreed development plan (i.e., there is a “firm intent” with no known
contingencies that will preclude development from proceeding at the appropriate time
and within a reasonable time frame).
For portfolio analysis and decision-making purposes, it is clearly important to be able to
distinguish between these three types of Undeveloped Reserves with different levels of
project maturity.
Fig. 2.4 illustrates how a typical project stage gate process can be related to the PRMS (e.g.,
status, classification, chance of development).
19
Fig. 2.4—Project stage gate sub-classes based on project maturity.
2.9 Economic Status

A third option for classification purposes is to subdivide Contingent Resource projects on the basis
of economic status into Economically Viable or Economically Not Viable Contingent Resources;
these sub-classes are based on assumed reasonable forecast conditions. In addition, the PRMS
indicates that, where evaluations are incomplete, such that it is premature to clearly assess the
project economics, it is acceptable to note that the project economic status is “undetermined”
(PRMS § 2.1.3.7.6). As with the classification options for Reserves that are based on Reserves
Status, this is an optional subdivision that may be used alone or in combination with Project
Maturity sub-classes.
Broadly speaking, one might expect the following approximate (not necessarily equivalent)
relationships between the two optional approaches (Table 2.2):
Project Maturity Economic

Sub-Class Status Sub-Class
Development Pending
Economically Viable Contingent Resources

Development On Hold
Development Unclarified Undetermined
Development Not Viable Economically Not Viable Contingent Resources
Table 2.2—Approximate Project Maturity sub-class relative to Economic Status sub-class.
20
2.10 Estimated Ultimate Recovery and Technically Recoverable Resources
The PRMS (2018) has introduced “Estimated Ultimate Recovery” and “Technically Recoverable
Resources” (EUR and TRR, respectively), commonly used terms in the industry, in its Glossary in
an effort to avoid confusion with prior PRMS terminology (PRMS § 1.1.0.8).
EUR is an estimate of recoverable quantities of petroleum (including quantities already
produced) from an accumulation(s), whether discovered or undiscovered. It is not a Reserve or
Resource category or class, but rather a way to simply describe the quantity estimated to be
recoverable based on defined conditions (for clarity, EUR must reference the associated technical
and commercial conditions applied; e.g., proved EUR is Proved Reserves plus prior production).
Care must be taken to ensure that consistent measurement (e.g., raw, marketable, sales quantities)
between cumulative and forecast quantities is used for EUR.
TRR, on the other hand, are the estimated remaining technically recoverable quantity from an
accumulation(s), whether discovered or undiscovered, and it is based on the volumes expected to be
recovered from the project(s) being implemented or planned. TRR is not constrained by commercial
conditions and does not include previously produced quantities (e.g., it is a forward-looking estimate
of remaining technically recoverable quantities). The term “TRR” appears to have originated in
governmental and geological assessments that were designed to reflect a nation’s ability to marshal
its hydrocarbon endowment in the event of a crisis. Therefore, TRR is an estimate of future
technically recoverable quantities and does not include historical cumulative production.
For an EUR, there may be a “technical recovery” estimate, which could differ from a TRR
estimate. This may happen, for example, when the technology applied in the EUR estimate is
established technology, while the TRR estimate also relies on technology under development.
“Technology under development” and “established technology” are considered currently available
technology for TRR and can both be used to estimate technically recoverable Contingent and
Prospective Resources. In all cases, any statement of EUR and TRR requires the conditions
associated with their usage to be clearly noted and documented.
The following simple example illustrates some of the concepts associated with developing an
EUR or TRR estimate. In Fig. 2.5, a well’s historical production rates have been fitted with a
hyperbolic decline forecast extrapolated to the technical limit (considering factors such as water
handling, lift capacity, etc.). On the right-side axis, cumulative production is recorded. At the
technical limit, the corresponding cumulative recovery is shown as the “Technical EUR” or, as
denoted on the cumulative production line, “EURtech.” However, the economic limit of production
is reached before the technical limit, and we have a “Commercial EUR” or “EURcom.”
Notice that the “Cumulative Production” line starts at the left-side origin (“Initial Production”)
and represents a running total recovery from the beginning; i.e., the EUR figure will include the
production prior to the “Evaluation Date” shown on the x-axis.
The area under the production forecast curve between the Evaluation Date and the Economic
Limit Date has been highlighted and designated as “Reserves.” (The quantity has satisfied the
economic criterion, and, for the purpose of this example, we assume that the other commerciality
criteria are met.) At the “Evaluation Date” on the x-axis, another cumulative line (“Future
Cumulative Production”) begins and is extrapolated, using the Production Forecast, to both the
Economic and the Technical Limits.
21
Fig. 2.5—Schematic comparison of Estimated Ultimate Recovery (EUR), Technically Recoverable Resources (TRR),
and Reserves.
The area under the production forecast curve between the Economic Limit Date and the
Technical Limit Date should be designated as “Contingent Resources.” As pointed out before, a
single project may not contain recoverable quantities in different resources classes under the
PRMS because this constitutes split classification. In a situation where the product pricing forecast
is unchanged, lowering the economic limit to the technical limit will necessitate some form of cost
reduction or facility modification, for example, thereby extending the economic life. This project
would normally require investment decisions (such as implementing new or changing out artificial
lift methods) and is not part of the existing project development plan.
In this example, the Economic and Technical Limits for the EUR case are assumed to be the same
for the TRR case, although they do not need to be, depending on, for example, the technology
employed. Again, assuming the two curves are based on the same applied technology, Fig. 2.5 has two
cumulative production lines, one for the EUR and one for the TRR (which excludes prior production),
and these lines are parallel but shifted by the cumulative recovery at the Evaluation Date.
Therefore, if we take the EURcom and subtract the cumulative production at the Evaluation
Date, then we arrive at the “Reserves” (if the production forecast is a 2P projection, then the
Reserves are the 2P quantities; in this situation, the EUR may be designated as “EUR 2P,” because
it is the sum of the cumulative recovery as of the Evaluation Date plus the 2P projection). Likewise,
if we subtract the cumulative recovery at the Evaluation Date from the EURtech, then we obtain the
TRR, as shown in Table 2.3 below, again using the same production forecast, applied technology,
etc., for both the EUR and TRR projections.
Furthermore, if Fig. 2.5 represents a project that has a cumulative production of 100 (units of
your choice) as of the evaluation date with a production forecast that predicts an EUR (to the
technical limit of production) of 250 units, then Table 2.3 indicates that the project’s TRR (at the
evaluation date) is 150 units. Considering economic parameters, the EURcom (to the economic
limit) is 175 units, which is 75 units less than the EURtech. The Reserves (being the remaining
commercial quantities) would then be 175 – 100 units, i.e., 75 units at the evaluation date.
22
Cumulative Production at Evaluation Date 100
EURtech 250
TRR 150 EURtech – Cumulative Production *
EURcom 175
Reserves 75 EURcom – Cumulative Production *
* For this example, this is not always true (see text).

All quantities must be measured on the same basis (e.g., raw, sales quantities, etc.), but “Reserves” are recommended to be sales
quantities per the PRMS.
Table 2.3—Comparison of recoveries.
Other adjustments that can be made to these projections include consideration of royalties,
license term, variation in applied technologies, other commercial criteria, and so on.
2.11 Example Application of Commercial Maturity Sub-Classes

Five coalbed methane exploration/appraisal wells were drilled in the southern part of a license
block in 2017. In 2018, a production pilot project using these wells commenced with the
commissioning of flowlines and a compression station selling gas to pressurized-gas filling
stations in the vicinity under a 15-year gas contract.
In early 2019, five new wells were drilled and completed in the northern part of the license
(Fig. 2.6). Coal characteristics as evaluated from logs were found to be similar to the southern
pilot wells. A decision was made to not acquire core data. The producing company decided to
complete these wells due to the similarity to the pilot geology, but without testing them. Contracts
to sell the gas from all five wells were negotiated. Compression units and flowlines were included
in the budget as part of the business plan.
However, development of the 2019 wells faced opposition from the communities living in this
remote area close to a national park. Public demonstrations and complaints filed with the local
council resulted in delays to the permit to produce and pending hearings.
Nevertheless, the company committed to drill 20 new wells in five years from 2020 to 2024 as
part of a Phase II development of the southern area. The first four wells were drilled in 2020, and
another six wells were scheduled to be drilled in 2021. The last 10 wells will be drilled thereafter.
23
Log data from the 2020 wells showed geological and petrophysical properties similar to the 2017
southern pilot wells. It was decided not to test the wells but to complete them as producers. As of the
31 December 2020 evaluation date, the first four wells were ready for production.
Fifteen-year sales contracts are being negotiated to secure gas sales from the first 10 wells
from Phase II. The wells will provide gas to several gas filling stations, and new gas stations have
been approved to be built in 2022 to take gas from the remaining 10 wells.
The following questions pertaining to the above scenario reference Fig. 2.7, which is
duplicated from PRMS Figure 2.1:
• Can the Entity Claim Any Reserves from the Five Southern Area (Pilot) Wells? This
development project is currently producing under a 15-year sales contract to an existing
market. Assuming the entity would not embark on development of these wells without
having satisfied all commerciality criteria, Reserves (maturity sub-class On Production)
may be assigned for the 15-year contract duration. Beyond that date, any further quantities
would be classified as Contingent Resources, with the maturity sub-class being either
Development Pending or Development On Hold, depending on the likelihood of
extending the gas contract.
• Can the Entity Claim Reserves from the Five Northern Area Wells? The wells have been
completed, and facilities have been committed in the budget; a market exists, as
evidenced by the contract negotiations; however, the public opposition poses a
contingency to development. Although the northern area project could be considered as
Reserves with a maturity sub-class of Approved for Development, the PRMS notes that
this sub-class requires that the project must not be subject to any contingencies, such as
outstanding regulatory approvals or sales contracts. Similarly, a requirement for the
Justified for Development sub-class states there is reasonable expectations that all
necessary approvals or contracts will be obtained and there must be no known
contingencies that preclude the development from proceeding. The public protest may
stop the project from moving forward, and as such, the northern area project should be
classified as Contingent Resources, sub-class Development On Hold. A Contingent
Resources, Development Pending sub-class is not recommended due to the prospect that
“critical contingencies…are reasonably expected to be resolved within a reasonable time
frame,” which, at this point, has not been determined.
• Can the Entity Claim Reserves for the Phase II 2020 Wells? How about for the Phase II
2021 Wells? Sales contracts are being negotiated and will likely be concluded without
contingencies; the entity has already drilled four wells and is committed to drill six 2021
wells, with a further 10 wells drilled thereafter. The first four wells, although not currently
on production, are capable of producing and selling petroleum to market, and therefore
they qualify as Reserves (Developed status) with the sub-class On Production. The second
six wells have yet to be drilled, and so they have Undeveloped Reserves status with the
sub-class Approved for Development, because all necessary approvals have been
obtained, capital funds have been committed, and implementation of the development
project is under way. It must be noted that these second six wells will be drilled at a tighter
well spacing than the previous four wells drilled in 2020 and the original five wells drilled
in 2017; therefore, the reserve assessment must account for the risk of well interference
and potential lower recovery per well resulting from the tighter spacing.
• Can the Entity Claim Reserves for the Undrilled Post-2021 Phase II Wells? The
development project is commercially viable at the time of reporting, based on the
24
reporting entity’s assumptions (Forecast Case). There is a firm intention to proceed with
development within a five-year window. The development plan is well defined, and there
is “Reasonable Expectation” that the sales contract required prior to project
implementation will be forthcoming. There are no known contingencies that could
preclude the development from proceeding. The last 10 Phase II wells have yet to be
drilled, and so they have Undeveloped Reserves status with the sub-class Justified for
Development. They do not meet the requirements for Approved for Development
established by the decision gate that specifies the project has reached a level of technical
and commercial maturity sufficient to justify proceeding with development. As stated
above, these remaining wells will be drilled as infills, and at tighter well spacing than the
current wells, so the reserve assessment must account for the risk of well interference and
resulting potential for lower recovery per well.
Fig. 2.7—Sub-classes based on maturity.
2.12 Acknowledgments
The authors wish to acknowledge the contribution of James G. Ross, author of this chapter in the
previous version of the Guidelines for Application of the PRMS. In addition, important feedback
and editorial effort have been provided by Charles Vanorsdale, Dan DiLuzio, Carolina Coll,
Bernard Seiller, and Ian McDonald.
2.13 Reference
Petroleum Resources Management System, Version 1.01. 2018. Richardson, Texas, USA: Society
of Petroleum Engineers.
25
Chapter 3
Seismic Applications
Chuandong “Richard” Xu (Chair)
Dan Maguire, Andrew Royle, David Johnston, Eric Von Lunen,
and Jean-Pierre Blangy
3.1 Introduction and Overview

Geophysical methods, and especially surface and borehole seismic surveys, have been used for
decades by the petroleum industry for exploration and development. As a cornerstone subsurface
discipline, the field of geophysics offers practical tools with which to describe and characterize the
subsurface, including the locations and the volumes of in-place hydrocarbons.
In general, geophysical applications contribute to the definition of the technical basis for the
estimation of resources and their volumetric uncertainty in the subsurface. Technical advances
continue to be made as a result of improvements in all facets of the seismic data value chain:
acquisition, seismic processing, imaging, advanced signal data analysis, inversion, and
interpretation. The quality of seismic data has improved through time, as much as the ability of
experienced professionals to associate seismic responses with reservoir and fluid properties under
both static and dynamic conditions. This has resulted in notable improvements in the reliability of
seismic technology to support the mapping and modeling of hydrocarbon accumulations.
Advances in areas such as full waveform inversion have resulted in higher-fidelity understanding
of the three-dimensional (3D) velocity field, which allows for improved estimation of the shape of
reservoirs and reduced uncertainty in depth predictions.
When seismic inversion and time-lapse [four-dimensional (4D)] seismic results are linked to
geostatistical modeling and to assisted history matching of dynamic models, geoscientists and
engineers are now able to build more robust ranges of uncertainty in the quantities of recoverable
hydrocarbons. This linkage enables engineers and geoscientists to more efficiently and accurately
build and test integrated geologic models and reservoir simulation models that agree with well,
seismic, and production data. These models, which inform the uncertainty space, are based on
multiple scenarios of 3D reservoir architecture and tie fluid movement in time through a field
development plan. They help in defining ranges of technically recoverable resources from a field,
based on the existing well stock, in addition to possible future production or injection wells.
This chapter introduces and explains the application of seismic methods in reserves and
resources assessment in line with the PRMS guidelines. There is a vocabulary associated with
geophysical concepts in addition to the application and interpretation of these concepts, which are
beyond the scope of this brief chapter. The reader is encouraged to refer to the Glossary at the end
of this volume, and to seek deeper understanding from their geophysicist colleagues, references
within this chapter, and the Society of Exploration Geophysicists as well.
3.2 Shared Earth Modeling

A Shared Earth Model (SEM), also called a Common Earth Model, is a multidisciplinary
computerized digital 3D representation of Earth’s subsurface containing a petroleum reservoir
system. It can be used to determine the size and uncertainty of total petroleum initially in place
(PIIP) in the system, predict production performance for wells, and optimize the field development.
This model is created by inputting all the available data from geology, geophysics, petrophysics,
26
and reservoir engineering, in different scales, and mathematically integrating these data sets.
Besides representing the subsurface as realistically as possible, the SEM should be updated
constantly to stay current as new data are acquired and imported.
The SEM can be subdivided into two types: a static model and a dynamic model. In general,
geologists, geophysicists, and petrophysicists mainly work on the geologic framework or static
model (geomodel), which is about geology, rock, and pore matrix. Reservoir engineers and
petrophysicists mainly work on the dynamic model (or reservoir model), which is about behavior of
fluids in the matrix. A reservoir simulation usually runs on these models to estimate the pore fluid
dynamics and to predict the field performance under a series of different producing strategies. (See
Chapter 6—Reservoir Simulation, herein, for more detail on the aspects of dynamic simulation.)
In the last decade, with the rapid advances of modeling software, computational hardware,
and information technologies, it is more common and practical for companies to build an SEM to
manage the development, performance, and life cycle of oil and gas reservoirs, as well as for the
purpose of improving estimation of reserves/resources and making decisions.
3.2.1 Static Model—Geologic Framework Architecture. A static model is often called a

geologic model. It integrates all available geological, geophysical, and petrophysical data and is
constructed to focus on the structure (including horizons, faults, folds, and unconformities), rock
properties (including lithology, clay content, porosity, pore fluids saturation, permeability, natural
fractures, and distribution), reservoir properties (including thickness, net-to-gross ratio, and
geometry), depositional environment and stratigraphy (including sequence stratigraphy, facies
distribution, and petroleum system), and other geological features pertinent to the model. Those
components are relatively static during the production period. A static model represents the
subsurface geology and container of hydrocarbons with calculatable properties and uncertainties
for resource estimation and classification.
A 3D seismic survey samples the 3D space of the subsurface evenly in both lateral and vertical
directions. A grid of multiple two-dimensional (2D) seismic lines samples the subsurface evenly
in the line direction and vertical direction but may have large gaps between lines. In contrast, a
drilled well samples the subsurface at a single location, with much higher resolution along the
wellbore trajectory but much less detection and resolution laterally away from the wellbore. Being
able to “see” the entire picture of the subsurface makes seismic surveys natural and effective tools
to define the medium- to large-scale features for constructing the framework of the static model
beyond the existing wells. After carefully correlating seismic data with well data and time-depth
conversion, the accuracy of depth from seismic interpretation could be as high as 1 to 2 m (3 to 6
ft) for clean and less noisy, shallow data.
The process of seismic inversion turns seismic trace amplitude information into rock
properties by using well data as input. The result from this process could be used to fill the
framework of the static model constructed from seismic interpretation. The advantage of seismic
inversion is that the rock properties in the inter-well space are transformed from seismic
measurements to petrophysically tied rock and fluid properties, rather than relying on a
mathematical interpolation of the well data, especially when two wells are far apart.
3.2.2 Dynamic Model—Reservoir Simulation. A dynamic model may be upscaled from the static
model to integrate all available reservoir engineering and petrophysical data. A good dynamic
model needs to be based on a good static model.
27
Reservoir simulation combines geological, geophysical, engineering, and production data to
model fluid saturation and pressure changes, primarily, as the reservoir is produced. These
dynamic reservoir models are typically used to influence field development decisions and to
estimate reserves, as further discussed in Chapter 6—Reservoir Simulation. During field
production, history matching is used to validate the underlying SEM, to fine-tune volume estimates,
and to better predict future reservoir performance.
Time-lapse (4D) seismic data, acquired during field production, are used to inform both static and
dynamic models. The 4D seismic data provide additional constraints on structure and static properties
such as lithofacies, reservoir continuity, and stratigraphy. In a process called seismic history matching,
4D seismic data also help to determine reservoir compartmentalization due to lateral facies changes
and faults, and to constrain properties such as permeability and fault transmissibility.
3.3 Volumetrics, Trap Geometry, and Gross Rock Volume

3.3.1 Defining Seismic Data. Volumetric analysis is one of the means in the PRMS (§ 4.1.2) by
which the PIIP may be calculated using reservoir rock and fluid properties. The term “volumetrics”
refers to a static calculation based on a geologic model dependent upon geometry to describe the
bulk volume of hydrocarbons in the reservoir. The uncertainty of the estimation depends on the
amount of data available and the quality of that data. Uncertainty is high in the early stages of
exploration but decreases as wells are drilled and the reservoir is developed.
A volumetric estimate of in-place volumes is the product of trapped reservoir area, net
reservoir thickness (see Chapter 5—Petrophysics, herein, for a discussion of net pay), porosity,
water or hydrocarbon saturation, and formation volume factor. At prediscovery (exploration) and
early postdiscovery (early appraisal) stages, with limited well information, reservoir (or
prospective reservoir) geometry interpreted from 2D/3D seismic surveys is likely the only direct
way that is available to assess PIIP. Global-average single values of porosity and saturation across
the reservoir, regional mapping, and analogous field data are commonly used in volumetrics
estimation with a range of uncertainties.
Defining trap geometry is essential for volumetrics. For all three types of conventional
reservoir traps (structural, stratigraphic, and combination), trap geometry consists of the top and
base of reservoir boundaries and the top and lateral seal/barriers and constraining or limiting faults,
forming the trap. All of these should be in the physical depth space.
A 3D seismic volume allows an interpreter to map the trap as a 3D grid of seismic amplitudes
reflected from acoustic/elastic impedance boundaries associated with the rocks and fluids in and
around the trap. Mapping traveltimes to selected acoustic/elastic impedance boundaries
(geoscientists often call these boundaries seismic horizons), displaying seismic amplitude
variations along these horizons, delineating isochrons between horizons, noting changes in
amplitude and phase continuity through the volume, and displaying time and/or horizon slices and
volumetric renderings of the seismic data in optimized colors and perspectives all contribute to the
detailed picture of the trap’s geometry.
Such seismic interpretations are made either in the time domain (and then converted to the
depth domain by time-to-depth conversion) or in a depth domain in a depth-migrated data set
(sometimes a depth-to-depth adjustment is needed when more wells are incorporated into or after
seismic depth-processing). Velocity data from wells, optionally supplemented with seismic
velocity data, are used to convert the horizons picked in time into depth and thickness. The
resolution of 3D seismic data typically ranges from 12.5 to 50 m laterally and from 8 to 40 m
28
vertically, depending on the depth and properties of the objective reservoir, as well as the nature
of the seismic survey acquisition parameters and the details of the subsequent processing.
It is important to appreciate that the relative uncertainty in predicting depth to a trapping
surface at a new location some distance from an existing borehole is much less than the errors in
predicting trap depth in an exploration setting prior to the drilling of the first well. That uncertainty
generally is tens to hundreds of meters because there is no borehole control on the vertical velocity
from Earth’s surface down to the trap. Effectively, the early wells help to “anchor” the 3D seismic
volume at several key points in measured depth, assisting the subsurface interpretation in further
refining the accumulation’s geometry into undrilled areas.
Gross Rock Volume (GRV) is defined as the total volume of rock within the hydrocarbon trap,
bounded by top and base of the reservoir, and any lateral barriers, above and below the hydrocarbon-
water contact. The fundamental and traditional role of seismic data in the past has been directed
towards the mapping and construction of a 3D geologic model of hydrocarbon traps and in establishing
the GRV for a reservoir. The seismic-interpreted trap geometry, calibrated by reservoir thickness from
well data, leads to the calculation of GRV, which is a critical factor in the calculation of PIIP.
The same procedure is necessary for each fault block of a fault-controlled, compartmentalized
reservoir and each pool of a stacked reservoir. Also, it is important to assess fault transmissibility,
by analyzing the fault configuration, fault complexity, and the juxtaposition of reservoir-to-
reservoir contact across the fault, and to compare this to a well’s performance, fluid and pressure
data in the surrounding area, along with the regional geologic characterization. The 3D seismic
data directly help to define, classify, and configure the middle- to large-scale faults.
In the case that reservoir formations are interbedded with nonreservoir formations under and
within the trap geometry, reservoir bulk volumes could be obtained by using the GRV and net-to-
gross ratio, after defining the reservoir rock by cutoff values of petrophysical parameters like
porosity, clay content, or a certain log measurement or log interpretation from wells. The net-to-
gross ratio often has a relationship to geological depositional environment, which could be
estimated from 3D seismic surveys, and hence the spatial distribution of net-to-gross ratio could
be estimated and then applied to estimate the reservoir bulk volumes.
Methods and technologies in geophysics advance rapidly. It is especially important to keep in
mind that seismic survey acquisition and the data processing history affect the subsurface imaging.
Recent advances in seismic data processing have improved velocity estimation (including
anisotropy), depth imaging, time-depth conversion, depth-depth conversion of depth-imaged data,
and structural attributes. Detailed seismic velocity analysis can provide a range of uncertainty on
GRV and hence impact ranges in resources volumes.
3.3.2 Uncertainty in GRV. The GRV of a field is defined by structural elements, such as depth
maps and fault planes, resulting from an interpretation based on seismic and well data. If the trap
volume under the seal is completely filled with reservoir rock, then the GRV of the trap is the same
as the reservoir bulk volume. Though ideal, this is generally not the case, and the thickness and
geometry of the one or more reservoir units within the trap have to be estimated to derive reservoir
bulk volume. Uncertainties in the GRV associated with seismic analysis, and hence uncertainties
in the in-place volumes, reserves, and production profiles, can arise from:
• Incorrect positioning of structural elements during the processing of the seismic data.
• Thinly layered reservoir and nonreservoir intervals below well and seismic resolution.
• Incorrect interpretation of trap geometry and reservoir geometry.
• Errors in the time-to-depth conversion.
29
An assessment of these uncertainties is an essential step in a field study for evaluation,
development, or optimization purposes. Generally, the relative uncertainty in predicting formation
depth on seismic data is inversely proportional to the number of wells in the seismic covered area;
i.e., the more wells used to calibrate the model, the less uncertainty there is in the predicted depth.
Lack of well control could lead to large uncertainties in general seismic interpretations, for
example, in a very early stage of exploration.
The accuracy of the estimate of the thickness (gross and net) of each reservoir is a critical
element in assessment of in-place volumes. Estimation of gross reservoir thickness is dependent
on the bandwidth and frequency content of the seismic data and on the seismic velocity of the
formation rocks. Broadband, high-frequency seismic data in a shallow clastic section where
velocity is relatively slow can resolve a much thinner bed than, for example, narrowband, low-
frequency seismic data deep in the subsurface in a fast carbonate section. Fortunately, geoscientists
can analyze seismic and sonic log data to estimate the thicknesses that can reasonably be measured
for particular formations/reservoirs under investigation. Again, the reasonableness of these ranges
can be compared to and constrained with regional data and other oil and gas fields found in analog
plays and basins.
Stacked reservoirs in a trap can be resolved individually, and separate reservoir bulk volumes
can be computed, if the reservoirs and their intervening seals can be interpreted separately and
individually based on the minimum thickness derived from the relevant tuning model. Under these
conditions, a deterministic estimate of reserves in each reservoir is possible. When the individual
reservoirs and seals are too thin to satisfy these conditions, then seismic modeling can be used to
get a general idea of the hydrocarbon reserves that might be present in a gross trapped volume. In
some circumstances, it may be possible to detune the seismic response of thin reservoirs to estimate
the total net or gross reservoir. The reliability of these calculations will depend on the formation’s
geological setting, such as bed thicknesses, spacing between beds, porosity variation, and quality
of seismic data.
3.4 Amplitude Variation With Offset and Direct Hydrocarbon Indicator Analyses
3.4.1 Introduction. The use of seismic reflection data to reduce uncertainties associated with
hydrocarbon assessment has been well documented since the mid-1970s. The term “direct
hydrocarbon indicator,” or DHI, refers to the use of seismic data to indicate the presence of
hydrocarbons. DHIs include flat spots, bright spots (strong amplitude reflections), and dim spots, as
well as more subtle changes in seismic phase (see Fig. 3.1). To maximize the confidence, DHIs must
be interpreted by experts and, in the context of a rock physics model, be calibrated to well data.
Fig. 3.1—Common Direct Hydrocarbon Indicators (DHIs), where twt indicates two-way traveltime.
30
Rutherford and Williams (1989) classified amplitude variation with offset/amplitude variation
with angle (AVO/AVA) anomalies into three classes, based on three reservoir types: Class 1
reservoirs have higher impedance than the surrounding rocks; Class 2 are reservoirs with very
small, either positive or negative, impedance contrast; and Class 3 are reservoirs with lower
impedance than the surrounding rocks. This was later refined by Castagna and Swan (1997), who
introduced a fourth AVO class, which corresponds to low-impedance reservoirs similar to Class 3,
but where the magnitude of reflectivity decreases with offset/angle (Fig. 3.2).
Fig. 3.2—Definition of four classes of amplitude variation with offset/angle (AVO/AVA) (from Castagna and Swan 1997).
When a known hydrocarbon accumulation is being appraised, seismic flat spots and/or seismic
amplitude anomalies can be used to increase confidence in the identification of fluid contacts when
the following conditions are met:
• The flat spot and/or seismic amplitude anomaly is clearly visible in the 3D seismic data
and is not related to imaging issues.
• Within a single fault block, well logs, pressure, well test, and/or performance data
demonstrate a strong tie between the calculated hydrocarbon-water contact (not
necessarily drilled) and the seismic flat spot and/or downdip edge of the seismic anomaly.
• The spatial mapping of the flat spot and/or downdip edge of the amplitude anomaly within
the reservoir fairway fits a structural contour, which usually will be the downdip limit of
the accumulation.
Seismic amplitude anomalies may also be used to support reservoir and fluid continuity across
a faulted reservoir provided the following conditions are met:
• Within the drilled fault block, well logs, pressure, fluid data, and test data demonstrate a
strong tie between the hydrocarbon-bearing reservoir and the seismic anomaly.
• Fault throw is less than reservoir thickness over (part of) the hydrocarbon-bearing section
across the fault, and the fault is not considered to be a major, potentially sealing, fault.
• The seismic flat spot or the seismic anomaly is spatially continuous and at the same depth
across the fault.
If these conditions are met, the presence of hydrocarbon in the adjacent fault block above the
seismic flat spot or seismic amplitude anomaly may be judged to be sufficiently robust to qualify
the hydrocarbon volumes. If these conditions are only partially met, the interpreter must consider
31
the increased level of uncertainty inherent in the data and appropriately classify the volumes based
on the uncertainty components.
Synthetic seismic AVO modeling (also called forward elastic seismic modeling) is the key tool
used to model the expected seismic amplitude response based on available well and/or analog data.
Perturbations to an AVO model are performed for various ranges in reservoir thickness as well as
reservoir lithology, porosity, net-to-gross ratio, overburden, and fluid content (hydrocarbon type,
saturation, and pressure). This exercise provides the seismic interpreter with a fundamental
understanding of possible scenarios that can give rise to various seismic responses, and the results can
be compared to the observed seismic traces. New probabilistic methods allow for the modeling of large
numbers of scenarios that are quantitatively matched to the observed seismic response, thus resulting
in a probability distribution function that ultimately gives a probability of hydrocarbon presence.
The presence of a robust DHI can have a substantial influence on the establishment of the risk
and/or volumes associated with a potential reservoir interval. The recommended practices in the
following sections demonstrate ways to quantify the uncertainty and improve the confidence of a DHI
associated with seismic amplitudes to aid in reducing uncertainty in the estimation of resource volumes.
It is noted that in many other examples, in which the seismic evidence (e.g., seismic DHI) is not
as convincing, other data sources (e.g., pressure data, well performance data, geologic deposition
model) will also contribute as part of an integrated analysis to achieve comparable confidence of the
hydrocarbon volumes below the lowest known hydrocarbons, as observed in the wells.
Common pitfalls in seismic amplitude interpretation from DHIs are generally associated with
variables including: low gas saturation; seismic data artifacts due to multiples, noise, and
anisotropy; unusual lithology (e.g., abnormally high porosity sands, coals, and other unusual
lithologic variations); and formation thickness (i.e., tuning and lateral variations).
3.4.2 Confidence Measures. Practitioners (Roden et al. 2012, 2014), as well as many oil and gas
companies, have compiled databases to quantify the success of using seismic amplitudes to identify
hydrocarbons prior to drilling. Scoring systems vary between companies and consortia, but they are
typically based on two criteria: confidence of the DHI and maturity of the analysis/data quality. Such
scoring systems can be used as a template to help interpreters characterize an undrilled accumulation.
• DHI confidence is established according to the following characteristics:
o Amplitude conformance to structure, which is the amplitude termination at the
closing contour of the structure.
o Presence of a flat spot associated with the hydrocarbon contact.
o AVO response and amplitude standout, matching the expected response from
synthetic seismic modeling.
o Match to local analogs (well control), where amplitude is supported by nearby wells
in similar reservoir zones.
o Quality control of seismic observations with AVO analysis using gathers, far-
offset/angle stacks, and/or windowed attributes.
• The DHI maturity of the available data and its interpretation are based on the following
characteristics:
o Seismic data availability and quality, including 2D/3D, prestack or poststack, and
time or depth data.
o Well-log data availability and quality, including acoustic (compressional and shear
sonic) logs, high-quality density logs, lithology interpretation, and fluid saturation
interpretation.
32
o Seismic data calibration, where seismic data are calibrated to well control to allow
quantitative comparisons of synthetic models to seismic observations.
o Interpretation, where the maturity of the interpreted strata and their depositional
environments are interpreted to allow appropriate amplitude extractions and other
seismic guided analysis, calibrated to well data for both geology and rock physics
perspectives.
Based on the established confidence and maturity characteristics, a DHI score can be
determined and refined over time. This allows for continuous learning and calibration. A consistent
peer-review process should be employed to establish a robust score and remove any bias in the
analysis. As the accumulation is developed and additional data are acquired, DHIs can be updated
and rescored if needed.
3.4.3 Evaluation of Resources Using Seismic and Well Data. The use of seismic data in the
evaluation of resources has been discussed in the literature by Ogilvie and Keyser (2004), Kloosterman
and Pichon (2012), Pichon et al. (2012), and Roden et al. (2012), among others. Seismic data can be
used for assessment of the structure, reservoir thickness/extent, fluid contacts, and, in many cases,
reservoir rock quality (such as low/high porosity and permeability). The presence of DHIs can have a
significant impact on methods used to assess Prospective Resources in the exploration stage, i.e., for
the aspect of area and thickness determination. Using seismic data as a support for reservoir extension
may allow for booking volumes beyond well spacing. An appropriate level of rigor is necessary to
assess Proved Reserves using seismic surveys, and seismic data must be integrated with other direct
indicators of hydrocarbons, but it is important to keep in mind that seismic data alone may not be
sufficient to define fluid contacts for Proved Reserves (PRMS Table 3). Strict criteria are required to
provide a body of evidence and to demonstrate that the seismic survey can establish a deeper contact
than that established by wells (Fig. 3.3). Note, however, that although the presence of hydrocarbons
below the lowest-known hydrocarbon in the figure may be inferred from the seismic signature,
evaluators cannot determine “reasonable certainty” in the absence of other data. A loss of amplitude
suggests the presence of a hydrocarbon-water contact in the cross section, above which may be inferred
hydrocarbons capable of economic producibility. When appropriately analyzed and applied, seismic
DHI can help to reduce the associated uncertainty of resources categories.
Once seismic observations have been calibrated and made to corroborate with DHIs (such as
from cores, logs, and wireline formation testing), DHI workflows should be deemed reliable, as
discussed above.
Fig. 3.3—Example showing where seismic amplitudes may aid in the booking of reserves below the lowest-known
hydrocarbon (LKH) down to the termination of seismic amplitude at the hydrocarbon-water contact (HWC)
(after Ogilvie and Keyser 2004), where HKH indicates highest-known hydrocarbon, GR indicates natural gamma ray log,
and Res indicates resistivity log.
33
3.4.4 Recommended Guidelines for Using Seismic DHIs. Effective DHIs must be derived from
the highest-quality seismic data available and must be shown to have a high degree of confidence
to delineate a Proved area. This can be achieved through confidence measures as discussed
previously. An appropriate level of sophistication and redundancy is necessary to evaluate
reserves/resources using seismic data (Ogilvie and Keyser 2004). The scoring measures should
be documented. If that is not possible, then the following criteria need to be clearly demonstrated
and documented to provide an appropriate “body of evidence.”
• Appropriate seismic data quality available: Angle/offset gathers and angle/offset stacks
must be available for proper analysis. It needs to be demonstrated that high-quality 3D
seismic coverage (depth-migrated volume with accurate velocity control) exists over the
entire structure. The reservoir zone also must be shown to have high signal-to-noise ratio
and appropriate angle coverage to enable appropriate observations of the expected AVO
response. The seismic image must be shown to be free of any noise affecting the ability to
utilize amplitudes to infer the presence of hydrocarbon reservoirs.
• Expected DHI response observed and validated: AVO synthetic modeling with
appropriate perturbations to the rock physics model (porosity, saturation, hydrocarbon type,
thickness) must be shown, using local well control. AVO synthetic models of local
reservoir analogs from nearby wells must be shown to directly match the observed seismic
responses with no ambiguity.
• Amplitude standout from background: The seismic amplitude signature of the
hydrocarbon response must be easily discernible from that of a brine amplitude response
with no ambiguity. For example, typical seismic amplitudes of hydrocarbon-filled sands
with Class 3 AVO (per Section 3.4.1) are on the order of two to three times greater than
the amplitude responses of the equivalent sands filled with brine.
• Amplitude conformance to structure: The downdip limit of the amplitude anomaly
should conform to depth structural contours (Fig. 3.4). Lateral velocity variations should
be taken into account or be shown to have a minimal impact. Time contours may be used
when insufficient formation velocity data are available; i.e., there not enough wells to
calibrate depth.
Fig. 3.4—Grading of amplitude conformance (after Roden et al. 2012).
34
• Continuous seismic amplitude response: Amplitude responses must be shown to be
continuous from the existing well control with productive hydrocarbons to the area of
structural conformance or “flat spot” (Fig. 3.5). No discontinuities should exist in the area
of amplitude to be booked. Discontinuities can indicate variable reservoir quality or
possible seismic data quality issues that affect the reliability of the amplitude interpretation.
• Pressure and fluid data: Seismically inferred fluid contacts must be integrated, and in
agreement, with reliable high-quality pressure or pressure gradient data (e.g., wireline
formation testing, regional mapping, and models of log and capillary pressure based on
heights above free-water level in the reservoir zone being characterized.
• Endorsement: It is imperative that the recommendation to book resources and reserves
using seismic interpretation has been reviewed and endorsed by a technical peer-review
team, internal reserves group (geoscientists and engineers), and/or independent consultants.
• Documentation: Geophysical techniques and applications must be well documented and
show that all the above criteria have been met.
Fig. 3.5—Grading of amplitude continuity (after Roden et al. 2012).
Fundamental analysis by an AVO/DHI expert should be part of the seismic interpretation

because not all anomalous amplitude events are DHIs, and not all geologic settings exhibit
DHIs equally.
After completion of a confidence assessment to evaluate their robustness and their uncertainty,
AVO/DHIs can be qualified and integrated into the evaluation of reserves/resources as key
elements in assessing fluid contacts as well as reservoir extension or compartmentalization.
A grading system was developed by an industry DHI consortium (Roden et al. 2014), based
on AVO characteristics when compared to drilling results. The DHI consortium concluded that the
correct interpretation of DHIs is a critical component in the risk analysis process for exploration
and development wells. Fig. 3.6 shows their statistics for AVO success rates, in terms of proximity
to well control and calibration to AVO modeling. Their findings demonstrate the decline in success
rate as the prospect distance from well control increases, or when the quantity/quality of seismic
input to the AVO model diminishes. The study noted that, “When a well-defined model with
reliable inputs closely matches the seismic, the success rate for [AVO] Class 3 reservoir prospects
is approximately 30% above average, and for [AVO] Class 2 reservoir prospects 40% above the
average success” (Roden et al. 2014, p. SC66) (AVO classes were discussed in Section 3.4.1.).
35
Fig. 3.6—For Class 2 and 3 AVO (see Section 3.4.1), the well success rates based on the proximity to well control (left) and
on performance of AVO modeling (right) (after Roden et al. 2014).
It is usually not possible to distinguish a fully saturated gas accumulation from a partially
saturated column (residual gas) using full-stack or conventional (two-term) AVO analysis, so this
may remain an unresolved risk.
In terms of the AVO/DHI prediction failure cases, the statistical data for unsuccessful
(both geologic and economic) wells, as reported by the above consortium (Roden et al. 2014),
have identified:
• Wet sands—about 49% of failures
• Low-saturation gas—about 23% of failures
• Lack of reservoir presence—about 17% of failures
• Tight reservoir—about 11% of failures
In addition, diagenetic changes that modify rock properties (e.g., facies transformation from
opal-A [amorphous] siliceous shale to opal-CT [α-cristobalite-tridymite] siliceous shale) have
been mistaken for flat spots.
In conclusion, the successful application of quantitative seismic interpretation techniques
requires a systematic and comprehensive methodology to interpret DHIs as well as domain
expertise in AVO interpretation to understand potential pitfalls. Detailed seismic-based reservoir
characterization is made possible by combining a rigorous assessment of the quality of the seismic
data with an understanding of the reliability of a rock physics model that is calibrated to the local
geology and nearby wells. In addition, peer reviews will help to place the AVO interpretation and
quality into the context of similar valid analogs and make appropriate estimates of model
uncertainty and potential resources/reserves categorization.
3.4.5 DHI Examples. 3.4.5.1 Example 1—Oil-Water Contact Evaluation. An example of fluid
contact evaluation using DHI is shown in Fig. 3.7 (Pichon et al. 2012). Well GF-1 was drilled in a
clastic environment and encountered two oil-bearing sand reservoir intervals and a gas-oil contact,
thus only have lowest-known hydrocarbon (oil) or “oil down to” for the two reservoirs. Furthermore,
the seismic survey was tied to the well, and seismic amplitudes within the interval showed excellent
conformance to structure at that level. A key outstanding question after this first well was to
determine the depth of the oil-water contact (OWC) and to define its degree of confidence.
Good quality high-resolution marine seismic data were available. The seismic data were
calibrated at the well, and a time-to-depth conversion was achieved. A detailed structural and
stratigraphic interpretation was required, because the reservoir thickness was less than the seismic
resolution. Synthetic seismic responses of fluid-substituted oil and water sands matched the actual
36
seismic responses very well. Furthermore, AVO calibration from nearby wells gave confidence in
the interpretation of the fluid contact from seismic data.
Fig. 3.7—Fluid contacts definition from seismic cross section and amplitude map (after Pichon et al. 2012), where GOC is
gas-oil contact, and WOC is water-oil contact.
The seismic interpretation of both upper and lower reservoirs indicated that they likely were
connected, and amplitude patterns were continuous. The seismic AVO attributes were conformable
with the depth contours and consistent with the spill limits for the structure. Furthermore, the
GF-1 well confirmed the connectivity of the upper and lower reservoirs through pressure gradient
analysis from a wireline formation test. In conjunction with the pressure data from a nearby,
downdip well wet in this reservoir, an OWC (WOC in Fig. 3.7) honoring the regional aquifer trend
was in agreement with the DHI depth interpreted from the seismic data.
The quality of DHI and consistency between seismic and nonseismic information provided a
high degree of confidence for both the gas-oil contact and OWC depth estimates. The
corresponding hydrocarbon volumes could then be qualified as Contingent Resources with their
associated range of uncertainty reflected in the C1, C2, and C3 categories until all other criteria
for Reserves were met.
3.4.5.2 Example 2—Gas-Water Contact and Compartmentalization Evaluation. Fig. 3.8
(Pichon et al. 2012) is an example of a turbidite reservoir that was interpreted to be a distal lobe
eroded by a channel. Well-2 was drilled within the lobe and discovered a full column of rich gas,
thus establishing the lowest-known gas, as no gas-water contact was encountered.
The seismic data have excellent quality and exhibit a high signal-to-noise ratio. The reservoir
section shows characteristic seismic attributes with a strong amplitude and, after calibration to well
data, good structural conformance to the depth map at top reservoir. Seismic data clearly indicate
the difference between hydrocarbon-bearing and wet sands (i.e., the amplitude attenuation
corresponds to a hydrocarbon-water contact), but rock physics modeling shows that oil having a
high gas-to-oil ratio cannot be discriminated from gas. A range in the hydrocarbon-water contact
37
was estimated by using the hydrocarbon pressure data from the well and aquifer pressures from
nearby wells. The seismic DHI-derived fluid contact depth falls within the depth range that was
estimated independently from pressure trends, the lowest-known gas, and the spill point. It is
concluded that the seismic contact estimation is reasonable and provides a narrower range of
uncertainty than the use of pressure data alone (Fig. 3.8, upper right).
Fig. 3.8—Compartmentalization analysis using a seismic amplitude map, seismic lines, and pressure data. Ranges in
contact uncertainties are shown for different data (upper right) (after Pichon et al. 2012), where HWC is hydrocarbon-water
contact, DHI is direct hydrocarbon indicator, GDT is gas down to, and WUT is water up to.
The main concern in this example is to understand the potential connection between the three
faulted blocks L, S, and G. Three major faults (F1, F2, and F3) are mapped, and minor faults are
also present. The seismic survey shows a continuous high amplitude that is consistent for all three
blocks. F2 appears to die out at line B, but both parts of the S block are connected via a relay zone.
In addition to the analysis of the seismic character across the three blocks, a connectivity
assessment was conducted, including the analysis of sand pressures, the analysis of faults (flexural
versus brittle), and the analysis of fault plane/juxtaposition. The conclusion was that blocks L and
S are very likely connected (high confidence) and that they share the same fluid and hydrocarbon-
water contact, but the connection between blocks S and G cannot be proven from the available
data, and this “disconnection” is shown by the red “X.” A similar disconnect is suspected across a
minor fault within block G. The resources/reserves recognized in the various fault blocks should
be assigned by taking into account the confidence in communication that has been assessed.
3.5 Seismic Inversion

3.5.1 Concepts and Definitions. Seismic inversion, in its many forms, is an industry-standard
technique that is the process of converting seismic traces to a vertical sequence of rock and fluid
38
properties at each trace location (Fig. 3.9). One may think of this like performing a medical
computerized tomography scan of the human body, but the object is changed to Earth’s subsurface:
It is a process to transform the remote-sensing measurements/signals of subsurface Earth into a
series of its physical properties, in a quantitative sense. The narrowest definition of seismic
inversion is the removal of the wavelet of each trace and the derivation of layer acoustic properties
(typically, acoustic impedance). The broadest definition of seismic inversion is the process of
reconstructing Earth properties by combining seismic and well data to estimate reservoir properties,
such as depositional facies, lithology, porosity, net-to-gross ratio, fluid type, and saturation, across
a seismic survey. There are typically relationships between the rock’s elastic properties (acoustic
and shear impedance) and reservoir properties (i.e., lithology, porosity, and pore fill) via localized
rock physics analysis/modeling, and hence, estimates of these parameters could be achieved by
seismic inversion, guided by rock physics analysis.
Fig. 3.9—Simplified schematic description of seismic impedance inversion

(after Society of Exploration Geophysicists Wiki 2022).
Seismic inversion can be done on full-stack data, partially stacked data (near-, mid-, and far-
angle/offset stack), AVO-derived attributes data, and prestack data. In all cases, the quality of the
track record and confidence ranges, either locally within the 3D volume or regionally, will need to
be considered when determining the reliability of seismic-based estimates. Additionally, a
relationship between acoustic impedance or elastic impedance and petrophysical properties must
be established at log-scale resolution. The inversion method should be considered as well as the
confidence in the well-based background model used for generating the low-frequency component.
Once properly calibrated with geological and petrophysical information from local wells,
seismic inversion is able to inform the spatial distribution of reservoirs, to provide accurate
information on stratigraphic variations, and to predict changes in formation properties away from
the wells. This information can be used to populate an SEM.
When geophysicists say impedance, they usually refer to acoustic impedance or P-impedance
(product of P-wave velocity and density), in contrast with shear impedance or S-impedance
(product of S-wave velocity and density). One common means of differentiation is that “acoustic”
means P-wave (compressional wave) only and “elastic” means P-wave (compressional wave) and
S-wave (shear wave) together.
Seismic inversion can be classified in different ways:
• By the type of input data, i.e., poststack or prestack.
39
• By the type of inversion algorithm, i.e., deterministic versus stochastic (probabilistic or
geostatistical).
• By the type of output, with either acoustic/elastic layer properties or reservoir/
petrophysical properties.
Inversion algorithms have inherent nonuniqueness, meaning that there is more than one model
that can give rise to the same seismic data. Because of this uncertainty, a probabilistic approach
can be followed to capture the full range of possible outcomes. The uncertainty analysis should
cover the nonuniqueness of the inversion process and the uncertainties arising from the rock
property model. The probabilities of the various outcomes can then be used as input to reserves
and resource volume assessments. However, estimating all the uncertainties in the process is
difficult. Use of this technology would need to be supported by a strong track record of successful
and unsuccessful cases in similar settings.
It is important to emphasize that the quality of seismic inversion is directly related to the
quality of the input, such as bandwidth, surface and/or near-surface layer influence, signal-to-noise
ratio, and prestack amplitude preservation during data processing. To avoid misleading results, it
is critical to quality control the seismic inversion process (data quality and inversion approach).
3.5.2 Well-to-Seismic Correlation via Synthetic Seismograms. Since well data are the input and
calibration for seismic inversion, a reasonably good correlation between the seismic trace and
synthetic trace generated from well logs should be established prior to performing the inversion
(Fig. 3.10). Usually, a few wells are set aside as blind tests for inversion when sufficient wells
are available.
Fig. 3.10—A well synthetic seismogram correlates with seismic data (van der Weiden et al. 2012), where DTC represents
P-wave velocity, RHOB is density, VSH is shale volume, PHIE is effective porosity, and SWT is total water saturation.
Well scenario modeling (1D and/or 2D) is needed to understand how changes in reservoir
parameters, e.g., reservoir thickness, lithology, porosity, fluid saturation (gas, oil, and water),
influence the seismic response.
40
3.5.3 Rock Physics Modeling/Analysis. Rock physics plays a fundamental role in seismic
quantification. It links elastic properties to reservoir properties (or petrophysical properties) via
core measurement and well-log information. Lithology (or matrix components), clay content,
porosity, pore fluid, pore pressure, and compaction affect a rock’s elastic properties. Modification
in elastic parameters leads to changes in seismic response. Therefore, seismic amplitude carries
information about reservoir properties and makes rock physics a bedrock on which to carry out
derivation and interpretation of reservoir properties from seismic data. With calibration from local
well logs, the combination of rock physics and seismic inversion is a robust method to obtain
seismic-driven reservoir properties and facies. In this process, uncertainty analysis must be
considered through statistical inferences.
To conduct rock physics analysis for quantitative seismic interpretation, it is recommended to
start with theoretical or empirical modeling. Rock matrix and pore-fluid properties must be given
as input to calculate the effective elastic moduli through a theoretical equation. A simple
theoretical effective elastic medium is useful to describe simple grain contacts like unconsolidated
siliciclastic reservoirs. If a reservoir underwent a diagenetic process, cementation models may be
suitable for rock elastic parameters. In carbonate reservoirs, pore type is an important aspect that
leads to more complex models (e.g., the differential effective medium used to calculate the dry
effective bulk and shear moduli for different geophysical pore types represented by the pore
aspect ratio).
With seismic inversion, one can derive the rock’s elastic properties (or geomechanical
properties) such as P-impedance, S-impedance, Vp-to-Vs ratio, Poisson’s ratio, and Young’s
modulus. It is important to understand how elastic properties are related to reservoir or
petrophysical properties. As an example, a quartz-rich turbidite reservoir may have different values
of P-velocity and P-impedance at different parts of the turbidite system. Unconsolidated rocks
have lower values of P-impedance than consolidated rocks, even if they have the same porosity.
Compaction, cementation, and diagenetic processes all change the relation between elastic and
petrophysical properties.
After demonstrating a relationship between rock elastic properties and reservoir properties at
a log scale, the geoscientist should demonstrate that the data quality of the seismic survey at the
reservoir level is good and that, for example, overburden effects do not obscure or distort the
imaging of the reservoir. It is critical for all the rock physics analysis used in seismic inversion to
be conducted under the condition that well synthetics (modeled seismic data derived from density
and sonic logs) adequately tie to the seismic data.
3.5.4 Seismic Impedance Inversion. Seismic impedance inversion is a mature and robust
technique that provides impedance layer properties at every seismic sample across a data set. Wells
are used to derive an initial impedance model. Because inversion accounts for wavelet effects, it
provides a superior definition of stratigraphic layers and an increase in vertical resolution.
A good example of a workflow that uses seismic technology to “generate a high-confidence
case which formed the basis for proved reserves” by integrating seismic and well data was shown
for the Gorgon gas field, offshore Western Australia (van der Weiden et al. 2012, p. 1056). The
referenced article is an example of how the subsurface team characterized the resource quantity
(actually Contingent Resources until all contingencies are met for 1C to become 1P), and the
process was used to document appropriate compliance with US Securities and Exchange reporting.
The workflow steps are as follows:
41
1. Define with reasonable certainty the “reservoir tank” upon integrating seismic, well log,
and wireline pressure data.
a. Establish well-to-seismic ties for each reservoir and define the top and base events on
seismic surveys. Most wells in this example have a high correlation factor for synthetics
with corresponding seismic traces, giving high confidence in seismic-to-well ties.
b. Establish confidence in the quality of seismic data sets, such as prestack time migration
(PSTM), prestack depth migration (PSDM), industry standard P-impedance inversion,
and near-, mid-, and far-angle stack data sets. All seismic data in this example show a
moderate to large fluid effect. The gas sands are soft and show Class 3 AVO response,
i.e., have lower acoustic impedance than the formation above them.
c. Map the reservoirs and their extent away from well penetrations on seismic data.
Through wedge modeling analysis (Fig. 3.11), the tuning thickness was determined
to be 28 m ± 4 m (~15 ms), and the reservoir becomes unmappable below a 5 m
thickness resolution. (The major reservoirs in this field ranged from 20 to 70 m in
thickness.) For each reservoir, up to 15 different thickness map scenarios were
generated. The highest and lowest thicknesses in the range of realizations were taken
as high and low cases, respectively.
Fig. 3.11—2D wedge model to predict the seismic tuning thickness and the detection limit of this seismic survey (van
der Weiden et al. 2012), where TWT is two-way traveltime, and GR is gamma ray.
2. Establish the internal continuity of the reservoirs with reasonable certainty.

a. Identify potentially offsetting faults through seismic interpretation and mapping, to
ensure that the reserves are not assigned to the adjacent reservoir blocks isolated by
major potentially sealing faults. Analyze the fault sealing potential using reservoir
juxtaposition plots and pressure data (Fig. 3.12).
b. Extend the reservoirs away from wells over long distances within the field using
seismic interpretation techniques such as amplitude extraction and geobody extraction.
c. Define seismic analogs from nearby fields with similar geology. (In the authors’
example, there was a positive track record of sand predictability in 40 out of 40
exploration and appraisal wells in a nearby field.)
42
3. Establish the reservoir properties and their continuity with reasonable certainty.
a. Define the indicator that separates the sands from the shales. Here, all the gas reservoirs
are lower-P-impedance sands encased in bounding shales. A separation of 500 P-
impedance units, as indicated by well logs, is a robust indicator of the sand line.
b. Use well-log information to generate models for seismic inversion on stacked data.
Show a good match between the seismic inverted impedance and the well-log
measured impedance to demonstrate that impedance is a reliable predictor of sands
away from well control.
c. Establish the relationship between P-impedance and reservoir properties. The linear
relationship between total porosity and P-impedance defined for the gas sands from
the wells can be applied to predict porosity using the seismic inverted impedance
volume, and to map the continuity and extent of the reservoirs (Fig. 3.13).
Fig. 3.12—Major and minor faults on seismic attributes and traverses (van der Weiden et al. 2012),
where TWT is two-way traveltime.
Fig. 3.13—(a) P-impedance versus porosity relationship for gas sands, where GR is gamma ray. (b) Sands with
P-impedance–derived porosity higher than 7% in the proved areas (van der Weiden et al. 2012), where PHIT is total porosity.
43
The integrated interpretation of the wells and the seismic impedance inversion results were
used to create a base-case reservoir static model. A full-field uncertainty analysis was performed
to identify the key static parameters such as reservoir thickness, time-depth conversion, porosity,
and gas saturation. To generate a high-confidence case that formed the basis for the low-estimate
Contingent Resources (1C), these uncertainties were combined with the expectation case static
model in a probabilistic way. Dynamic simulations of the field were established to demonstrate
economic producibility.
Both reflection and inversion seismic data were used exhaustively to demonstrate the
continuity of reservoirs beyond the radius of investigation seen by the well tests. As a consequence,
a larger resource accumulation was established with fewer wells.
3.5.5 Stochastic Inversion. In recent years, prestack stochastic seismic inversion has gained
considerable attention and application. The seismic inversion process is inherently nonunique,
which means that there are numerous elastic property models that fit the same seismic data with
equal probability. Interpretation geophysicists must deal with this nonuniqueness problem and
discard scenarios that are not realistic.
During stochastic inversion, the large model space is constrained with geological information
and is sampled to identify geologically consistent solutions. Stochastic seismic inversions use a
Bayesian framework to determine the impedance probability density functions from seismic and
well information. Four key steps are usually followed during stochastic seismic inversions:
geological model building, well data upscaling, prior model building, and variogram modeling.
Prestack stochastic inversions generate equally probable, high-frequency layers of elastic
impedances with layer thicknesses that could be as small as 1 ms two-way time (roughly equivalent
to 1 to 1.5 m for midrange formation P-wave velocity). Such fine-scaled layer definition ensures
that results are closer to the desired vertical (well-log) scale for static geomodels.
Using the stochastic inversion impedances and the rock physics model between impedance
and porosity (similar to Fig. 3.13a), large numbers of porosity realizations are generated
mathematically, e.g., through collocated sequential Gaussian simulation. The multiple porosity
solutions allow uncertainties to be accounted for that are evaluated in a suite of numerical models
to obtain a Best Estimate (P50) geomodel.
Consider the crossplot in Fig. 3.14, where elastic properties from well-log data are colored by
interpreted lithofacies. In a first analysis, geophysicists will look for visual facies discrimination
by determining if it is possible to distinguish facies for a range of elastic properties that can be
derived from seismic data. In the example shown here, two elastic properties are key to
discriminating lithofacies: acoustic impedance and the compressional (P-wave) to shear (S-wave)
velocity ratio (Vp/Vs). Geophysicists will then define the probability density functions for each of
the four facies and apply these to the seismic inversion to achieve seismic-derived facies, which
can be further populated into each cell of an existing geomodel by proper scaling.
In summary, seismic inversion’s utility in support of reserve/resource estimation is best suited
for reservoir property mapping (inferring reservoir presence, thickness, saturation, and reservoir
quality). Due to its inherent nonuniqueness, seismic inversion works best where the field properties
are well constrained, e.g., a sand reservoir deposited in an offshore deepwater setting encased in
shales and where good-quality well control is available for calibration. In this instance, linking the
elastic properties measured by the seismic survey to well-derived relationships for sand presence,
quality, and thickness can result in a robust interpretation that can drive the assignment of reservoir
properties in 3D away from well penetrations. Additionally, if the acoustic signature of an oil/gas-
44
bearing reservoir is distinctly different from a water-bearing reservoir, it may be possible to
confidently map a field’s hydrocarbon-water contact to aid in quantifying hydrocarbons below the
lowest-logged elevations. Fig. 3.15 is an illustration of an estimated OWC and associated
indicative oil saturations, as they were derived from inverted data, around an injection well.
Fig. 3.14—Lithofacies classification using elastic properties. (a) Crossplot of acoustic impedance versus Vp/Vs, colored by
facies. (b) Probabilistic approach for seismic classification: for each facies, a probability density function is modelled. (c)
3D view of (b) (after Teixeira et al. 2017).
Fig. 3.15—Illustration of reservoir discrimination, oil-water contact (OWC), and indicative oil saturation from normal
seismic reflectivity (top), inverted Vp/Vs (middle), and prestack inversion for relative oil saturation (bottom) with calibration
along an injection well trajectory and the penetrated OWC (Nasser et al. 2016).
45
3.5.6 Uncertainty Assessment. In assessing volume variance in a discovered resource, seismic
information has often been used to evaluate changing container shapes that could come from
variability in the modeled velocity field used for depth conversion. The volume uncertainty in such
an approach is driven by structural uncertainty. In the past, this GRV uncertainty was often the
main “seismic” uncertainty that was considered in resource variance analysis.
With modern rock physics inversion workflows, it is possible to make reasonable quantitative
estimates in variance between the base-case stratigraphic model and other possible spatial
distributions of the reservoir rocks and their associated properties. In this way, the seismic data
can also inform the uncertainty in reservoir properties away from wells.
When developing relationships by crossplotting elastic properties versus reservoir properties
of interest such as net-to-gross ratio, facies, and porosity, there is often sufficient overlap or
nonuniqueness to allow multiple models to be selected to distribute the properties in the geomodel
in a manner that honors the observations on seismic data.
Using the modeled relationships between elastic properties and reservoir properties, it is
possible to generate multiple realizations for a property distribution that can be ordered on a
property of interest. This approach includes more variance than typical “deterministic” inversions.
In Fig. 3.16, 500 realizations were made and sorted by sand volume. Statistical analysis estimated
P10, P50, and P90 values from the distribution, thus allowing for selections of high, best, and low
cases. A single realization that has a volume at the P50 level is shown for illustration.
Fig. 3.16—Illustration of distribution of 500 sand probability runs and one example P50 proxy case
(Moyen and Doyen 2009).
Quantifying uncertainty is a fundamental step in seismic inversion and seismic classification.

Statistical rock physics methods use statistical techniques to quantify uncertainties in theoretical
and empirical rock physics relations. For the same range of elastic properties, there may be many
46
associated facies. Instead of circling a region in an elastic parameter crossplot and associating it
with a given facies in a deterministic approach, the uncertainty of a seismic classification should
be considered. For each facies, one can describe a probability distribution function as a conditional
probability (Fig. 3.14b). Using Bayes’ theorem, an a posteriori probability distribution function is
constructed and applied sample-by-sample in the seismic cube. As result, a seismic-derived
probabilistic facies cube is obtained, illustrated in cross section in Fig. 3.17. Well facies are not
considered hard data, but seismic probability cubes are consistent with well information used
for calibration.
Fig. 3.17—Seismic section showing reservoir facies probability from a stochastic seismic inversion. Facies displayed on
well path: reservoir (yellow) and nonreservoir (green). Wells A3 and A4 are blind tests (after Teixeira et al. 2017).
A probabilistic seismic cube is a seismic attribute obtained from coupling rock physics
properties and seismic inversion. As an attribute, it is subjected to geophysical interpretation to
achieve its value for seismic reservoir characterization. A posteriori probabilistic information
allows the geoscientist to estimate uncertainty and, hence, to derive several scenarios from seismic
data. For example, a conservative approach would consider the reservoir facies to include all areas
with a probability higher than 80% to 90%. Conversely, an optimistic approach may incorporate
as reservoir facies all regions with probabilities higher than 20% to 30%. This kind of information
can be used in uncertainty analysis to define the PIIP scenarios (low, best, and high estimates).
It is important to emphasize that areas assigned with a high probability of good reservoir/facies
through seismic inversion cannot be considered as discovered based on seismic data alone if they
are disconnected from an area that has well penetrations. (Per the PRMS § 2.1.1.1, “A discovered
petroleum accumulation is determined to exist when one or more exploratory wells have
established through testing, sampling, and/or logging the existence of a significant quantity of
potentially recoverable hydrocarbons and thus have established a known accumulation.”) Instead,
these volumes are considered as undiscovered and should be categorized as Prospective Resources.
The shallowest reservoirs that were penetrated by wells as shown in Fig. 3.17 were wet and were
not considered in PIIP. Similarly, the deeper reservoirs that were not drilled by any of the wells
cannot be included as discovered PIIP until they are drilled.
Well-test interpretation provides key data that are used to characterize how fluid-flow
properties may change around a well. These changes can be associated with modifications in
47
petrophysical properties that are connected to elastic properties through rock physics. In such
cases, seismic attribute interpretation can be a very useful tool to support dynamic data
interpretation and thereby mitigate risks and reduce uncertainty due to ambiguity of seismic
interpretation (Fig. 3.18).
Fig. 3.18—Interpretation of static data integrated with dynamic data. Seismic information from facies (left) matches well
testing interpretation (right). This consistency helps to mitigate risk and reduce uncertainty (after Teixeira et al. 2017).
As shown in the example (Fig. 3.18), a best estimate reservoir probability map was extracted
from the probability reservoir facies cube obtained by the statistical rock physics process. It should
be noted that in the region of well A, and in the region of proposed well B, there is a high
probability for the occurrence of reservoir facies. However, the seismic inversion indicates that the
probability of finding reservoir facies drops off quickly away from well A. This suggests we may
find nonreservoir facies at a short distance from well A. Similarly, the interpretation of a well test
from well A shows that there is a significant permeability reduction (possibly a flow barrier) away
from the well. When data that are combined from different independent sources (such as static and
dynamic data) lead to the same understanding of the reservoir, the integrated interpretation
becomes more robust, and confidence in the interpretation increases substantially.
3.6 Seismic Surveillance

A third general application of 3D seismic analysis is monitoring changes in reservoir properties
(fluid saturation, pressure, temperature, porosity) that occur due to production. This application
highlights seismic differences at various stages of production and is often called time-lapse or
4D seismic.
Seismic surveillance is possible when one:
• Acquires a baseline seismic data set, preferably before the start of production or before a
change in field depletion strategy (e.g., water injection).
• Acquires additional 3D seismic data sets (called monitor surveys) after the baseline.
48
• Observes amplitude and traveltime differences between the baseline and monitor surveys
that can be related to reservoir property changes through rock physics analysis and
seismic modeling.
Time-lapse seismic data influence resource estimation and classification in several ways by:
• Illuminating previously unidentified reservoir areas, impacting estimates of net reservoir
volume.
• Identifying swept and/or unswept hydrocarbon, resulting in resource reclassification.
• Improving time-depth relationships, which are then used to update GRV.
• Improving static and dynamic reservoir models used to estimate recoverable
hydrocarbons.
• Identifying depleted and/or undepleted reservoirs, resulting in resource reclassification.
• Establishing whether different fields are in pressure communication and if undeveloped
resources are being accessed by existing wells.
The reliability of 4D seismic data is determined by:
• The similarity of the baseline and monitor seismic images outside of the produced
reservoir area. This is called repeatability and is a measure of the signal-to-noise ratio of
the 4D seismic data. More repeatable data are more reliable.
• The accuracy of the rock physics model, derived from a relevant set of well-log and core
data, which links the physical changes related to the hydrocarbon recovery process to the
changes in the rock’s elastic properties. The interpretation of larger elastic property
changes is generally more reliable than interpretations for smaller changes.
• The number of monitor surveys. Multiple repeat surveys reduce interpretation uncertainty
by separating signal from noise, allowing fluid fronts to be tracked with more confidence.
Repeatability, and thus reliability, of 4D seismic data is controlled by the similarity in baseline
and monitor survey geometry and processing. Changes in the seismic image that result from
acquisition and processing differences may be mistaken for production-related time-lapse changes.
Repeatability can be quantified using multiple measures, most notably the normalized root mean
square difference calculated for each trace between two seismic surveys (Kragh and Christie 2002).
Normalized root mean square difference measures the percent nonrepeatability of the seismic
surveys. Away from facilities such as platforms, modern marine 4D seismic surveys can be highly
repeatable and thus very reliable. Onshore 4D surveys are generally less repeatable mainly because
of time-dependent variations in the near surface.
The ability to relate the changes in the seismic response that result from production to the
changes in fluid saturation, pressure, and other rock properties depends on the accuracy of the rock
physics model, which is typically derived from well-log data and supplemented by core data. There
can be uncertainties in the prediction of the 4D seismic response, especially for changes in
reservoir pressure. However, with proper calibration and sufficient understanding of the reservoir’s
depletion processes, rock physics models can be used for 4D seismic modeling to aid interpretation
and for updates of reservoir geologic and flow simulation models. If the change in seismic response
that occurs due to production is greater than the nonrepeatability of the data, then the 4D signal is
considered to be robust.
There are many examples of 4D seismic signals, including:
• Water displacement of hydrocarbons as a result of aquifer influx and/or water injection
is common in many published case studies (such as Example A below).
• Amplitude and traveltime changes can often delineate original and producing fluid
contacts.
49
• Gas injection and/or gas cap expansion also result in changes in the seismic response of
the reservoir.
• Demonstrating the movement of water or gas in the reservoir can highlight zones of
greater reservoir quality and permeability pathways and can delineate fluid-flow barriers.
• Pressure increases, particularly associated with injection wells, can also result in 4D
seismic changes.
• If reservoir pressure declines below the bubble point, liberated gas can sufficiently change
the seismic response, illuminating a reservoir that may not have been seismically visible
on the baseline data.
• As reservoir pressure declines further, the reservoir may compact, reducing its thickness
and porosity. As shown in Example B, compaction is often inferred from 4D seismic
changes in the overburden and underburden as they geomechanically respond to the
change in reservoir thickness.
In general, the seismic surveillance tool is useful in time-lapse mode as a check on the validity
of the assumptions in the geologic model that is used in a reservoir simulation of fluid flow.
Because seismic monitoring is more spatially specific than pressure monitoring, estimation and
extraction of reserves can be optimized over time by using the seismic data to guide detailed
simulations of depletion and to resolve contradictions between the seismic survey and the reservoir
model. In general, the incorporation of time-lapse seismic results prompts geologic model updates
that usually improve production history matches.
3.6.1 Seismic Surveillance Example A: 4D Interpretation and 3D Inversion Lead to

Reassessment of the Resources and Reserves. In this first example (Johnston and Laugier 2012),
the key uncertainties in resource estimation for the Paleocene reservoir in the North Sea Ringhorne
Field were the effective reservoir thickness and the area above the inferred original oil-water
contact (OOWC), determined by logs and extrapolated using seismic data. Much of the sand
thickness was below seismic resolution. An added complexity was that the reservoir was nearly
transparent on the preproduction baseline stacked seismic data. Reservoir interpretation was
complicated by seismic interference from an underlying strong top chalk reflection. The nearby
flat areas of the reservoir in the north and south of the field were particularly sensitive to depth
conversion, adding to uncertainty in volumetrics. Water sweep from a regional aquifer resulted in
a robust 4D (difference) seismic signal, which can be observed to encroach on the structure as a
result of production from wells C07, C09, and C10 (Fig. 3.19).
The extent of the 4D swept volume outside the originally interpreted OOWC is a clear
indication of reservoir volumes that were not included in the initial assessment. The new (true)
OOWC, defined by the downdip limit of the 4D geobody, provided a strong constraint on seismic
velocity and time-depth conversion, particularly in the northern part of the field, away from well
control. The 4D data provided the justification to lift and expand the flank of the reservoir.
Evidence from the 4D data also suggested increased sand thickness in certain areas. These
interpretations of swept oil were validated by matching the swept volume obtained by the 4D
seismic data with the produced volume through material balance calculations.
In addition to the differencing of the 4D seismic stacked data, prestack AVO elastic inversion
of the baseline 3D seismic survey was used to better characterize the reservoir and extend its
pinchout limit further updip. Based on the 4D interpretation and the 3D inversion, a reassessment
of resources led to a 40% increase in the resource base for the reservoir.
50
Fig. 3.19—Water sweep geobody (dark blue), extracted from the 4D amplitude difference, displayed on the base reservoir
time surface. Also shown are the pre-4D interpreted original oil-water contact (OOWC) in seismic two-way time (light blue)
and updip pinchout (red). The revised pinchout interpretation, based on elastic inversion of the baseline 3D survey, is
shown in purple. Note that the 4D response extends outside the initial OOWC interpretation. The time-depth relationship
for the reservoir was adjusted so that the OOWC was conformable to the downdip limit of the 4D signal
(modified from Johnston and Laugier 2012).
Of that increase, 60% was applied to Proved and Probable Reserves (P1 and P2) based on the
increased in-place volumes (from the combination of lifting the reservoir above the OOWC in
certain areas, thicker reservoir in some places, and a shallower updip limit of the trap), continued
strong reservoir performance, and planned infill wells. (The development project was based on a
final investment decision and plan for the infill wells.) A second 4D monitor survey confirmed the
locations of two planned infill wells. Probable Reserves were assigned where data control or
interpretations of available data were less certain than in the Proved Reserves areas, and they were
estimated as an incremental quantity as opposed to a cumulative (2P) assessment.
The remaining 40% of the increase in the resource base was applied to the Contingent
Resources class based on additional infill drilling potential and the possible development of attic
oil and/or thin reservoir sands. No incremental Possible Reserves (P3) were assigned due to the
fact that the geoscience and engineering data, although able to define the area and vertical reservoir
limits from which production should be economic, did not cover an area with a defined,
commercially mature project. The area updip of the existing project did not have a development
plan in place at the time of the assessment and was classified as Contingent Resources.
51
3.6.2 Seismic Surveillance Example B: Evaluate the Stress Change in the Overburden. In this
second example, pressure depletion of the high-pressure, high-temperature (HPHT) North Sea
Shearwater Field, in combination with the high compressibility of the reservoir rock, resulted in
compaction of the reservoir sandstone and led to displacements, deformations, and stress changes
in the overburden rock due to stress arching (De Gennaro et al. 2017). The stress changes in the
overburden resulted in significant wellbore integrity issues, which led to a decision to shut in
damaged wells and postpone an infill drilling program. In order to reinstate drilling, the operator
began a study to better understand the overburden based on the integrated evaluation of all
available subsurface and drilling data, including information obtained from geochemical analyses,
geomechanical modeling, and 4D seismic surveys.
The 4D seismic data recorded compaction within the reservoir and stress changes outside of
the reservoir. Shifts in seismic arrival time between baseline and monitor surveys, which could be
robustly measured to a high degree of precision, were most diagnostic of these changes.
Overburden and underburden reflections in the monitor survey were delayed in time relative to the
baseline because of extension and a decrease in the vertical stress, which resulted in a decrease in
seismic velocity (Fig. 3.20a). The derivative of the time shifts, called the time strain, indicated
zones with the greatest changes in velocity (Fig. 3.20b). The time strain also showed an increase
in velocity within the reservoir due to compaction.
Fig. 3.20—(a) 4D time-lapse shifts between the baseline and monitor surveys showing the overall slowdown of seismic
velocities between the surveys. (b) 4D time-lapse strains between the baseline and monitor surveys indicating the
locations where the largest rate-of-change within the time shifts occurred. Red colors indicate a deceleration of seismic
velocities (e.g., in the overburden), whereas blue colors indicate an acceleration of seismic velocities (e.g., related to
compaction in the reservoir), after De Gennaro et al. (2017).
Analysis of the 4D data suggested that the overburden was not as homogeneous as previously
thought and that localized zones existed where enhanced permeability and flow potential through
52
conductive fractures was possible. In addition, comparisons of the 4D seismic results to predictions
based on geomechanical modeling were utilized to update and improve the geomodel, which was
then used to more confidently estimate volumetrics and to design the reinstatement wells.
3.6.3 Seismic Surveillance Example C: Evaluate Fault transmissibility and Reservoir

Compartmentalization. One of the key applications of 4D analysis is in the observation of
barriers and/or baffles. Because 4D data record changes through time due to production, it is a
dynamic surveillance tool. For example, the analysis of the time shifts at the top of the reservoir
at the North Sea Elgin Field (Fig. 3.21) indicated that some faults could act as pressure barriers,
compartmentalizing the reservoir, while other faults allowed communication between producing
and undrilled areas (Taylor et al. 2007). In Fig. 3.21, the fault block to the southeast showed very
little 4D response, suggesting that it was isolated from the producing wells and was undrained.
The east-west fault that formed the northern limit of that fault block was likely a barrier.
Conversely, the area to the north of the field, which was located across a bounding fault, showed
a significant 4D response and was therefore likely to be in pressure communication with the
producing wells. One should keep in mind, however, that this area was located near the platform,
and thus the seismic data would have poorer repeatability and would be less reliable compared to
elsewhere in the field.
Fig. 3.21—Measured 4D time shifts above a compacting reservoir. Reflections in the monitor survey are delayed in time
relative to the baseline survey as a result of extension in the overburden due to compaction of the reservoir (modified from
Taylor et al. 2007), where TWT is two-way traveltime.
A coupled geomechanical and reservoir engineering model of the field was created and used
to generate synthetic 4D time shifts. The model was calibrated by matching the synthetic time
53
shifts to the observed time shifts. The improved model helped to determine the actual stresses
within and around the reservoir and to better predict the field behavior. The ultimate goal of the
model update was to define new drilling targets and mitigate casing integrity risk. These
observations and hence updates of overburden and reservoir modeling, along with other data,
resulted in a decision to resume infill drilling in the field.
3.6.4 Seismic Surveillance Example D: Monitor Water and Gas Sweep Efficiency. In this
fourth example, an Angolan deepwater field consisted of unconsolidated turbiditic sand systems
(Berthet et al. 2015). Understanding heterogeneities within these systems is essential to locate infill
wells. Oil was produced using water and gas injection to support the reservoir pressure. The
primary purpose of the 4D seismic surveys was to monitor the water and gas sweep efficiency.
However, because the reservoir was initially undersaturated, depletion drove the pressure below
the bubble point. This resulted in free gas in the reservoir that was detectible on the 4D monitor
surveys, delineating areas of the field that were in pressure communication with the depleted areas.
Fig. 3.22a shows the 4D response related to the gas effects, i.e., a decrease in velocity within the
reservoir. The spatial limits of the 4D signal could be interpreted to define reservoir baffles (faults
and sedimentological limits such as a channel abandonment or erosive surface), which were
incorporated into the reservoir model. Transmissibility multipliers were then optimized in a
seismic history match in the dynamic simulation model to minimize the difference between the
initial model 4D velocity change (Fig. 3.22b) and the observed velocity change (Fig. 3.22c).
Fig. 3.22—(a) 4D seismic response showing a decrease in velocity between the baseline and monitor seismic surveys as a
result of gas exsolution in the reservoir. (b) Simulated 4D response based on the initial dynamic reservoir simulation
model. (c) Simulated 4D response based on the history-matched simulation model. Green bars indicate perforated
intervals (modified from Berthet et al. 2015).
Analysis of the 4D response data indicated that the water injector did not support the oil
producer because of a sedimentological baffle. Pressure support was noted to the north through
several sources, as noted in the paper. The seismic interpretation provided a good pressure match
and improved the match to the water-cut. The 4D data showed an area where gas had not come out
of solution (no color change in Fig. 3.22a), contrary to the initial model prediction. Thus, there
was potential for an infill well (either a producer or an injector) in this area (subject to further
simulation scenarios).
3.7 Conclusion
Seismic methods have been used in petroleum exploration and production for decades and have
achieved significant success, especially in finding new fields and better characterizing old fields,
54
credited to their spatial coverage and continuous and even sampling of the 3D subsurface. The
application of seismic data in petroleum resources management has a wide spectrum, from an
intensive usage of the high-quality 3D and time-lapse (4D) seismic data for all the classes of
reserves and resources in the classification framework for an offshore oil/gas field to sparse seismic
usage in 2D applications in older fields. Eventually, any seismic-driven new discovery in the
exploration stage is a case of turning Prospective Resources to Contingent Resources.
An SEM is the place where all the disciplines work together. The application of seismic
information, after calibration with well data, is to provide a 3D static model of the structure and
stratigraphy, including the reservoir geometry and faults, leading to the estimation of GRV in a
reservoir and then PIIP. The uncertainties of these estimates can be reduced as measurements from
more wells are incorporated. In certain cases, seismic data can show the OWC and gas-water
contact by AVO and/or DHI techniques. The formation’s acoustic/elastic properties and reservoir
properties can be estimated via seismic inversion guided by well data and then used in 3D static
Earth model building for reservoirs. The 4D seismic data can be used to monitor the changes in
reservoirs due to production, and then to indicate the sweep efficiency and identify the unswept
areas, which contribute to building the dynamic Earth model, understand the depletion sequence,
and optimize the field development plan.
The examples presented in this chapter intend to show illustrative scenarios of seismic
application in resources management. Since each 3D seismic survey can be different in acquisition
time, equipment, ground conditions (on land), near-surface conditions, and data processing, the
quality of seismic data can vary significantly. It is important for reserves evaluators to consult with
their colleague geophysicists to utilize the seismic data properly and effectively for reserves
assessment purposes, and to discuss questions on geophysical theories and concepts. The goal is
to adequately maximize the value of seismic data in the entire chain of petroleum resources
management while characterizing the controlled quality and range of uncertainty.
3.8 Acknowledgments
The authors thank the Society of Exploration Geophysicists (SEG) Board of Directors for their
encouragement and support, and the SEG Oil and Gas Reserves Committee for taking the lead as
a group. The authors wish to acknowledge Jean-Marc Rodriguez, author of this chapter in the
previous version of the Guidelines for Application of the PRMS. The authors also thank Rob
Stewart, a former SEG president, for insightful discussions, Maria Capello as the SEG board
liaison, and Annabella Betancourt as the SEG staff liaison.
3.9 References
Berthet, P., Solans, P., Selva, F. et al. 2015. Production History Match of an Angolan Deepwater
Field Turbiditic Channel Using 4D Seismic Information. Paper presented at 77th
Conference and Exhibition, EAGE, Madrid, Spain, 1–4 June 2015. Th N101 04.
Castagna, J. P., and Swan, H. W. 1997. Principles of AVO Crossplotting. The Leading Edge 16
(4): 337–342. https://doi.org/10.1190/1.1437626.
De Gennaro, S., Taylor B., Bevaart M et al. 2017. A Comprehensive 3D Geomechanical Model
Used to Deliver Safe HPHT Wells in the Challenging Shearwater Field. Paper presented at
51st US Rock Mechanics/Geomechanics Symposium, San Francisco, California, USA,
June 2017. ARMA 17-735.
Johnston, D. H. and Laugier, B. P. 2012. Resource Assessment Based on 4D Seismic and
Inversion at Ringhorne Field, Norwegian North Sea. The Leading Edge 31 (9): 1042–1048.
55
Kloosterman, H. J. and Pichon, P.-L. 2012. The Role of Geophysics in Petroleum Resources
Estimation and Classification—New Industry Guidance and Best Practices. The Leading
Edge 31 (9): 1034–1040.
Kragh, E. and Christie, P. 2002 Seismic Repeatability, Normalized RMS, and Predictability. The
Leading Edge 21 (7): 640–647. http://dx.doi.org/10.1190/1.1497316.
Moyen, R. and Doyen, P. M. 2009. Reservoir Connectivity Uncertainty from Stochastic Seismic
Inversion. Paper presented at 2009 SEG Annual Meeting, Houston, Texas, October 2009.
SEG-2009-2378.
Nasser, M., Maguire. D., Hansen. H. J. et al. 2016. Prestack 3D and 4D Seismic Inversion for
Reservoir Static and Dynamic Properties. The Leading Edge 35 (5): 415–422.
Ogilvie, J. S. and Keyser, W. L. 2004. Seismic Considerations for Classifying Proved
Resources/Reserves: West Africa Example. Paper presented at 2004 SEG Annual Meeting,
Denver, Colorado, October 2004. SEG-2004-0478. .
Pichon, P.-L., Delahaye. S., Fabre. G. et al. 2012. DHI Support for Resources Evaluation:
Confidence Assessment Examples. The Leading Edge 31 (9): 1060–1065.
Roden, R., Forrest, M., and Holeywell, R. 2012. Relating Seismic Interpretation to
Reserve/Resource Calculations: Insights from a DHI Consortium. The Leading Edge 31
(9): 1066–1074.
Roden, R., Forrest, M., Holeywell, R. et al. 2014. The Role of AVO in Prospect Risk
Assessment. Interpretation 2 (2): SC61–SC76.
Rutherford, S. E. and Williams, R. H. 1989. Amplitude versus Offset Variation in Gas Sands.
Geophysics 54: 680–688.
Society of Exploration Geophysicists. 2022. SEG Wiki, https://wiki.seg.org/wiki/Main_Page.
Taylor, N., Ben-Brahim, L., Tindle, C. et al. 2007. Elgin-Franklin 4D Seismic—Encouraging
Results for Reservoir Monitoring & Development Planning. Paper presented at 69th EAGE
Conference and Exhibition, June 2007. A026.
Teixeira, L., Cruz, N., Silvany, P. et al. 2017. Quantitative Seismic Interpretation Integrated with
Well-Test Analysis in Turbidite and Presalt Reservoirs. The Leading Edge 36 (11): 931–
937. https://doi.org/10.1190/tle36110931.1.
van der Weiden, R., Nayak, P., and Swinburn, P. 2012. Seismic Technology Supporting
Reserves Determinations: Gorgon Field, Australia. The Leading Edge 31 (9): 1050–1058.
56
Chapter 4
Assessment of Petroleum Resources

using Deterministic Procedures
Danilo Bandiziol (Chair)
Dominique Salacz, Jes Christensen, Joel Turnbull, and Oluyemisi Jeje
4.1 Introduction
This chapter provides additional guidance to the Petroleum Resources Management System
(PRMS) § 4.1 and § 4.2 regarding the application of the three broad categories of deterministic
analytical procedures for estimating the range of recoverable quantities of oil and gas. These
categories are: analogy, volumetric analysis, and production performance analysis methods, which
include material balance, reservoir simulation, decline curve analysis, and other techniques.
In the early days of a project’s life when the project may not have advanced much beyond the
development plan, modeling generally will be simple (such as through the use of analogs or
volumetric analysis) and then gradually increase in complexity as the project matures. Depending
on the scale of the investment, it is not uncommon to have advanced models available (even full-
scale reservoir simulations) and ready at sanction. However, it is more common to see a
progression of these models throughout the life of the project as the evaluating entity gathers more
data and information about the reservoirs being produced. Once the project is fully mature, and
only little further investment (financial and/or data acquisition) is anticipated, evaluators may use
simpler models such as decline analysis, which rely mainly on the production performance of
existing wells.
If the project is incremental (PRMS § 2.3.0.1) to an already existing project (with an existing
resource recovery estimate already attached, often referred to as the base profile), it is important
to evaluate the resources for the addition and the existing project together (to take into account
effects such as interference amongst wells or back-pressure effect in the production facility) and
to compare the resulting profile with the base profile. Sometimes an incremental project will
accelerate the production of resources from the base profile, but the acceleration may be
accompanied by a deceleration (relative to the base profile) at later time. This is generally a
positive situation because it improves the economics, but only the added resources (difference
between the total profile and the base profile) of the new project should be counted in the PRMS
matrix when booking reserves and resources. Care must also be taken if incremental projects are
booked as additive to a base profile/project because these projects may not be independent.
Simpler models used for estimation of resources during the early phases of a project (such as
exploration, appraisal, and sometimes initial development periods) can be derived by estimating
initially in-place volumes using static-data-based volumetric methods and the associated recovery
efficiency derived from analog development projects, or by using analytical methods. As these
models generally are not very detailed, their execution may require less computer power and will
therefore be a relatively quick exercise. It is common to incorporate these early efforts into multiple
models or scenarios, whether deterministic or probabilistic (please see Chapter 7—Probabilistic
Resources Estimation for further details).
57
Once on production, resources (now likely to be classified as reserves) may be estimated using
dynamic-data-based performance analysis (such as reservoir simulation or advanced material
balance models).
The PRMS accepts two deterministic approaches to resource/reserves estimation: the
“incremental” and the “scenario” (or “cumulative”) methods (PRMS § 4.2.1.1). When both are
applied to the same project, they should arrive at comparable results (PRMS § 2.2.2.11), especially
when aggregated to the field level; they are simply different ways of thinking about the
same problem.
In the incremental approach, more emphasis is placed on the experience and professional
judgment of the evaluator in the estimation of discrete quantities to be assigned to each category
(e.g., P1, P2, and P3 for incremental reserves). When performing volumetric analyses using the
incremental approach, a single value is adopted for each parameter based on a well-defined
description of the reservoir to determine the in-place resources, or reserves, volumes. This method
works best when there is insignificant interference among the low, best, and high estimates.
In the scenario/cumulative approach, three separate analyses are prepared to bracket the
uncertainty through sensitivity analysis (i.e., values estimated by three plausible sets of key input
geoscience and engineering data parameters). These scenarios are designed to represent the low,
best, and high realizations of original in-place volumes and associated recoverable petroleum
quantities (PRMS § 2.2.1.4). Depending on the stage of commercial maturity, these scenarios
underpin the PRMS categorization of Reserves (low, best, high or 1P, 2P, 3P) and Contingent
Resources (low, best, high or 1C, 2C, 3C) of the projects applied to discovered petroleum
accumulations or Prospective Resources (low, best, and high or 1U, 2U, 3U) of the undiscovered
accumulations with petroleum potential.
The advantages of a deterministic approach are the following: It describes a specific case
where physically inconsistent combinations of parameter values can be identified and removed
(i.e., the specific combination of static and dynamic parameters must physically exist); it is easier
to prepare; it is easier to audit; and the approach has a history of use with estimates that generally
can be reliable and reproducible.
The guidance in this chapter is focused on the deterministic methods where the range of
uncertainty is captured primarily using a scenario approach. The reader is referred to Chapter 7—
Probabilistic Resources Estimation, herein, for guidance on applying probabilistic methods.
4.2 Project Life Cycle

Fig. 4.1 illustrates the range of uncertainty of estimated ultimate recovery (EUR) for any petroleum
project, which is expected to decrease over time as the accumulation is discovered, appraised (or
delineated), developed, and produced, with the degree of uncertainty normally decreasing at each
stage. Once discovered, the length of each period depends on many factors, such as the size of the
accumulation (which affects the duration of the appraisal and construction periods) and the
development design capacity in terms of annual reservoir depletion rate. For example, projects
with lower depletion rates will support a relatively longer plateau period followed by a longer
decline period, and the opposite is also true. While the “best estimate” is conceptually illustrated
as remaining constant, there may be significant volatility in this estimate in actual projects over
the course of the field appraisal and development life cycle.
58
Exploration and Appraisal Phase I
Production Phase II
High Estimate
Range of Uncertainty
Estimated Ultimate
Best Estimate
Ultimate Recovery
Recovery (EUR)
Abandonment
Low Estimate (UR)
Cumulative
Production
Plateau Period
Daily Oil Rate
Early Late
Time Time
IIa IIb IIc (Decline Period)
I
Economic Limit
Time
Analogy and Analytical Methods
Volumetric
Methods Reservoir Simulation and Material Balance Methods
Production Performance Trend (PPT) Analysis
Fig. 4.1—Change in uncertainty and examples of assessment methods over the project’s exploration and production
life cycle.
The selection of the appropriate method with which to estimate reserves and resources
changes for any petroleum recovery project over its life cycle, and the accuracy of estimates
depend largely on the following factors:
• The type, quantity, and quality of geoscience, engineering, and economics data available
and required for both technical and commercial analyses.
• Reservoir-specific geologic complexity, hydrocarbon type, recovery mechanism, stage of
development, and maturity or degree of depletion. (For unconventional reservoirs, the
reader is referred to Chapter 10—Unconventional Resources Estimation herein.)
• Partnership agreements—Many petroleum accumulations are developed by joint
ventures, with one company conducting operations. While generally the operator will
have a preferred modeling type/approach, and the other partners either may adapt to the
operator or supplement provided estimates with their own independent work, it is good
practice to have multiple model types to increase the reliability of the estimate. Other
types of agreements may only address a shared cost with the operator, and in those cases,
all evaluations are done independently of the other interest holders in the project.
• The anticipated size of investment (higher investment normally requires more initial work
to reduce uncertainty, e.g., by applying more complex models).
Although certain estimation techniques are better suited to a particular stage in the project as
per the application ranges indicated by the arrows at the bottom of Fig. 4.1, the reader should not
consider the indicated limits as strict. The darker areas of the bars in Fig. 4.2 generally conform to
the stages at which each technique is most applicable.
59
Pre-Drill Post-Drill
Prospect Explore Appraise Delineate Produce
Typical Resource Classification
Prospective Contingent Resources Reserves
Resources
Analogy
Volumetrics
Material Balance
Simulation
Decline Curve Analysis
Fig. 4.2—Change in assessment methods over the project’s exploration and production life cycle.
For example, simulation is most applicable in a project where sufficient data have been
acquired to enable a history match of performance. However, a near-field prospect might already
be covered by an existing geocellular model and have sufficient offset wells to provide meaningful
input to the subject model. Likewise, a material balance approach is helpful in the initial stages to
test scenarios, and it is also useful with postproduction data.
Depending on the amount and quality of historical pressure, production, and other reservoir
performance data available, a combination of reservoir simulation, material balance, and
production performance trend (PPT) analyses (such as decline curve analysis) can be used not only
to estimate directly the recoverable petroleum, but also to estimate the petroleum in-place
quantities (by the first two methods only) and thereby provide a useful second check and validation
of estimates obtained earlier by volumetric methods.
4.3 Analogy-Based Assessment

According to Hodgin and Harrell (2006, p. 2), an analog reservoir is one that is “at a more advanced
stage of development than the reservoir of interest and thus may provide concepts or patterns to
assist in the interpretation of more limited data.” The PRMS added further emphasis when
evaluating analogs to incorporate the development plan elements, which are often a material
consideration in the resultant applicability of the proposed analogous reservoir(s).
Although there are several suggestions in the literature as to how an analog should compare
with the reservoir in question, the suitability of the analog fundamentally depends on the purpose
of comparison. The PRMS allows some flexibility in the choice of the analog, but certain
fundamentals must be considered.
The evaluator should start by considering four aspects: the depositional environment, the fluid
type, the reservoir drive mechanism, and the development type. Should any of these factors be
materially different between the subject reservoir and the proposed analog, it is highly likely that
the analogy will be poor. Once these four aspects have been considered, the remainder of the
aspects (PRMS § 4.1.1.2) can be reviewed in detail.
Subsequently, the petrophysical properties (termed as the “familiar six” by Sidle and Lee
2010) should be considered explicitly: porosity, permeability, permeability distribution, net
thickness, continuity, and hydrocarbon saturation.
60
In certain resource classifications, additional mandatory requirements for the analog are
stipulated, as described below.
Note that, apart from certain exceptions pointed out below, the analog reservoir does not need
to be in the same geographic area as the subject reservoir, nor should geographic proximity be the
foremost consideration in assigning analogy (PRMS § 4.1.1.5).
4.3.1 Purpose of Analogs. Fundamentally, there are two reasons why analogs are used in resource
management: classification of resources and estimation of resources. The former typically has
more stringent expectations of the analog than the latter.
4.3.1.1 Classification of Resources. The PRMS system explicitly references four situations
where analogs may be used to justify classification in a given resource class:
• Economic producibility without a flow test (PRMS Table 1: Reserves): In this situation,
the analogous reservoir should be in the same area.
• Applicability of workover, treatment, and changes of equipment (PRMS § 2.3.1.1):
Incremental recovery from such treatments may be classified dependent on analogous
reservoirs.
• Improved recovery (PRMS § 2.3.4.3): Commerciality may be justified based on a
reservoir with analogous rock and fluid properties where a similar improved recovery
project has been successfully applied.
• Unconventional Contingent Resources without a flow test (PRMS § 2.4.0.4): In this
situation, log and core data and nearby producing analogs provide evidence of potential
economic viability.
4.3.1.2 Estimation of Resources. The extent to which analogs are used in the estimation of
resources will depend largely on the maturity of the project in question. For Prospective Resources
or predevelopment Contingent Resources, there will be limited information on the reservoir in
question, so some fundamental questions will need to be answered with the help of analogs,
perhaps in combination with other analytical methods, including:
• Recovery efficiencies (primary/secondary/tertiary)
• Production rates and profiles
• EUR/well
• Reserves life index (defined as reserves divided by average or current production rate)
As the project becomes more mature, it is likely that alternative analytical techniques will be
used, but analogs still may be necessary to fill in the blanks, at least prior to establishing
meaningful production trends.
Note that the estimation of a resource quantity may well impact the resource classification
(e.g., a review of the profile or EUR/well might indicate questionable commerciality).
4.3.2 Methodology for the Application of Analogy. The procedure for using analogy should be
applied in a consistent, rigorous manner [the following is modified from Hodgin and Harrell (2006)]:
1. Determine the purpose of the analog and explicitly define the parameters that will impact
the validity of the analog. For resource classification, there may be some mandatory
parameter requirements.
2. Identify multiple analogs and compare the results with reference to the input parameters.
As described by Sidle and Lee (2010), should there be any outliers in terms of
performance, they need to be fully understood before progressing.
61
3. Evaluate the subject reservoir in light of the analog. In general, the subject reservoir
should be better than the analog reservoir, at least “in the aggregate” for recognizing
Proved Reserves. For example, for economic producibility, the porosity of the analog
reservoir may be better than the subject reservoir, but a more germane test would be the
reservoir transmissivity (Kh/µ, where Kh is horizontal conductivity, and µ is viscosity),
which should be better in the subject reservoir than in the analog reservoir.
4. Establish the proof of analogy under the relevant Reserves definitions to a mature
reservoir and recovery process.
The analog normally will help to generate the best estimate, or 2P, but it will inform the 1P
and 3P estimates too. If there are any weaknesses in the analog, this should give the evaluator
thought to widen the resource range. Such examples could be:
• Few applicable analogs
• Analogs with one or more parameters better than the subject reservoir
• Large range of “output values” (e.g., EUR/well) among the analogs
• A difference in scale between the analogs and the subject reservoir (e.g., small subject
reservoir, large analog reservoirs)
The Canadian Oil and Gas Evaluation Handbook [Society of Petroleum Evaluation Engineers
(Calgary Chapter) 2019] gives some useful advice on certain specific reserves parameters
(Canadian Oil and Gas Evaluation Handbook § 6.4.4.2). On areal assignments, when the analog
is being used to estimate the areal extent, it should be noted if progressively smaller pools are
encountered in a mature area, so the subject area should be reviewed on a regional basis. On
recovery factors, key elements such as abandonment pressures and fluid displacement efficiencies
must be considered, but the Canadian Oil and Gas Evaluation Handbook suggests that geographic
proximity, while desirable, is not essential. On performance characteristics, analogs can help
greatly, not only on the explicit decline, but also to understand the reservoir engineering aspects
such as water loading of gas wells in the future.
4.3.3 Application of Analogy to an Oil Example. Consider the case of a small offshore oil
discovery in a mature province with two appraisal wells in the reservoir. No core has been taken,
and no flow test has been conducted. The evaluator would like to estimate the recovery efficiency
(RE). It is assumed the development will be a subsea tieback to a host platform operated by a third
party. Gas lift will be supplied by the host platform.
The first question to be answered is whether there is an active aquifer or not. Absent a well
test, it will be necessary to use analogs. Using the methodology outlined above, the following steps
should be taken:
1. The purpose of the analog is to determine whether a natural aquifer will provide adequate
pressure support and thereby a reasonable RE, without recourse to the use of injection
wells. The analog will only be valid if it is in the same geological setting in the same
reservoir and it has similar petrophysical parameters, the same development scheme, and
similar fluid types. Given that a natural aquifer waterflood is the primary recovery
method, there are no additional requirements in the use of analogs for resource
classification.
2. Next, it will be necessary to gather multiple analogs by relying on internal, commercial,
regulatory, or other public data. Assuming there have been multiple developments in the
geological setting, how many developments have relied on natural aquifer support, and
how many have required injection wells? Have any of the analogs run into difficulty from
62
a recovery point of view? If there are any fields that are outliers (either they chose water
injection wells or relied on natural aquifer waterflood and ran into problems), then these
need to be understood in detail before progressing.
3. Assuming for this example that the natural aquifer is the norm in the geological setting,
that any outliers can be explained as being not relevant, and that the regional fields are
true analogs, then it will be necessary to compare petrophysical parameters between the
analog(s) and subject reservoir. This may be problematic if the regional analog data are
not public. In this instance, it may be necessary to cast the net wider using analog data
away from the regional area.
4. Following this process, other key questions must be answered using analogs, including
economic producibility, production rates, production trends, etc. A test against the
Reserves definitions must be made at this point.
Starting with the first two points, several analogs are gathered, concentrating on aspects that
would impact the aquifer support. Table 4.1 outlines the relevant parameters from analogs with
aquifers. Note that “overpressure” can indicate limited geological extent. Analog reservoirs are
also chosen based on magnitude/ranges of measurable reservoir factors such as permeability,
porosity, transmissivity, etc., as indicated earlier in this section.
Reservoir
Deposition Type Quality Extent Overpressure
Subject Deepwater marine Turbidite Excellent Continuous None
Analog A Deepwater marine Turbidite Excellent Continuous None
Analog B Shallow marine Sheet deposits Excellent Limited None
Analog C Deepwater marine Turbidite channels Excellent - fair Channelized None
Analog D Shallow marine Tidal channels Good - fair Channelized None
Analog E Shallow marine / deltaic Elongate / lobate Excellent - fair Continuous None
Analog F Shelf marine Shelf deposits Good Limited Significant
Faulting Aquifer
Density Throw Seal Strength Size Support
Subject Limited Sub-seismic Unknown Unknown Unknown Unknown
Analog A Limited Sub-seismic None Excellent Large Excellent
Analog B Limited Sub-seismic None Fair Fair Fair
Analog C Significant 10 m max None Good Large Good
Analog D Limited Sub-seismic None Good Large Good
Analog E Some 20 m max Sealing Poor Large Poor
Analog F Limited Sub-seismic Baffle Good Fair Fair
Table 4.1—Example analog parameters for aquifer support.
The next step is to interrogate the data, identify any outliers, and determine whether they are
relevant to the subject reservoir. In the example above, the only reservoir with poor aquifer support
(Analog E) is one with seismically identifiable fault throws, which we know is not the case in our
reservoir. However, there are two reservoirs (Analogs B and F) with fair support, but with limited
sand extent (identifiable through regional correlations).
In this instance, there is no strong evidence against good aquifer support, but the analogs are
few, and there are significant differences between the analog reservoir and the subject reservoir. It
63
is suggested that the evaluator assume a low case scenario, which has “fair” aquifer support, while
the best estimate is based on “good” aquifer support.
The analyst must employ their judgement to consider two options, either: to plan for injection
wells to be included in the low case, with a similar RE to the best case but with significantly more
capital expenditure, or to assume a reduced RE for the low case and add a new, separate project for
incremental resources attributable to water injection in the event of finding poor aquifer support.
In this example, Analog B could be a good candidate for the low case scenario (based on the
fair aquifer support and fair aquifer size rationale; Analog E has poor support but a large aquifer
size, plus some seismically identifiable faults); Analog C could be a good analog for the best case
scenario (choice based also on the aquifer support rationale, as well as on the reservoir
characteristics); and Analog A could be used for the high case scenario (based on analogy of
reservoir characteristics associated with excellent aquifer support).
4.4 Volumetric Analysis

Volumetric analysis, as described in the PRMS § 4.1.2.1, “uses reservoir rock and fluid properties
to calculate PIIP [petroleum initially in place] and then estimate that portion that will be recovered
by a specific development project.” Volumetric analysis is the most widely used technique for
early stage developments, but there can be many pitfalls associated with its use.
It is essential to consider the range of uncertainty throughout the estimation process, whether
via deterministic or probabilistic methods. Within deterministic procedures, it is normal practice
to consider low, best, and high estimates, but it is also advisable to consider alternative scenarios,
particularly for the fundamental assumptions, including horizon interpretation, velocity modelling,
structural interpretation, depositional environment, and geological unit correlation.
Consider the well and reservoir schematics in Fig. 4.3 below. Without additional data, such
as a detailed sequence stratigraphic model, there are numerous possibilities for correlation of the
flow units.
Fig. 4.3—Example of interwell correlation uncertainty.
The issue of reservoir continuity is a major concern in the volumetric estimation of petroleum
initially in place (PIIP). Geologic understanding of the depositional environment(s) and the use of
pressure and fluid data are key factors in the establishment of reservoir continuity or
compartmentalization. A field example of how correlations can change as more data are acquired
was discussed in Ray et al. (2010). In this field, a change in correlation following initial
development drilling moved a significant portion of the PIIP from a lower-quality unit to a higher-
quality unit, thereby dramatically increasing reserves and resources.
The general volumetric equation may be expressed in terms of EUR by:
EUR (STB or scf) = PIIP (STB or scf) × RE (fraction of PIIP), ...................................... (4.1a)
64
dependent upon average values for area (A), net pay (h), porosity (φ), initial water saturation (Swi),
and initial hydrocarbon formation volume factor (Bhi) for oil (RB/STB) or gas (Rcf/scf), where R
indicates reservoir conditions. The generalized classic volumetric equation for the PIIP [oil
initially in place (OIIP) or gas initially in place (GIIP)] is given by:
PIIP (STB or scf) = Ahφ(1 − Swi)/Bhi,....................................................................................... (4.1b)
where oil or gas volumes are in barrels or cubic feet, abbreviated as STB and RB or scf and Rcf,
respectively, representing the measurements at standard surface and reservoir conditions, based on
prevailing pressure and temperature standards.
4.4.1 Estimating Volumetric Parameters. Volumetric analysis should be generated by an

integrated technical team, including geoscience, reservoir and production engineering, and
facilities/process engineering. The first step is to quantify the static parameters, followed by an
assessment of the RE based on dynamic parameters. The RE is also a function of the anticipated
depletion strategy, and it can be either calculated or supported through the use of analogs.
4.4.1.1 Static Parameters. The gross rock volume is often the parameter with the greatest
uncertainty, and it must be explicitly defined. An area-depth map, whether generated traditionally
or as an output from geocellular models, is an excellent tool for estimating gross rock volume
ranges. The interpreter should consider alternative horizon and fault picks from seismic data as
well as alternative velocity modelling if there is any doubt in these areas. For regions of dense well
control, this will be less of an issue.
Other static parameters (porosity, net-to-gross ratio, hydrocarbon saturation) can then be
estimated from available data and/or supplemented by analogs. It is advisable, however, that
thought be given to the depositional setting, since a mathematical average of the wells may not be
appropriate in channelized systems, for example. Similarly, parameters may degrade further from
the sedimentological source, or they may degrade with greater depth due to diagenesis.
Geological conceptual models capturing the deposition of the sand bodies, particularly when
there are multiple/stacked sands, should be created. The way in which each sand body connects to
another (laterally and possibly vertically) will be critical in understanding PIIP and RE. Unless the
region is very well understood and shows strong correlation throughout, or if there are
unambiguous geological or geochemical markers, then it is expected that the sand correlation will
be nonunique and should be explicitly considered as an uncertainty. In addition to depositional
reservoir architecture, faulting within the field must also be considered because it will similarly
have an impact on RE as well as resource category.
Care must be exercised with the estimation of hydrocarbon phase contacts. Pressure gradient
measurements yield invaluable information, but all possibilities must be considered outside of well
control. For example, the existence of a thin oil rim may not be encountered with initial drilling,
yet it might make a significant impact on the development scheme, investment scope, and ultimate
recovery. A gas cap above the highest known oil is also possible, although the degree of
undersaturation should provide an indication of the potential for a gas cap. It should also be
recognized that the quality (i.e., representativeness) of the pressure-volume-temperature data is
paramount in establishing appropriate hypotheses. Representative pressure-volume-temperature
data are required for dynamic simulation and material balance calculations as well as for
volumetric estimates, which are dependent upon accurate formation volume factors for PIIP
assessment, considering the surface separation conditions.
65
Attention should also be paid to petrophysical cutoffs. Not all the microscopic hydrocarbon pore
space will contribute, and it is important for PIIP and RE to be calculated on the same basis. There
are arguments for and against using cutoffs, but this question should be explicitly considered. Further
discussion on petrophysical cutoffs may be found in Chapter 5—Petrophysics herein.
4.4.1.2 Dynamic Parameters. It is useful to break down RE into its component parts as
discussed in the subsections below and suggested by Smalley et al. (2009). The aim of their paper
was to maximize efficiency for a producing field, but it is also useful in early estimates.
• Pore-Scale Displacement Efficiency: This refers to the microscopic (displacement) efficiency
of the reservoir, and it also serves as the maximum theoretical RE obtainable under waterflood.
For example, microscopic RE for a waterdrive oil reservoir may be given by:
1−𝑆𝑆𝑤𝑤𝑤𝑤 − 𝑆𝑆𝑜𝑜𝑜𝑜
𝑅𝑅𝑅𝑅𝜇𝜇 = , .............................................................................................................. (4.2)
1− 𝑆𝑆𝑤𝑤𝑤𝑤
where:
REµ = microscopic RE of oil, fraction
𝑆𝑆𝑤𝑤𝑤𝑤 = initial water saturation, fraction
𝑆𝑆𝑜𝑜𝑜𝑜 = residual oil saturation, fraction.
Rough estimates for the fluid saturations can be made through initial core and log
measurements or through judiciously chosen analogs. This displacement is a function of rock-fluid
and fluid-fluid interaction properties, such as wettability, fluid mobility, relative permeability, and
capillary pressure. In some textbooks, REµ is referred to as “displacement efficiency” with the
symbol 𝐸𝐸𝐷𝐷 .
• Areal Sweep Efficiency: This refers to the areal connectedness of the producing well; i.e.,
the producing well may not be connected to other parts of the reservoir as a result of
faulting or the fact that the various sands are not stratigraphically connected. Smalley et
al. (2009) referred to this as “drainage efficiency.” Thought should be given to the
depositional environment (sand continuity, sand consistency, net-to-gross ratio), the
mobilities of the fluids, and the degree of faulting seen at seismic and core scale.
For early stage estimates, analog EUR/well gives a very good sense that drainage efficiency
has been appropriately considered. Some textbooks give areal sweep efficiency the symbol 𝐸𝐸𝐴𝐴 .
• Vertical Sweep Efficiency: Vertical sweep efficiency concerns the fraction of the vertical
section of the reservoir actually contacted by displacing fluid(s). This parameter is also
affected by fluid mobilities, in addition to vertical heterogeneity, any gravity segregation
forces, the quantity of injected fluid(s) if any, and so on. In some textbooks, vertical
sweep efficiency is denoted by the symbol 𝐸𝐸𝑉𝑉 .
• Volumetric Sweep Efficiency: In the literature, volumetric sweep efficiency is the
product of vertical and horizontal sweep efficiencies, and it represents the percentage of
the oil that is moved to the producing wells within the drained area. This would be a low
number for depletion-drive reservoirs and a higher number for waterdrive reservoirs.
4.4.1.3 Recovery Efficiency. The total RE, often referred to by the symbol 𝐸𝐸𝑅𝑅 , is the
multiplicative product of the microscopic displacement and volumetric sweep efficiencies or:
𝐸𝐸𝑅𝑅 = 𝐸𝐸𝐷𝐷 × 𝐸𝐸𝐴𝐴 × 𝐸𝐸𝑉𝑉 . ........................................................................................................... (4.3)
66
As mentioned previously, RE may be calculated, although it is acknowledged that this
estimate is inferior to RE determined from actual, lengthy, and reliable performance data. It must
be stressed that RE is a function of many parameters, such as the reservoir drive mechanism, any
improved oil recovery applications, the rock and fluid properties (e.g., wettability, relative
permeability, etc.), “complex situations” (discussed later), development characteristics, and even
the project economics.
For example, within the industry, there are empirical RE formulae generated “in-house” by
larger producing companies based on the reservoirs developed and producing by the company.
These REs are typically predicated on the study of RE performed by the American Petroleum
Institute in 1967 (American Petroleum Institute Subcommittee on Recovery Efficiency 1967; Arps
1968) and are expressed in the following form:
φ(1−𝑆𝑆𝑤𝑤 ) 𝐵𝐵 𝑘𝑘𝜇𝜇𝑤𝑤𝑤𝑤 𝐶𝐶 𝑝𝑝 𝐸𝐸
𝑅𝑅𝑅𝑅 = 𝐴𝐴 × � � ×� � × (𝑆𝑆𝑤𝑤 )𝐷𝐷 × � 𝑖𝑖 � , ........................................................... (4.4)
𝐵𝐵𝑜𝑜𝑜𝑜 𝜇𝜇𝑜𝑜𝑜𝑜 𝑝𝑝𝑎𝑎
where:
RE = RE of oil, fraction
A, B, C, D, E = displacement-dependent coefficients (see Table 4.2)
φ = effective porosity, fraction
𝑆𝑆𝑤𝑤 = initial water saturation, fraction
𝐵𝐵𝑜𝑜𝑜𝑜 = initial oil formation volume factor, reservoir barrels per stock tank barrel
(= bubblepoint formation volume factor for solution gas drive reservoirs
without primary gas cap)
k = arithmetic average of absolute permeability, Darcies
𝜇𝜇𝑤𝑤𝑤𝑤 = initial water viscosity, centipoise (= 1 for solution gas drive)
𝜇𝜇𝑜𝑜𝑜𝑜 = initial oil viscosity, centipoise (= 𝜇𝜇𝑜𝑜𝑜𝑜 for solution gas drive at bubblepoint
pressure)
𝑝𝑝𝑖𝑖 = initial reservoir pressure, psig (= bubblepoint pressure for solution gas drive)
𝑝𝑝𝑎𝑎 = abandonment reservoir pressure, psig.
Drive Mechanism A B C D E
Solution gas 0.41815 0.1611 0.0979 0.3722 0.1741
Waterdrive 0.54898 0.0422 0.077 - 0.1903 - 0.2159
Table 4.2—Recovery efficiency displacement-dependent coefficients.
“Complex situations” may include reservoirs that are nonvolumetric, stratified, fractured, or
overpressured, those characterized by thin oil rims with large gas caps and/or underlying aquifers,
dual-porosity/dual-permeability reservoirs, and so on. The fundamental equation above was not
formulated using examples of such reservoirs.
The reader should also be aware that the API (API 1984, p. 6) followed up the 1967 study
with a statistical review cautioning “against continued use of the (1967) correlations…to predict
recovery or recovery efficiency for any one reservoir.” Nevertheless, the 1967 equations are widely
used still, with local or company-specific variation as noted.
There are a number of other RE correlations available in practice, such as that developed by
Guthrie and Greenberger (1955), formulated as
67
𝑅𝑅𝑅𝑅 = 0.272 log(𝑘𝑘) + 0.256(𝑆𝑆𝑤𝑤 ) − 0.136 log(𝜇𝜇𝑜𝑜 ) − 1.538(𝜙𝜙) − 0.00035(ℎ) + 0.114,
............................................................................ (4.5)
where k is in millidarcies (md), but all other parameters are in the same units as in Eq. 4.4. This
correlation was developed for waterdrive sandstone reservoirs specifically. The use of recovery
factor correlations is cautioned, because they were developed for specific conditions with limited
databases using values for the key parameters that are subject to uncertainty.
Finally, several published empirical correlations that can be used to estimate RE can be found
in Cronquist (2001), Walsh and Lake (2003), and Craig (1971). However, it should be emphasized
that even a rough estimate of RE from a near-analog or a value determined by using a physically
based analytical method is preferable to using empirical correlations.
4.4.2 Application of Volumetrics to Oil Example. Let us consider the same example as presented
previously (Section 4.3.3), consisting of an offshore oil discovery with two appraisal wells. The
most usual method for determination of resources at this stage of development would be through
a combination of volumetrics and analogs, although a simple material balance model might also
be appropriate.
For the volumetric analysis, the first step is to establish the parameters that should be included
to define the range of PIIP and technically recoverable resources. A common mistake is to make the
assumption that the uncertainty range is limited to traditional parameters such as gross rock volume,
porosity, and the like, whereas more fundamental questions such as depositional environment,
reservoir correlation, top structure pick, and velocity model should be explicitly considered.
Some example parameters are shown below in Table 4.3 for our offshore oil example. Note
that there is a mix of static and dynamic parameters. The evaluator assumes that, while the horizon
interpretation is unambiguous (hence a single interpretation), the velocity model is highly
uncertain, and there are three possible scenarios or characteristics. These scenarios will be used in
defining the low, best, and high estimate cases.
Scenarios
Parameter Uncertainty Aspect Low Mid High
Depositional environment Regionally understood Shallow marine / Shallow marine Shallow marine
deltaic
Correlation Two geological models Correlation style “A” Correlation style “A” Correlation style “B”
Horizon interpretation Unambiguous Interpretation 1 Interpretation 1 Interpretation 1
Velocity model Significant uncertainty Methodology 1 Methodology 2 Methodology 3
Compartmentalization Low uncertainty Central fault sealing Central fault baffle Central fault open
Phase contacts Significant uncertainty Oil-water contact = Oil-water contact Oil-water contact at
Oil-Down-To between Oil-Down-To spill point
position and spill
Table 4.3—Fundamental uncertainty parameter table.
Next, a standard volumetric calculation is performed. The two key elements in Table 4.4
below that will have the most impact in the static calculation are the velocity model, which
generates differing top structure maps, and the treatment of the contact or lowest known
hydrocarbon (LKH). The column “low/best” highlights the relative difference between input
parameters. (Since this is an oil reservoir, the term OIIP has been used in place of PIIP. GRV is
gross rock volume, NTG is net to gross, So is oil saturation, and Bo is oil formation volume factor.)
68
Scenarios
Parameter Units Source Low Best High Low/Best
GRV (spill) MMRB Velocity model 1 550 887 1,842 62%

GRV (intermediate) MMRB Velocity model 2 413 665 1,107 62%
GRV (LKH) MMRB Velocity model 3 275 444 739 62%
NTG fraction Analogs 0.80 0.89 0.97 90%
Porosity fraction Analogs 0.27 0.29 0.31 93%
So fraction Analogs 0.81 0.85 0.91 95%
Bo rb/STB Analogs 1.07 1.10 1.16 97%
OIIP MMSTB 45.0 132.7 434.6 34%
Table 4.4—Example volumetric calculation: static parameters, where MM indicates million.
The dynamic aspects are considered next. Again, the two key uncertainties in Table 4.5 below
are correlation style (the degree of compartmentalization), which will affect the connectedness,
and the treatment of the faulting.
Recovery Efficiency Scenarios

Parameter Units Source Low Best High Low/Best
Pore/micro % Inputs from analogs 75% 76% 78% 98%

Drainage (open) % Correlation 80% 85% 93% 94%
Drainage (baffle) % Correlation 78% 83% 91% 94%
Drainage (sealing) % Correlation 70% 75% 83% 93%
Sweep % Range from analogs 53% 62% 71% 85%
Overall % Range from analogs 28% 39% 52% 71%
Tech resource MMSTB 12.6 52.2 223.8 24%
Table 4.5—Example volumetric calculation: dynamic parameters, where MM indicates million.
Note that the parameters that have the biggest influence are the fundamentals in the table (such
as the velocity model and the correlation) rather than the individual petrophysical parameters.
For the example above, the overall low case is a simple product of all the low input variables,
for clarity of the procedure only. In reality, the analyst should consider the combinations of
parameters that would represent a reasonable low case.
4.5 Performance-Based Methods

Performance-based methods include material balance, reservoir simulation, and PPT analysis. PPT
analyses are commonly used methods to directly estimate the technically recoverable resources for
oil and gas wells, reservoirs, and specific development (or recovery) projects. PPT analyses are
traditionally known as decline curve analyses (DCAs) but may also be considered to include
techniques that have found acceptance in unconventional resources assessment, such as rate
transient analysis and production decline analysis. (The reader is referred to Chapter 10—
Unconventional Resources Estimation, herein.) The theory behind traditional DCA was
crystallized in 1945 by Arps (1945), and it still remains one of the most prevalent and dependable
methods for production forecasting today, under specified flow conditions.
69
Historical production performance trends observed in mature wells, reservoirs, or projects
may generally be extrapolated to the cumulative production at the economic limit, and this will
provide a reasonable assessment of the technically recoverable resources when lift type and well
design are also considered.
To better comprehend the limitations of PPT analysis, Harrell et al. (2004) pointed out the
following conditions under which production decline trends would provide acceptable projections
of production profiles and resulting reserves estimates for the asset under study:
• Production conditions, methods, and the overall production strategy are not changed
significantly over the projected remaining producing life.
• The reservoir has been fully developed, and the well count is relatively stable. There is
no new project that may change the behavior of the reservoir.
• Wellbore interventions and other remedial work can be classified solely as maintenance.
This also implies that the evaluator will have sufficient reliable production data for the
analysis, and that the production for all phases, including oil, gas, condensate, and water, is well
established, with an emphasis on understanding water breakthrough maturity when an active
waterdrive is present.
PPTs are not only reservoir-specific, but they also depend on the particular reservoir
management, production practices, and facility constraints used. Any significant change in these
practices could easily lead to erroneous results. Therefore, the validity of production profiles
projected using DCA depends on both the quality and quantity of the past production data and also
on the evaluator’s professional experience gained through working on many hands-on assessments
and reconfirmations of results over time with actual performance, including the use of appropriate
analog reservoirs.
4.5.1 Material Balance. Material balance data include production and injection history, volume-
weighted average reservoir pressures, and reservoir-specific relevant fluid and rock properties, all
as a function of reservoir pressure and temperature. Independent of the volumetric methods, the
material balance methods can be used to directly and simultaneously estimate in-place volumes,
the relative size of a gas cap, and/or the water entering the reservoir through influx or injection.
The results of material balance analysis are considered more reliable when longer performance
histories, high-quality production data, and high-quality stabilized reservoir pressures are used. A
well-established and reasonable assumption is that use of the material balance analysis to estimate
in-place volumes is often considered valid if the cumulative production exceeds 10% of the
original in-place volume {Cronquist 2001, 30; or Canadian Oil and Gas Evaluation Handbook §
2.9 [Society of Petroleum Evaluation Engineers (Calgary Chapter) 2019]}, provided the
development of the accumulation is such that the pressures used in the analysis represent an
average over the entire reservoir. Uncertainty in the estimate is expected to decrease over time as
historical production performance data cover at least the early production period (shown in Fig.
4.1 as IIa) and beyond. A material balance analysis using all available data and information
typically should result in a best estimate of the in-place volume.
Technical principles and definition of the terms involved in developing the conventional
material balance equation (MBE) applicable to any oil or gas reservoir (i.e., black or volatile oil
and retrograde or nonretrograde gas) and applications may be found in Walsh and Lake (2003) and
Towler (2002). Modern flowing and dynamic material balance analyses developed by Mattar and
McNeil (1998) and Mattar and Anderson (2005) may also be used. The main difference between
a flowing material balance and a classic static material balance lies in the source of the pressure
70
data. The flowing material balance uses flowing pressures to estimate in-place volumes instead of
buildup or other static pressure measurements. One significant benefit to the flowing material
balance is that, so long as the data are of sufficient quality and frequency, it is not necessary to
shut in well or field production in order to obtain a good estimate of gas initially in-place (GIIP).
While the flowing material balance technique is more computationally intensive, the methodology
is generally intuitive to someone accustomed to classical material balance analysis.
A general discussion of the use of material balance methods follows, but more detail on special
topics is provided in Appendix A to this chapter.
4.5.1.1 Technical Principles of Oil Material Balance. The general MBE is commonly
expressed as:
𝑁𝑁𝑝𝑝 �𝐵𝐵𝑜𝑜 +�𝑅𝑅𝑝𝑝 −𝑅𝑅𝑠𝑠 �𝐵𝐵𝑔𝑔 �−�𝑊𝑊𝑒𝑒 −𝑊𝑊𝑝𝑝 𝐵𝐵𝑤𝑤 �−𝐺𝐺inj 𝐵𝐵𝑔𝑔inj −𝑊𝑊inj 𝐵𝐵𝑤𝑤𝑤𝑤
𝑁𝑁 = 𝐵𝐵𝑔𝑔 𝑆𝑆𝑤𝑤𝑤𝑤 𝑐𝑐𝑤𝑤 +𝑐𝑐𝑓𝑓
, ................................. (4.6)
(𝐵𝐵𝑜𝑜 −𝐵𝐵𝑜𝑜𝑜𝑜 )+(𝑅𝑅𝑠𝑠𝑠𝑠 −𝑅𝑅𝑠𝑠 )𝐵𝐵𝑔𝑔 +𝑚𝑚𝐵𝐵𝑜𝑜𝑜𝑜 � −1�+𝐵𝐵𝑜𝑜𝑜𝑜 (1+𝑚𝑚)� �∆𝑝𝑝
𝐵𝐵𝑔𝑔𝑔𝑔 1−𝑆𝑆𝑤𝑤𝑤𝑤
where
N = initial oil in-place, STB
Np = cumulative oil produced, STB
Bo = oil formation volume factor at current reservoir pressure, RB/STB
Rp = cumulative produced gas-oil ratio, scf/STB
Rs = current solution gas-oil ratio, scf/STB
Bg = current gas formation volume factor, RB/scf
We = cumulative water influx, RB
Wp = cumulative water produced, STB
Bw = water formation volume factor, RB/STB
Ginj = cumulative gas injected, scf
Bginj = gas formation volume factor of the injected gas, RB/scf
Winj = cumulative water injected, STB
Bwi = injected water formation volume factor, RB/STB
Boi = oil formation volume factor at initial reservoir pressure pi, RB/STB
Rsi = gas solubility at initial pressure pi, scf/STB
m = ratio of gas-cap gas volume to oil volume, RB/RB
Bgi = gas formation volume factor at pi, RB/scf
Swi = initial water saturation, fraction
cw = water compressibility coefficient, psi−1
cf = formation compressibility coefficient, psi−1
∆p = change in average reservoir pressure, (pi − p).
There are several practical aspects to consider when using oil material balance techniques that
will affect the quality of analysis results, a few of which are outlined in the sections below. For
additional detail on applying the methodologies presented, it is suggested to consult a textbook
such as Dake (1986, 2001) or Ahmed (2001).
71
4.5.1.2 Technical Principles of Gas Material Balance. In volumetric gas reservoirs, there is
by definition no (or insignificant) aquifer water influx, and the volume of initial hydrocarbon pore
volume will not significantly decrease from its initial value, but it will remain constant during
reservoir pressure depletion. Therefore, with no adjoining aquifer or water influx (We = 0), no
water production (Wp = 0), and no injection of gas (Ginj = 0), the generalized conventional MBE
for a volumetric, normally pressured gas reservoir reduces to (Lee and Wattenbarger 1996):
GBgi = (G − Gp )Bg, ......................................................................................................... (4.7)
where
G = GIIP, scf
Bgi = gas formation volume factor at pi, RB/scf
Gp = cumulative gas produced, scf
Bg = current gas formation volume factor, RB/scf.
Assessment of abnormally pressured gas reservoirs will necessitate the introduction of water
and formation expansion terms (i.e., ∆Vw + ∆Vf ) that must be added in Eq. 4.7; the reader is
referred to Lee and Wattenbarger for further detail. Otherwise, Eq. 4.7 may be rewritten to:
𝐵𝐵𝑔𝑔𝑔𝑔
Gp = G (1 − ). ............................................................................................................ (4.8)
𝐵𝐵𝑔𝑔
𝑧𝑧𝑧𝑧𝑝𝑝𝑠𝑠𝑠𝑠
The gas formation factor (Bg) can be calculated using Bg = , where standard surface
𝑇𝑇𝑠𝑠𝑠𝑠 𝑝𝑝
pressure (psc) and temperature (Tsc) conditions are 14.65 psia and 519.67°R (60°F), subject to local,
contractual, or regulatory variations.
It is common practice to express this relationship in terms of average reservoir conditions by
combining and rearranging the different elements to yield this well-known equation applicable
only to volumetric gas reservoirs:
𝑝𝑝𝑖𝑖
𝑝𝑝 𝑝𝑝𝑖𝑖 𝑧𝑧𝑖𝑖
= − 𝐺𝐺𝑝𝑝 , ............................................................................................................. (4.9)
𝑧𝑧 𝑧𝑧𝑖𝑖 𝐺𝐺
where
pi, p = average reservoir pressure (psia) at reservoir datum, and “i” stands for initial
T = average reservoir temperature at reservoir datum (°R)
zi and z = gas compressibility factors evaluated at pi and T and any p and T, respectively
G = GIIP (scf)
Gp = cumulative gas production (scf) at reservoir pressure (p).
Eq. 4.9 simply asserts that in volumetric gas reservoirs, the gas production, and therefore the
ultimate recovery under natural pressure depletion, is a direct function of the expansion of the free
gas initially in place. Furthermore, Eq. 4.9 suggests that a plot of (p/z) vs. Gp should yield a straight
line with an intercept (pi/zi) and a slope of [−(pi/zi)/G], from which the GIIP = G relationship can
be estimated. When an abandonment pressure is assumed (pab), the z factor at that pressure (zab)
and a corresponding EUR may be estimated.
72
An example of a (p/z) plot for a conventional, normally pressured volumetric gas reservoir,
with the variables mentioned above, is shown in Fig. 4.4.
Fig. 4.4—Conceptual plot of p/z versus cumulative gas production.
4.5.1.3 Considerations When Applying Material Balance to Gas Reservoirs. There are
several practical aspects to consider when using gas material balance techniques that may improve
the quality of analysis results, some of which are outlined in Appendix A. For additional detail on
applying the methodologies presented, the reader may consult a textbook such as Lee and
Wattenbarger (1996) or another source as presented in the references.
4.5.1.4 Material Balance: Comparison with Results from Other Methods. Once a material
balance analysis has been completed, it should be compared against other methods used to estimate
in-place volume, especially volumetrics. If there are conflicts between the material balance result
and other estimation methods, an attempt at reconciliation should be made to avoid over- or
underestimation of in-place volumes and technically recoverable resources.
Here are a few examples of situations that may arise:
• If the volumetrics-estimated in-place volumes are higher than the material balance–
estimated in-place volumes, and enough oil and/or gas has been produced from the
reservoir to provide confidence in the material balance results (generally at least 10% of
the in-place volume), perhaps there is compartmentalization in the field resulting in the
current wells only accessing a portion of the total volume in the area. In other words, the
MBE is indicative of the contacted area only.
• For a gas reservoir, if the volumetrically estimated GIIP volumes are lower than the
material balance estimated GIIP, and the initial reservoir pressure is high, it may suggest
that this is a case of an overpressured reservoir, and the p/z analysis is overestimating
the GIIP.
• If the technically recoverable resources estimated by production decline analysis are
significantly less or greater than the value estimated by material balance, perhaps an
incorrect value was applied as an abandonment pressure in performing the material balance.
73
• If the recovery factor estimated by material balance differs greatly from the values
obtained from analogs, perhaps a closer look at the analogs’ parameters is required to
explain the differences. In many cases, the reservoir properties may be a good match, but
there are differences in the recovery process that may not have been considered.
4.5.1.5 Application of Material Balance to Oil Example. The example oil project is a black-
oil reservoir, initially undersaturated (i.e., no gas cap) with partially active water influx, which was
developed by peripheral downdip water injection to supplement reservoir energy and to help
maintain a constant reservoir pressure 100 to 200 psia above the bubblepoint pressure. The project
produced 220.8 million STB of oil (15.4% OIIP of 1,429.6 million STB estimated), 126 billion scf
of solution gas, and 80 million STB of water and injected 385 million STB of produced and supply
water into the aquifer below the original oil-water contact. Furthermore, above the bubblepoint,
solution gas-oil ratios are equal (Rs = Rsi = Rp), as all gas produced at the surface would be dissolved
in the oil. The straight-line MBE (Havlena and Odeh 1963, 1964) is suitable for this particular
case, and its details may be found in the references cited.
𝐹𝐹𝑝𝑝 (𝑊𝑊𝑒𝑒 +𝑊𝑊inj 𝐵𝐵𝑤𝑤 )

= 𝑁𝑁 + , ............................................................................................ (4.10)
(𝐵𝐵𝑜𝑜𝑜𝑜 𝑐𝑐𝑒𝑒 Δ𝑃𝑃) (𝐵𝐵𝑜𝑜𝑜𝑜 𝑐𝑐𝑒𝑒 Δ𝑃𝑃)
where
Fp = cumulative net reservoir withdrawal, RB
N = oil initially in-place, STB
Boi = oil formation volume factor evaluated at pi and T
ΔP = change in pressure from pi to p, psia
We and Winj = water influx (in RB) and water injection (in STB), respectively
Bw = water formation volume factor, RB/STB
ce = effective saturation-weighted compressibility of the reservoir system,
(Soco + Swcw + cf)/(1 − Sw), psi−1
So = oil saturation, fraction
co = oil compressibility coefficient, psi-1.
This MBE represents reservoir depletion under a combined waterdrive (i.e., water influx
and/or downdip water injection into the aquifer) that is effective and strong enough to maintain
average reservoir pressure above the bubblepoint pressure. Since water is injected into the aquifer
at the periphery, it is treated as part of the water aquifer irrespective of how much of the water
actually enters the oil zone and helps to displace oil or how much of it enters the aquifer.
Eq. 4.10 suggests that a same-scale Cartesian plot of the left-hand side vs. the second term of
the right-hand side should yield a unit slope intercepting the ordinate at N (i.e., PIIP), as in
Fig. 4.5.
Data necessary for this plot can be generated at each timestep as follows: At any time period
with an appropriate ΔP, (1) the Fp and ce data can be calculated using the relevant relationships
and measured production and injection data, (2) the unsteady-state water influx theory of Van
Everdingen and Hurst (1949) may be used to estimate dimensionless influx rates (WD), and (3)
then the water influx (We) can be calculated.
Based on the average reservoir pressure observed and production and injection performance
data recorded over an 8-year period (the first-year data were erroneous, out of scale, and excluded),
the terms in Eq. 4.10 were calculated and plotted (Fig 4.6) in the same fashion as in Fig. 4.5.
74
Fig. 4.5—Material balance plot for oil reservoir with waterdrive.
With the variations shown in the plotted data, three parallel straight lines were drawn with a
unit slope supporting the value of the dimensionless radius, 𝑟𝑟𝐷𝐷 = 5 (defined as a ratio of the aquifer
radius to reservoir radius), and bracketing the degree of uncertainty in the measured and interpreted
data. These minimum, most likely, and maximum interpretations of PIIP (in this situation, oil,
hence OIIP or N) values of 1,300, 1,600, and 2,000 million STB were assumed to represent the
low, best, and high estimates, respectively.
Over the past 8 years, the ongoing peripheral waterflood project confirmed similar
performance to the analogs nearby, one of which had been since converted to a CO2 pilot, while
another was recently initiated, with encouraging performance similar to the initial response seen
in the first CO2 pilot. Recovery potential due to the implementation of a CO2 pilot in this project
would be incremental and subject to development contingencies.
Assessment OIIP for the Unsaturated Example Oil Reservoir

12,000
Estimate N (MMSTB)
8
10,000 High 2,100
Best 1,600
Low 1,300 6
7
F/(Boi Ce ∆P), MMSTB
5 3
8,000
2 4
6,000
For rD= (rA/rR) = 5
2, 3, 4 .... Time (years)
4,000
2,000
0
0 2,000 4,000 6,000 8,000 10,000 12,000
(We + Winj Bw)/(Boi Ce ∆P), MMSTB
Fig. 4.6—Assessment of OIIP by material balance methods (early production stage), where MM indicates million.
75
Based on these low, best, and high estimate OIIPs, the respective Technically Recoverable
Resources (under the ongoing peripheral waterflood project) and Contingent Resources (under a
proposed CO2 miscible project) were calculated as presented in Table 4.6.
Table 4.6—Assessment using oil material balance methods (early production stage), where MM indicates million,
B indicates billion, and BOE is barrels-oil-equivalent.
4.5.1.6 Application of Material Balance to Gas Example. Fig. 4.7 depicts the p/z vs. Gp
performance plots for this example volumetric wet-gas reservoir. Because of variations in the
observed data under pressure depletion, it was possible to draw three different straight lines
bracketing the potential degree of uncertainty in the measurement and interpretation of the
historical data. These low, best, and high estimate interpretations of PIIP (in this case, gas,
therefore GIIP) estimates from Fig. 4.7 may be used for assigning different reserves categories of
1P, 2P, and 3P, respectively, based on an estimated (p/z) economic limit of 1,500 psia. The
resulting implied volumetric reservoir RE is calculated to be about 75% to 76% of GIIP. Estimates
are further supported by the following factors and considered reasonable because: (1) the reservoir
has been established to be volumetric with nonretrograde gas; (2) it is fully delineated and
developed with a best estimate GIIP of 1,800 billion scf using volumetric methods; (3) it has
already produced 316.2 billion scf, which is more than 17.6% of this volumetric GIIP or 21.1% of
the low GIIP estimate from Fig. 4.7; and (4) the project economics based on these three different
scenarios are all determined to be viable with discounted cash flow rates of return exceeding 20%.
76
p/z (psi)
Fig. 4.7—Gas reserves assessment by material balance methods (late production stage), where OGIP is original gas
in-place, and B indicates billion.
Based on the initial condensate gas ratio of 30 STB/million scf raw gas (with a gross heating
value of 1,100 Btu/scf) and a recovery factor of 60% original condensate in-place from the nearby
analog reservoirs, the in-place and reserves estimates for this gas reservoir are summarized in
Table 4.7. Note that the recoverable raw gas volumes (in terms of both scf and therefore barrels-
oil-equivalent, BOE) summarized in Table 4.7 must be reduced by approximately 20% for the
surface loss to yield their residue sale gas equivalents or Reserves (EUR), consisting of 3.2% for
the shrinkage of condensate reserves and 16.8% for the subsequent processing to remove
non-hydrocarbons (8.1%) and recovery of C2 plus natural gas liquids (8.7%). For more detail,
readers should refer to Chapters 11—Production Measurement & Operational Issues and 12—
Resources Entitlement and Recognition, herein, on production measurement, reporting, and
entitlement issues.
It is a common practice to determine whether “gas compression” is economically viable and can
be used to lower wellbore backpressure to help gas wells produce at lower average reservoir
abandonment pressures (or associated p/z economic limits) and thus provide additional reserves.
The wellbore backpressure is the sum of the backpressure imposed by the sales gas pipeline
and the pressure drops in the gas gathering system at the surface and in the tubing string in the
wellbore. A gas well will stop flowing when the average reservoir pressure drops to this wellbore
backpressure. This “no flow” average reservoir pressure, and therefore its (p/z) value, does not
necessarily represent the economic limit because the wellbore-imposed backpressure can be
reduced by designing and installing an optimal gas compression facility (with an optimal
compression ratio) at the point of sales (or plant feed) to significantly reduce the sales gas pipeline-
imposed backpressure.
77
Table 4.7—Gas reserves assessment by material balance methods (late-production stage).
The economic limit (p/z) of 1,500 psia for this example reservoir represents a point where the
value of production is just equal to the operating cost of producing the project under pressure
depletion without compression. It is a deep gas reservoir, and although gas compression is
expected to reduce the economic (p/z) limit to as low as 1,000 psia, it is uneconomic because the
value of incremental gas reserves realizable is determined to be less than the capital and operating
costs of installing and running the compression facility. Thus, the incremental project volumes
associated with compression are considered as Contingent Resources pending future updates for
cost reduction and/or higher gas prices.
4.5.2 Reservoir Simulation. Traditional material balance analysis using analytical procedures is
routinely being performed during reservoir simulation studies. A history-matched numerical
simulation is a powerful tool, capable of modelling the complex interactions of production and
injection from multiple wells, changing pressure-volume-temperature properties, and changing
pressure gradients within a field to provide a rigorous material balance–based solution and, of
course, to compare alternative development scenarios. Reservoir simulation is further discussed in
Chapter 6—Reservoir Simulation herein.
4.5.3 Decline Curve Analysis. Decline curve analysis is well known and widely used because it
provides a visual illustration of the historical production performance of a well, a group of wells,
or a reservoir. The established trend can be extrapolated beyond the economic limit to estimate
petroleum reserves as well as subeconomic remaining resources. Review, derivation, and
understanding of these governing equations (summarized in Table 4.8), representing each decline
model, are very important for correct use and application of traditional DCA.
78
Table 4.8—Governing equations for decline curve analysis.
DCA is based on the solution of the following generalized differential hyperbolic equation
defining the nominal decline rate (D) as the fraction of “change in production rate with time (t)” as:
1 𝑑𝑑𝑑𝑑
𝐷𝐷 = − � � = 𝐾𝐾𝑞𝑞 𝑏𝑏 , ................................................................................................. (4.11a)
𝑞𝑞 𝑑𝑑𝑑𝑑
where
D = nominal decline rate (related but not equal to the slope of the tangent to the line) at
time (t); a fraction of production rate (Qt) with a unit of reciprocal time (1/t) per
79
month, year, etc., consistent with the units of production, frequently expressed as
percent per year
q = production rate (STB/D, STB/month, or STB/yr)
b = hyperbolic decline exponent
K = integration constant.
The reciprocal of the nominal decline rate is called the “loss ratio”:
1 𝑞𝑞
=− . ................................................................................................(4.11b)
𝐷𝐷 𝑑𝑑𝑑𝑑⁄𝑑𝑑𝑑𝑑
The hyperbolic decline exponent, b, is defined as the derivative of the loss ratio:
𝑑𝑑 𝑞𝑞
𝑏𝑏 = �− � ........................................................................................................ (4.11c)
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑⁄𝑑𝑑𝑑𝑑
The hyperbolic decline model is not only the most common decline trend observed in the actual
performance of oil and gas wells and reservoirs, but it also represents the most general and
challenging decline trend with two unknown parameters of initial nominal annual decline rate (Di)
and decline exponent (b). Moreover, the decline exponent (b) may have any value except b = 0 and
b = 1, which represent the special cases defined by exponential and harmonic models, respectively.
Note that the exponential and harmonic models are just specific cases of the hyperbolic model.
When originally conceived, the Arps-style approach assumed boundary-dominated flow, with an
upper limit of 1 for the harmonic decline case. Investigations since this original formulation have
demonstrated the presence of “hyperharmonic” exponents associated with transient flow regimes
in unconventional reservoirs. For example, the exponent observed for transient linear flow is 2,
while that for transient bilinear flow is 4. This is described further in Chapter 10—Unconventional
Resources Estimation, herein.
Initial nominal decline rate (Di) is the nominal (or continuous) decline rate corresponding to
the initial production rate at which decline begins (Fig. 4.8). It is a tangent to the initial production
rate at a given time. The ratio of nominal decline rate at any time t (i.e., Dt) to initial decline rate
(Di) when production decline first begins is proportional to a power b (except 0 and 1) of the
respective production rates and defined by:
𝐷𝐷𝑡𝑡 𝑞𝑞𝑡𝑡 𝑏𝑏
=� � . .................................................................................................................. (4.12)
𝐷𝐷𝑖𝑖 𝑞𝑞 𝑖𝑖
Di is related to the initial effective (or stepwise) decline rate (di), which is a step function rather
than a continuous function between two consecutive rates, represented as the secant of a rate-
versus-time arc connecting the two rates by the following relationships:
𝐷𝐷𝑖𝑖 = − ln(1 − 𝑑𝑑𝑖𝑖 ) or ....................................... 𝑑𝑑𝑖𝑖 = 1 − 𝑒𝑒 −𝐷𝐷𝑖𝑖 [exponential decline],...(4.13a)
and
1 1 1
𝐷𝐷𝑖𝑖 = �(1−𝑑𝑑 )𝑏𝑏 − 1� or 𝑑𝑑𝑖𝑖 = 1 − (1+𝑏𝑏𝑏𝑏 )1/𝑏𝑏 [hyperbolic decline]............................. .(4.13b)
𝑏𝑏 𝑖𝑖 𝑖𝑖
80
Fig. 4.8—Nominal compared to effective decline rate (IHS Markit 2022).
1
For example, if Di = 0.25/yr and b = 0.5, then 𝑑𝑑𝑖𝑖 = 1 − = 0.21/yr.
(1+ 0.5×0.25)1/0.5
The effective decline rate, di, is the one petroleum engineers are most familiar with, as it is
typically read directly from a rate-vs.-time plot. Many commercially available DCA packages rely
on the effective decline rate (secant) in their analysis; usually, there is not much difference between
the nominal and effective decline rates except in the case of unconventional reservoirs, where
initial steep declines have a significant impact—the larger is the Di, the greater is the difference
with di.
Rate of decline depends on several factors, such as the reservoir depletion rate, the reservoir
maturity, the average reservoir pressure, the reservoir rock and fluid properties (magnitude and
distribution), and the reservoir management and production practices.
The decline trend analysis of production rate vs. time advanced by Arps (1945) employs
hyperbolic equations similar to Eq. 4.11a, which are summarized in Table 4.8.
It has been widely reported that the value of “b” varies with reservoir drive mechanism.
Although the development of unconventional reservoirs in North America has resulted in observed
“b” values significantly exceeding one (e.g., b = 2 for linear flow and b = 4 for bilinear flow; see
Chapter 10—Unconventional Resources Estimation herein), the following generally applicable
values may be expected (Table 4.9):
Value of Decline Exponent (b) Governing Reservoir Drive Mechanism

0 Single-phase liquid (oil above bubblepoint)
0 Single-phase gas at high pressure
0.1–0.4 Solution gas drive
0.4–0.5 Single-phase gas
0.5 Effective edge waterdrive
0.5–1.0 Commingled layered reservoirs
b>1 Observed in unconventional reservoirs
Table 4.9—General hyperbolic exponents related to reservoir drive mechanisms (after IHS Markit 2018).
81
The governing rate-time relationship of a general hyperbolic decline model (Table 4.8) is:
𝑞𝑞
𝑞𝑞𝑡𝑡 = (1+𝑏𝑏𝐷𝐷𝑖𝑖 1⁄𝑏𝑏
. ......................................................................................................... (4.14a)
𝑖𝑖 𝑡𝑡)
Eq. 4.14a may also be rewritten as:

1 1
𝑙𝑙𝑙𝑙𝑙𝑙 𝑞𝑞𝑡𝑡 = 𝑙𝑙𝑙𝑙𝑙𝑙 𝑞𝑞𝑖𝑖 − log(1 + 𝑏𝑏𝐷𝐷𝑖𝑖 𝑡𝑡) = 𝑙𝑙𝑙𝑙𝑙𝑙 𝑞𝑞𝑖𝑖 − log(1 + 𝐶𝐶𝐶𝐶), ........................................ (4.14b)
𝑏𝑏 𝑏𝑏
where C = bDi, which is an unknown constant (refer also to Table 4.8). For a correct value of C,
Eq. 4.14b suggests that a log-log plot of q vs. (1 + Ct) should yield a straight line with a slope of
m (= −1/b) and intercept of qi at time zero.
Given the initial production rate at the onset of decline (qi) and other oil or gas production data
observed over the decline period, DCAs have been largely an exercise in curve fitting to establish
characteristic straight lines and/or type curves and conducting nonlinear regression analysis to
simultaneously estimate the correct values of these two unknown parameters Di and b.
With estimated correct values of these two unknowns, the cumulative production at any given
time can now be directly calculated using the following relationship:
𝑞𝑞𝑏𝑏 (1−𝑏𝑏) (1−𝑏𝑏)
𝑁𝑁𝑝𝑝𝑝𝑝 = 𝑁𝑁𝑝𝑝𝑝𝑝 + (1−𝑏𝑏)𝐷𝐷
𝑖𝑖
× �𝑞𝑞𝑖𝑖 − 𝑞𝑞𝑡𝑡 �, ........................................................................ (4.15)
𝑖𝑖
where qi and q represent the production rate at initial time i and any time t, respectively, and Npi
and Npt represent the cumulative oil production at the initial (i) production rate (qi) or from time
zero (t = 0) up to the time t.
It is difficult to determine the correct value of the decline exponent (b) while attempting to
estimate two unknowns (C and b) simultaneously. It is quite possible to have the same b but
different Di values matching the same decline trend that extrapolates to different estimates of
reserves, and this must be taken into account to quantify the uncertainty associated with
the analysis.
Ideally, one would estimate the nominal decline (Di) first and then perform a simple trial-and-
error procedure iterating on this single insensitive decline exponent b to evaluate and establish the
best-matched decline trend that corresponds to the best value of b. Also, b can be calculated as
follows: from Eq. 4.14b, when Ct >> 1 (which happens soon enough), we can determine b from
the slope of a log-log plot of q vs. t. This is almost always possible for wells in transient flow
(when D is large), whereas it might be more difficult for wells in boundary-dominated flow (when
Di is much smaller, and so time to reach Dibt >> 1 is longer).
In most cases, the well count in a producing asset (reservoir, field, etc.) changes with time.
The conditions may not be stable enough to perform the DCA on a reservoir basis. Also, as the
reservoir produces, wells will die, and the slope of the declining asset will change to reflect only
the remaining wells, a situation known as “survivor bias.” To remedy this situation, DCA should
be performed on a well-by-well basis (or completion-by-completion basis). The engineer will need
to use their judgment to extrapolate production trends despite inherent errors that will be caused
by noise in the data, inaccuracies in the b and D parameters, or even potential changes in dynamic
conditions that may not be obvious by looking at the data.
Well-by-well DCA, originally performed manually, is most often carried out using relevant
software. However, the results provided by commercially available software should always be
82
scrutinized in light of the characteristics of the reservoir as well as the depletion drive mechanisms,
which must be understood by the evaluating engineer. Production down-time, curtailment,
stimulation, platform maintenance activity, and other factors impacting production at capacity
must be understood. Furthermore, particularly in the case of hydraulically fractured wells, early
time cleanup must be excluded from the decline analysis. Determination of the end of the cleanup
period is often assisted by analyzing secondary phase ratios such as the gas-oil or water-oil ratios.
The analysis is typically completed based on plots such as the following (example shown is a
hydraulically fractured, low-permeability oil well):
Fig. 4.9 shows a log-log rate vs. cumulative production plot (left) and the same plot when the
production rate is adjusted by normalizing for pressure drop [i.e., q/(pi – pbhf)] on the right. No
trend can be identified without pressure-normalizing the rate. However, this technique requires the
availability of bottomhole flowing pressure data at intervals similar to the production data.
Fig. 4.9—Rate vs. cumulative oil (left); pressure-normalized rate vs. cumulative oil (right),
where BOPD indicates barrels of oil per day.
Fig. 4.10 illustrates the use of secondary ratios in identifying the portion of the production
history that extends beyond the cleanup period and thus may be used for DCA. In this example,
the evaluator defined the first 7 days as the cleanup period, based on the stabilization of the oil
production rate and the gas-oil ratio. However, the water-oil ratio (WOR) has not yet stabilized,
although it is declining favorably. The figure also includes the choke setting to ensure that
production fluctuations are not due to the changing of choke settings.
Fig. 4.11 is a typical semi-log plot of primary phase production rate vs. time. Again, this is an
example of a low-permeability oil reservoir, and the production history is quite short; nevertheless,
the figure demonstrates the way in which such a routinely used plot can be employed to determine
the decline rate under an exponential depletion mode. Semi-log plots have been used as the main
DCA tool for decades and are quite easy to generate and analyze in commercial software as well
as spreadsheet programs.
83
Fig. 4.10—Rate vs. time showing cleanup time, choke size, and secondary ratios, where GOR is gas-oil ratio, WOR is
water-oil ratio (both dimensionless), and WC is water cut (as a fraction).
Fig. 4.11—Rate vs. time, semi-log, primary phase, exponential fit. LR = loss ratio.
While the foregoing discussion has been on per-well DCA, DCA may be conducted for a
group of wells (such as for a unitized project); however, in doing so, the details of the underlying
individual well performances will be lost, and the aggregate production history likely will provide
misleading forecasts. This will be mentioned in Section 4.5.3.3. In addition, a curve fit of past
performance to extrapolate future production does not necessarily yield the Proved (1P) Reserves
forecast, as mentioned in Section 4.5.3.5. The evaluator needs to be aware that, for each DCA-
84
generated fit of past performance, there is a range of uncertainty for the predicted performance that
may be expressed in terms of 1P, 2P, and 3P quantities, as shown in the example of Section 4.5.3.4.
4.5.3.1 Supplementary Decline Curve Analysis Techniques. There are other well-established
production performance analyses that may be used to predict recoverable volumes based on trends
exhibited for a well and/or a reservoir even before the production rate begins to decline. These
reservoir drive-specific analyses were briefly discussed by Cronquist (2001). Salient points of
these methods may be summarized as follows:
• Cumulative Gas Production vs. Oil Production Trends: For oil reservoirs with solution
gas drive, a semi-log plot of log Gp vs. Npt may develop a trend that could be extrapolated
to estimate oil recovery, with the maximum Gp being equal to the original solution gas
in-place (GIIP = Rsi × OIIP).
• Water Cut, Oil Cut, or WOR vs. Cumulative Production Trends: These performance
trends have been found to be particularly useful in analyzing an oil reservoir producing
with waterdrive or producing with downdip water injection or a pattern waterflood. The
established trend is extrapolated to the economic water cut (fw) or WOR to obtain EUR
under the prevailing production method with which the trend has been established. It may
be useful to note the following reported observations:
o A plot of “log fw (or fo) vs. Npt” (i.e., semi-log plot) trend may turn down at small
values of fo, but this will occur earlier for light oils and later for viscous oils (Brons
1963).
o A semi-log plot of (WOR + 1) and total fluids withdrawal (Fp) vs. time (t) may help
to define oil rate trend (Purvis 1985). Additionally, a semi-log plot of “(WOR + 1)
vs. Npt” tends to be linear at WOR values less than 1 and therefore may help to define
performance trends at low values of WOR or water cuts.
o Ershaghi and Omoregie (1978) and Ershaghi and Abdassah (1984) recommended
that a plot of {ln[(1/fw) – 1] – (1/fw)} vs. Npt should be linear. However, they noted
that due to the inflection point of the fw vs. Sw curves, the method will work only at
higher water cuts when fw > 50%.
o For reservoirs with an active waterdrive or that are under waterflooding operations,
a semi-log plot of cumulative oil production (x-axis) versus oil cut can yield an
excellent extrapolation (at high water cuts) to the EUR. Generally, this method works
best for fw > 50%.
Actual PPT analyses require a thorough understanding of their semitheoretical technical bases
and the well-established and widely used methods and procedures. However, the correct
application of these procedures is not straightforward. One could easily and incorrectly obtain an
excellent match but end up with unreliable reserves estimates.
4.5.3.2 Type Curve Analysis. Fetkovich (1980) introduced a robust means with which to
match performance data to empirical type curves for the purpose of estimating not only decline
parameters but also certain productivity properties, such as skin or flow capacity. This approach
relies on the assumption that there are basically two flow regimes during the producing life of a
well, namely, a transient period (when the drainage radius is expanding, usually in early time) and
semi-steady-state or depletion stage (when the boundaries have been reached, and pressure in the
drainage area declines in an approximately uniform fashion). The approach is further predicated
on single-phase, slightly compressible, radial flow from a homogeneous reservoir.
Other forms of type curve analysis exist, such as Carter’s (1985) gas type curves and the gas-
to-liquid equivalent type curves of Palacio and Blasingame (1993), but discussion of the selection
85
and use of type curves is beyond the scope of this work. The reader is referred to the sources at the
end of this chapter.
4.5.3.3 Pitfalls and Potential Issues Associated with DCA/TCA. During any DCA exercise,
the evaluator also will need to consider carefully all the elements that will affect the behavior of
the wells. Such parameters will include, for example:
• Physical cutoff rates and maximum water cuts: Production trends can be extrapolated to
very low production rates [or very high basic sediment and water, (BSW)]; however, the
expected vertical flow performance of the wells must be taken into account (through lift
curves, for example) to ensure that the technical forecast is truncated according to
physical constraints. Likewise, water handling can constrain production rates due to
facility limitations.
• Test data vs. actual production data: The evaluator should review the data and filter
carefully any element that will affect the production trend. This includes potential
downtimes, allocation issues, validity/thoroughness of field measurements, noise, etc.
• Well aggregation: While DCA may be performed best on a well-by-well basis (or even
completion-by-completion basis), this may be an issue for the well aggregation in large
assets. The deterministic sum of P90s will be very pessimistic, and the sum of P10s will
be very optimistic. (The reader is referred to Chapter 8—Aggregation of Reserves,
herein.) What may appear to be a reasonable low case on a well-by-well basis may lead
to an unrealistic forecast when many low cases are aggregated.
• Individual vs. group level inputs: DCA likely will be misleading if performed for a group
of wells where production is reported, or at least supplied to the evaluator, at the group
level. This may be the case for a waterflood unit, in which the aggregate production trend
does not identify well workovers or rapidly rising individual well water cuts, for example,
or an asset that is operated by another entity that does not provide individual well
performance data.
• General decline trends and external parameters: The evaluator should understand the
general decline trend and avoid the influence of external parameters such as choke
changes. For example, a well may have a very low decline rate because the choke is
opened slightly on a regular basis until it reaches the maximum choke setting. Similarly,
some companies may set an annual depletion rate for their reservoirs or fields, and the
wells are choked back to adhere to this practice.
• All produced phases and constraints: The evaluator should consider all produced phases
in the DCA and any facilities or regulatory constraints. The evaluator should not merely
focus on the major phase production forecast, and a restriction may be caused by
secondary phase production.
• Outliers: Some reservoirs cannot easily be evaluated with DCA. Fractured carbonate
reservoirs may produce dry oil at a plateau rate for a long time, but the well will die very
quickly when water breakthrough starts. Low-permeability reservoirs cannot easily be
evaluated because the b factor will vary with depletion stage, and it can be greater than 1
in unconventional assets.
• Production metrics: Producing days vs. calendar days trends will be very different.
Calendar day production will introduce external parameters, shutdowns, etc. PPTs are not
only reservoir specific, but they also depend on the specific reservoir management and
production practices. Any significant change in these practices could easily shift and
change the previously established decline trends and invalidate their extrapolations.
86
Therefore, proper application of these procedures, to a large extent, depends on the
experience and skill levels of the professional evaluators and their ability to judge the
reasonableness of results obtained by comparing them to known analogs and/or other
performance-based methods.
4.5.3.4 DCA Application Example. A well of interest is located in the New Mexico side of
the US Permian Basin. The laminated sand-shale turbidite reservoir produces black oil under
solution gas depletion drive. Crude quality is 39 °API with a solution gas-oil ratio of 350 scf/STB
and initial formation volume factor of 1.25 RB/STB. Water production exists from the start, but
data from more mature wells indicate that the wells do not water out regardless of structural
position (this is a stratigraphic trap). Thirty-six months of capacity production has yielded the
semi-log rate-time relationship plotted in Fig. 4.12.
Fig. 4.12—Semi-log rate-time graph.
A curve fit of the data should represent the most likely, or 2P, forecast (Fig. 4.13). Honoring
the most recent data (approximately the last 18 months) results in a hyperbolic fit with a b exponent
of 0.75 and an initial nominal decline rate of 0.015/D (loss ratio of 88.6%/yr). The evaluator checks
the exponent against empirical values and finds that multilayered reservoirs usually may be
matched with values between 0.5 to 1.0. As a further check, the evaluator notes that core data
analysis characterizes the reservoir as extremely heterogeneous, with a Dykstra-Parsons
coefficient of 0.87.
However, the first half of the data is not matched as well. This may be explained as a revision
in decline as the more permeable layers have depleted to the point that the lower-permeability
layers contribute a greater percentage of the total flow rate at the later time. If the evaluator believes
that the late-time data actually represent all layering, a high-side projection results in Fig. 4.14. In
this case, the b exponent has been increased to 0.9, and the initial nominal decline rate has been
increased to 0.022/D (loss ratio of 90.4%/yr).
87
Fig. 4.13—Hyperbolic curve fit.
Fig. 4.14—High-side hyperbolic fit.
Finally, again looking only at the late-time trend, an exponential trend may also be fit, with
the resulting trendline (correlation coefficient of 0.977) shown in Fig. 4.15. This exponential
decline is 36.7%/yr and can serve as the 1P projection.
A comparison of the three profiles, now designated as 1P, 2P, and 3P, is depicted in
Fig. 4.16. A slight adjustment was made to the initial rates at the beginning of the projections.
88
Fig. 4.15—Exponential late-time fit.
Fig. 4.16—1P, 2P, and 3P projections.
The evaluator has independently verified that the economic limit, on a per-well basis, is
1 barrel of oil produced per day (BOPD), under the defined operating conditions of primary
(solution gas drive) recovery only. A terminal decline rate switching from hyperbolic to an 8%/yr
exponential decline has been observed in other wells producing from the reservoir, but the switch
point occurs beyond the remaining economic life of the well in question. Consequently, the
reserves (and EUR, including the prior production) for the three cases may be summarized as in
Table 4.10.
89
Case Reserves* EUR
1P 83 15,820
2P 4,371 20,108
3P 4,423 20,160
* Reserves to 1 BOPD Economic Limit
Actual Cum. Production = 15,737 STB
Table 4.10—Summary of 1P, 2P, and 3P projections.
The reserves evaluator will note the difference between the 1P and both the 2P and 3P
reserves. The 2P and 3P projections have economic lives of 6.2 and 8.3 years, respectively,
compared with 3.1 years for the 1P; however, the real effect is the initial instantaneous decline rate
of about 29%/yr for the 2P and 3P estimates relative to 37%/yr for the 1P estimate. The variance
increases steadily as the hyperbolic decline rate progressively becomes more shallow, nearing 9%
at the economic limit.
4.5.3.5 Uncertainty. As explained above, until the wells or the field have reached the end of
field life, there will always be some uncertainty inherent to this analysis. Even with good-quality
data, the b and D parameters may not be quantified satisfactorily for a number of reasons. In the
example above, the evaluator may get different outcomes despite a very steady trend compared to
most actual data observed.
It is important to note that, even an No Further Activity (NFA) case computed with DCA may
not correspond to a “Proved Developed Producing” case. The DCA should always be
representative of the full range of uncertainty, and the evaluator should be able to quantify a 1P
NFA (PDP), 2P NFA, and 3P NFA case. This also means that all flowing wells should contribute
to the 1P, 2P, and 3P cases, or, incrementally, that all flowing wells should have a Proved,
Probable, and Possible component.
However, uncertainty may be much lower for flowing wells than for new development
projects (Fig. 4.17). Using a single trend for the DCA instead of a full low-best-high range would
suggest that the uncertainty is so small such that the approximation 1P = 2P = 3P can be used.
While this may be practical in a very mature asset, it may not be realistic.
Fig. 4.17—Range of potential reserves narrows as uncertainty decreases.
90
Finally, DCA will provide very little information in the early stage of a field’s life when few
production data are available, and uncertainty is driven by the geology (gross rock volume, net-to-
gross ratio, porosity, etc.). In late life, performance curves may be used to assess the reservoir
production as geology is better understood and uncertainty is driven by fluid dynamics, connected
zones, pressure communication, etc.
4.6 Summary and Conclusions

Consistent with the PRMS guidelines on petroleum resources and reserves definitions,
classification, and categorization, different deterministic assessment methods and procedures have
been illustrated to estimate oil and raw gas resources and reserves.
Key takeaways from this chapter are as follow:
• Petroleum resources assessment is and must be a continuous ongoing technical process
supported by good practices and collaborative efforts across many disciplines.
• Petroleum resources assessment should use the methods most suitable for analyzing the
data available, including static geoscientific and engineering approaches as well as
dynamic actual production performance (both subsurface and facilities considered), and
these methods should be carried out by a collaborative multidisciplinary team of expert
evaluators.
• Assessment of subsurface petroleum resources is complex and subject to many
uncertainties in static and dynamic reservoir parameters coupled with regulatory,
operational, and economic uncertainties. Although exceptions will continue to exist, the
quantity of reliable data and degree of certainty in the estimates of PIIP and EUR are
expected to increase over time.
• Material balance methods rely upon pressure data gathered from wells. Ideally, collected
data should have sufficiently long testing times to reach stabilized buildup pressures. For
shorter shut-in durations, pressure points used for analysis should be the result of
extrapolation from pressure transient analysis of the acquired data or a comparable
correction method that can be used to estimate an appropriate stabilized pressure.
• A good practice when employing material balance is to first review reservoir pressure
data as a function of time through use of a plot (see Appendix A to this chapter). This is
especially helpful in low-permeability reservoirs, where pressures may not equilibrate
throughout the whole reservoir, causing multiple trends to appear. In cases where
significant pressure differentials exist, some means of averaging is needed to estimate a
uniform trend for the reservoir as a whole.
• In gas material balance analysis, initial p/z trends may suggest a linear relationship, but
care must be taken to understand the reservoir drive mechanism. Several drive
mechanisms may cause significant deviation from a straight p/z relationship over the life
of the reservoir and can lead to materially incorrect GIIP estimates.
• Irrespective of project maturity and the amount and quality of performance data available,
the reliability of the resource estimates largely depends on the ability of experienced
reserves evaluation professionals, not only to know the most appropriate methods to use,
but also to exercise prudent judgment, ensuring the reasonableness and validity of these
estimates by always comparing them with those estimated using different methods and/or
with known analog reservoirs.
91
• Use of the full PRMS classification and categorization matrix provides a standardized
framework for characterizing the estimates of marketable hydrocarbon volumes
according to their associated risks and uncertainties.
4.7 Acknowledgments
We wish to acknowledge the contributions and inputs of Charles Vanorsdale, Monica Clapauch
Motta, Rich DuCharme, Dan DiLuzio, and Dr. John Lee. A special acknowledgment also goes to
Yasin Senturk, author of this chapter in the previous version of the Guidelines for Application of
the PRMS.
4.8 References
Ahmed, T. 2001. Reservoir Engineering Handbook, second edition. Houston, Texas, USA: Gulf
Professional Publishing.
Ahmed, T. and McKinney, P. D. 2005. Advanced Reservoir Engineering. Oxford, UK: Elsevier
Inc.
Alberta Energy Regulator (AER). 2013. Directive 040: Pressure and Deliverability Testing Oil
and Gas Wells. Calgary, Alberta, Canada: Alberta Energy Regulator.
American Petroleum Institute (API) Subcommittee on Recovery Efficiency. 1967. A Statistical
Study of Recovery Efficiency. Dallas, Texas, USA: API Bulletin D14, American Petroleum
Institute.
American Petroleum Institute (API) Subcommittee on Recovery Efficiency. 1984. A Statistical
Study of Recovery Efficiency, second edition. Dallas, Texas, USA: API Bulletin D14,
American Petroleum Institute.
Arps, J. J. 1945. Analysis of Decline Curves. In Transactions of the American Institute of
Mechanical Engineers, Vol. 160, 228–247. SPE-945228-G. Richardson, Texas, USA:
Society of Petroleum Engineers. https://doi.org/10.2118/945228-G.
Arps, J. J. 1968. Reasons for Differences in Recovery Efficiency. Paper presented at the
Symposium on Petroleum Economics and Evaluation, Dallas, Texas, March 1968. SPE-
2068-MS.
Brons, F. 1963. On the Use and Misuse of Production Decline Curves. Producers Monthly
(September 1963): 22–25.
Carter, R. D. October 1985. Type Curves for Finite Radial and Linear Gas Flow Systems:
Constant Terminal Pressure Case. SPE J. 25 (05): 719–728. SPE-12917-PA.
Cole, F. W. 1969. Reservoir Engineering Manual, 285. Houston, Texas, USA: Gulf Publishing
Company.
Craig, F. F. 1971. The Reservoir Engineering Aspects of Waterflooding, Vol. 3. Richardson, Texas,
USA: Monograph Series, Society of Petroleum Engineers.
Cronquist, C. 2001. Estimation and Classification of Reserves of Crude Oil, Natural Gas, and
Condensate. Richardson, Texas, USA: Society of Petroleum Engineers.
Dake, L. P. 1986. Fundamentals of Reservoir Engineering, ninth edition (and previous versions).
Amsterdam, The Netherlands: Elsevier Science Publishing Company, Inc.
Dake, L. P. 2001. The Practice of Reservoir Engineering, revised edition. Amsterdam, The
Netherlands: Elsevier Science Publishing Company, Inc.
Ershaghi, I. and Abdassah, D. 1984. A Prediction Technique for Immiscible Processes Using
Field Performance Data. J Pet Technol 36 (4): 664–670. SPE-10068-PA.
https://doi.org/10.2118/10068-PA.
92
Ershaghi, I. and Omoregie, O. 1978. A Method of Extrapolation of Cut vs. Recovery Curves.
J Pet Technol 30 (2): 203–204. SPE-6977-PA. https://doi.org/10.2118/6977-PA.
Fetkovich, M. J. 1980. Decline Curve Analysis Using Type Curves. J Pet Technol 32 (6): 1065–
1077. SPE-4629-PA. https://doi.org/10.2118/4629-PA.
Guthrie, R. K. and Greenberger, M. H. 1955. The Use of Multiple-Correlation Analyses for
Interpreting Petroleum-Engineering Data. Proc., Drilling and Production Practice, 130–137.
Washington, DC, USA: American Petroleum Institute.
Harrell, R., Hodgin, J., and Wagenhofer, T. 2004. Oil and Gas Reserves Estimates: Recurring
Mistakes and Errors. Paper presented at the SPE Annual Technical Conference and
Exhibition, Houston, Texas, USA, 26–29 September. SPE-91069-MS.
https://doi.org/10.2118/91069-MS.
Havlena, D. and Odeh, A. S. 1963. The Material Balance as an Equation of a Straight Line. J Pet
Technol 15 (8): 896–900. SPE-559-PA. https://doi.org/10.2118/559-PA.
Havlena, D. and Odeh, A. S. 1964. The Material Balance as an Equation of a Straight Line—Part
II, Field Cases. J Pet Technol 16 (7): 815–822. SPE-869-PA. https://doi.org/10.2118/869-PA.
Hodgin, J. E. and Harrell, D. R. 2006. The Selection, Application and Misapplication of
Reservoir Analogs for the Estimation of Petroleum Reserves. Paper presented at the SPE
Annual Technical Conference and Exhibition, San Antonio, Texas, USA, September 2006.
SPE-102505-MS.
IHS Markit. 2018. Rate Transient Analysis. https://cdn.ihs.com/www/pdf/Rate-Transient-
Analysis.pdf (accessed 6 June 2022).
IHS Markit. 2022. Decline Analysis Theory.
https://www.ihsenergy.ca/support/documentation_ca/Harmony/content/html_files/reference_
material/analysis_method_theory/decline_theory.htm (accessed 6 June 2022).
Lee, J. and Wattenbarger, R. A. 1996. Gas Reservoir Engineering, Vol. 5. Richardson, Texas,
USA: Textbook Series, Society of Petroleum Engineers.
Mattar, L. and Anderson, D. 2005. Dynamic Material Balance. Paper presented at the Canadian
International Petroleum Conference (56th Annual Technical Meeting), Calgary, Alberta,
Canada, 7–9 June. PETSOC-2005-113.
Mattar, L. and McNeil, R. 1998. The “Flowing” Gas Material Balance. J Can Pet Technol 37 (2).
PETSOC-98-02-06. https://doi.org/10.2118/98-02-06.
Moghadam, S., Jeje, O., and Mattar, L. 2011. Advanced Gas Material Balance in Simplified
Format. J Can Pet Technol 50 (01): 90–98. SPE-139428-PA.
Palacio, J. C. and Blasingame, T. A. 1993. Decline-Curve Analysis Using Type Curves: Analysis
of Gas Well Production Data. Paper presented at the Low Permeability Reservoirs
Symposium, 26–28 April. SPE-25909-MS.
Petroleum Society of the Canadian Institute of Mining. 2004. Determination of Oil and Gas
Reserves. Calgary, Alberta, Canada: The Petroleum Society of the Canadian Institute of
Mining, Metallurgy and Petroleum.
Pletcher, J. L. 2002. Improvements to Reservoir Material-Balance Methods. SPE Res Eval &
Eng 5 (01): 49–59. SPE-75354-PA.
Poston, S. W. and Berg, R. R. 1997. Overpressured Gas Reservoirs. Richardson, Texas, USA:
Purvis, R. A. 1985. Analysis of Production-Performance Graphs. J Can Pet Technol 24 (4): 44.
PETSOC-85-04-03. https://doi.org/10.2118/85-04-03.
93
Ray, F. M., Pinnock, S. J., Katamish, H., and Turnbull, J. B. 2010. The Buzzard Field: Anatomy
of the Reservoir from Appraisal to Production, Vol. 7, 369–386. London, UK: Petroleum
Geology Conference Series, Geological Society, London.
Sidle, R. E. and Lee, W. J. 2010. An Update on the Use of Reservoir Analogs for the Estimation
of Oil and Gas Reserves. SPE Econ & Mgmt 2 (02): 80–85. SPE-129688-PA.
Smalley, P. C., Ross, B., Brown, C. E., Moulds, T. P., and Smith, M. J. 2009. Reservoir
Technical Limits: A Framework for Maximizing Recovery from Oil Fields. SPE Res Eval &
Eng 12 (04): 610–629. SPE-109555-PA.
Society of Petroleum Evaluation Engineers (Calgary Chapter). 2019. Reserves Definitions and
Evaluation Practices and Procedures. In Canadian Oil and Gas Evaluation Handbook
(COGEH), consolidated third edition (Online October 2019). Calgary, Alberta, Canada:
Society of Petroleum Evaluation Engineers.
Towler, B. F. 2002. Fundamental Principles of Reservoir Engineering, Vol. 8. Richardson,
Texas, USA: Textbook Series, Society of Petroleum Engineers.
Van Everdingen, A. F. and Hurst, W. 1949. The Application of the Laplace Transformation to
Flow Problems in Reservoirs. J Pet Technol 1 (12): 305–324. SPE-949305-G.
Walsh, M. P. and Lake, L. W. 2003. A Generalized Approach to Primary Hydrocarbon
Recovery, Vol. 4. New York City, New York, USA: Handbook of Petroleum Exploration and
Production Series, Elsevier.
Walsh, M. P. and Raghavan, R. 1994. The New, Generalized Material Balance as an Equation of
a Straight Line: Part 1—Applications to Undersaturated, Volumetric Reservoirs. Society of
Petroleum Engineers. Paper presented at the Permian Basin Oil and Gas Recovery
Conference, Midland, Texas, USA, 16–18 March. SPE-27684-MS.
94
Appendix A—Special Considerations for Material Balance Analyses
4A.1 Diagnostic Plots

A good practice when employing material balance methods is to first review reservoir pressure
data as a function of time in a plot. Major events that have occurred during the life of the reservoir
should be noticed as changes in the trend of the data, and they should be annotated on the plot or
included in the documentation supporting the estimate, as appropriate. For example, if the reservoir
pressure drops below the bubblepoint in a solution gas reservoir, then this will be evident as a
decrease in the rate of pressure decline, as shown in Fig. A-1.
Fig. A-1—General pressure-time plot for a solution gas drive reservoir (Dake 1986), where GOR is gas-oil ratio.
Other noteworthy items include the following, some of which are presented in Fig. A-2:
• There may be measurement errors, which may be evident in pressures that are too high
or too low when compared with the main pressure-depletion trend. This includes cases
where pressures have inadequately built up prior to measurement.
• There may be measurements that are taken from a separate reservoir or fault block that
exhibit a different, nonparallel depletion trend.
• There may be evidence of pressure support, causing a flattening or bending upwards of
the pressure data, especially in later times.
• There may be significant historical operational changes, including field shut-ins for
maintenance or abrupt production changes.
95
Bubblepoint?
Fig. A-2—Example of effects in a pressure-time plot.
Diagnostic plots can greatly enhance the material balance analyses. Care must be taken to use
the correct plot to represent the active drive mechanism because interpretation of an incorrectly
selected plot will give misleading results. For example, Fig. A-3a represents a gas cap drive
material balance analysis plot using the Havlena-Odeh approach, in which the in-place oil volume
is represented by the slope of the line, whereas in Fig. A-3b, which is used for waterdrive
reservoirs, the in-place oil volume (N) is represented by the y-axis intercept. In addition, an
appropriate straight-line analysis must be created, as indicated by the various possible incorrect
analyses indicated in Fig. A-3a and A-3b.
b.
a.
Fig. A-3—Oil material balance analysis plots: (a) estimating gas cap size; (b) estimating aquifer size (Dake 1986).
The terms in the figure above are defined as follows:
𝐹𝐹 = 𝑁𝑁𝑝𝑝 �𝐵𝐵𝑜𝑜 + �𝑅𝑅𝑝𝑝 − 𝑅𝑅𝑠𝑠 �𝐵𝐵𝑔𝑔 � + 𝑊𝑊𝑝𝑝 𝐵𝐵𝑤𝑤 ,
𝐸𝐸𝑜𝑜 = (𝐵𝐵𝑜𝑜 − 𝐵𝐵𝑜𝑜𝑜𝑜 ) + (𝑅𝑅𝑠𝑠𝑠𝑠 − 𝑅𝑅𝑠𝑠 )𝐵𝐵𝑔𝑔 ,
𝐵𝐵𝑔𝑔
𝐸𝐸𝑔𝑔 = 𝐵𝐵𝑜𝑜𝑜𝑜 � − 1�,
𝐵𝐵𝑔𝑔𝑔𝑔
96
where:
F = production term in the material balance equation, bbl
Eg = expansion of gascap gas term in material balance equation, rb/STB
Eo = expansion of oil and its originally dissolved gas term in the material balance equation,
rb/STB
m = ratio of gas-cap gas volume to oil volume, bbl/bbl
We = cumulative water influx, bbl
Np = cumulative oil produced, STB
Bo = oil formation volume factor at reservoir pressure, rb/STB
Rp = cumulative produced gas-oil ratio, scf/STB
Rs = current gas-oil ratio, scf/STB
Bg = current gas formation volume factor, bbl/scf
Wp = cumulative water produced, STB
Bw = water formation volume factor, bbl/STB
Boi = oil formation volume factor at initial reservoir pressure pi, rb/STB
Rsi = gas solubility at initial pressure pi, scf/STB
Bgi = gas formation volume factor at pi, bbl/scf.
4A.2 Low-Permeability Reservoirs

While permeability is not a direct input into material balance equations, it is important to consider
the effect of reservoir permeability on pressure. In a low-permeability reservoir (and also, in certain
conditions, in a reservoir developed with large well spacing), pressures may not equilibrate
throughout the whole reservoir. Fig. A-4a presents an example of a high-permeability reservoir
where all the pressures have equilibrated for each measurement, while Fig. A-4b presents a low-
permeability case where each well has a different pressure trend.
Fig. A-4—Comparison of high- and low-permeability reservoir pressure-time plots (Dake 2001).
In cases where significant pressure differentials exist, some method of averaging individual
well pressure declines is needed to estimate a uniform trend for the reservoir as a whole. One
simple approach recommended for oil material balances by Dake (2001) is to assign drainage areas
to each well and then estimate the average pressure decline by volume weighting of the pressures
for each drainage area.
97
4A.3 Volatile Oils
It is assumed that there is pressure equilibrium in a reservoir for each measured pressure data point,
and that fluid samples are representative of in-situ conditions. For a volatile oil system, more
caution is required around those assumptions due to the dynamic nature of compositions and
increased liquid dropout from produced gases. Conventional material balances with standard
laboratory pressure-volume-temperature (black-oil) data tend to underestimate oil recovery for
volatile oils (Ahmed 2001).
A volatile oil analysis requires use of a modified form of the oil material balance, such as that
presented by Walsh and Raghavan (1994). As reservoir pressure drops below the bubblepoint, and
there is a release of gas, there will also be an amount of volatilized oil present in the gas. The MBE
is modified by including the ratio of the volume of stock-tank oil to surface gas existing in a vapor
phase at reservoir conditions (Rv), representing the volatile gas-oil ratio.
4A.4 Material Balance Pressure Data Quality

Estimates of average reservoir pressure rely upon data gathered from wells. In an ideal situation,
all pressure measurements taken and appropriately corrected would represent the true average
reservoir pressure. This is generally a reasonable assumption for gas reservoirs where sufficiently
long testing times have allowed for a stabilized buildup pressure to be measured. As a guideline,
a pressure can be considered to be stable if evidenced by an increase in pressure of no more than
2 kPa/hr (0.3 psi/hr) over a 6-hour period (Alberta Energy Regulator 2013).
For shorter shut-in durations, where pressures may not be stabilized, pressure points used for
analysis should be the result of extrapolation from pressure transient analysis of the acquired data
or a comparable method, to estimate an equivalent stabilized pressure. If there is no evidence
provided of the length of shut-in of pressure data, a wider range of uncertainty should be
considered to encompass the possibility of insufficiently built up pressures.
All pressure data should be appropriately corrected to a common datum depth, such as the
reservoir midpoint in a well or the field-wide average depth of a reservoir. If wellhead pressure
measurements are used, then they must be properly corrected to estimate reservoir pressures at the
datum depth, considering fluid property variations that may occur throughout the length of the
wellbore. A common method of wellhead-bottomhole pressure drop estimation is to divide the
wellbore into segments and estimate the pressure drop for each segment, assuming each segment
has uniform fluid properties due to its relatively short length.
4A.5 Abandonment Pressure and Compression

The bottomhole flowing pressure of a well is the sum of a delivery point pressure (typically a plant
inlet pressure or sales pipeline pressure) and the pressure drops that occur in the gas gathering
system on the surface and in the tubing string in the wellbore. A gas well will stop flowing when
the average reservoir pressure approaches the minimum bottomhole flowing pressure at which the
system is able to operate. This minimum pressure is also known as the abandonment pressure.
Gas well abandonment pressure (pab) depends on surface and/or subsurface constraints as well
as economics. At the surface, the constraint is typically the pipeline gathering pressure; in the
subsurface, the reservoir permeability has a significant influence on the abandonment pressure
(low permeability usually results in higher pab, while higher permeability may result in lower pab).
In the absence of a predetermined pab, there are rules-of-thumb as a function of the completion
depth that may be applied for initial conventional gas reserves estimates. One such rule-of-thumb
for Canadian reservoirs (Petroleum Society of the Canadian Institute of Mining Institute 2004) is
98
to use 1500 kPa per 1000 m of depth (approximately 66 psi/1,000 ft). Similarly, a value of 100
psi/1,000 ft of depth has long been used for US fields in Texas and Oklahoma. These figures do
not reflect the benefits of compression.
The abandonment pressure and its corresponding (pab/zab) value do not necessarily represent
the economic limit because the minimum achievable flowing pressure can be reduced by
subsequent design and installation of a compression system in the gas delivery system (typically
installed at the plant or at the well). It is a common practice to determine whether gas compression
is economically viable and can be used to lower well flowing pressures to help gas wells achieve
lower abandonment pressures (and associated pab/zab values) and thus increase EUR. An example
showing an increase in EUR after installation of surface compression can be observed in
Fig. A-5.
Fig. A-5—Conceptual plot showing the effect of compression on EUR (Dake 2001).
Due to a reduction in flowing pressure, compression may significantly increase the gas
production rate of a well, which needs to be considered when modelling gas forecasts and
estimating economic viability.
4A.6 Nonlinear p/z Relationships

The key assumption of straight-line p/z analysis of a volumetric reservoir is that the system behaves
like a single tank; i.e., there are minimal pressure variations within the reservoir and no other
sources of addition pressure support. Although initial data points may show a linear trend,
depending on the drive mechanism present in the reservoir, there may be a deviation from a linear
p/z analysis trend, such as that shown in Fig. A-6.
Note that for most of the mechanisms shown in Fig. A-6, it can be observed that the curves
initially appear to exhibit a linear relationship. If only the early data are used for p/z analysis
estimates, then this would lead to erroneous GIIP estimates.
If the reservoir drive mechanism can be correctly identified, then appropriate correction
factors may be applied to either adjust or relinearize the p/z plot. In many cases, a more
complicated solution, such as the solving of multiple simultaneous material balances, may be
required—a task often undertaken using reservoir simulation.
99
Fig. A-6—Conceptual p/z plot showing nonlinear curves due to various reservoir drive mechanism types. CBM stands for
coalbed methane (Moghadam et al. 2011).
4A.7 Overpressured Reservoirs

An overpressured reservoir, which includes abnormally pressured and geopressured reservoirs, is
defined as a reservoir with a high pressure gradient, generally above 0.5 psi/ft (Ahmed and
McKinney 2005). In an overpressured reservoir, the initial formation compressibility may be the
same order of magnitude as the gas compressibility. Initially, the formation contributes a
significant amount of energy to the reservoir. As the pressure in the reservoir is depleted, the
compressibility of the gas increases much faster than the formation compressibility, until the effect
of the formation compressibility is negligible.
Material balance analysis of an overpressured reservoir yields two distinct slopes: a shallow
trend in the pressure range where the formation expansion plays a significant role and a steeper
trend when gas expansion is the single dominant production mechanism. A p/z analysis of initial
data would significantly overestimate GIIP for an overpressured reservoir, as shown in Fig. A-7.
Fig. A-7—Typical overpressured gas reservoir dual-slope p/z plot (Poston and Berg 1997), where cg indicates
compressibility of gas.
100
4A.8 Waterdrive Reservoirs
As pressure declines in a reservoir, there may be influx of water from an aquifer, which acts to
provide pressure support. This is generally shown as initial volumetric behavior followed by an
upwards bend in the p/z analysis line. The strength of the pressure support system can be inferred
by the time at which the bend in the analysis line is evident after initial production and the
magnitude of the deviation from volumetric behavior.
While the p/z analysis line may indicate the presence of a significant waterdrive mechanism,
it generally does not allow for identification of the magnitude of the drive. Cole (1969) suggested
a simple plot of GpBg/(Bg − Bgi) vs. Gp that would provide an indication of the strength of a
waterdrive depending on the shape of the resulting curve, as shown in Fig. A-8.
Fig. A-8—Cole plot comparing different waterdrive efficiencies (Pletcher 2002).
Due to gas being trapped by encroaching water, recovery factors for gas reservoirs with a
significant waterdrive component tend to be 30% to 50% lower than their dry gas expansion
counterparts; this effect may be partially mitigated in high-pressure and high-temperature (HPHT)
reservoirs. Identification of the correct pressure support mechanism may be helpful in avoiding
overestimation of recoverable volumes.
4A.9 Retrograde Condensate Reservoirs

When performing a material balance on a gas condensate reservoir, it is good practice to use a two-
phase gas deviation (z) factor when estimating volumes of gas below the dew point and at higher
condensate yields.
The two-phase z factor is used to account for liquid dropout in the reservoir as pressure
declines and changes occur in fluid composition in the reservoir. Generally, if only single-phase
compressibility factors are used, then there will be an underestimation of gas and condensate
recoverable volumes. Fig. A-9 shows a conceptual example of how the two values may differ,
although the shape of the curves will vary based on composition and in-situ conditions.
In Fig. A-9, the two curves would provide similar results when the pressure is close to initial
conditions. As the pressure in the reservoir is depleted, condensate dropout commences, which
leads to a divergence in the behavior of the z factor curves.
101
Fig. A-9—Example of equilibrium and two-phase gas deviation factors for a gas condensate reservoir
(Lee and Wattenbarger 1996).
Two-phase z factors may be obtained from laboratory tests or estimated from fluid
composition through use of an equation of state. Specifically, two-phase z factors are measured
from constant-volume depletion tests (Lee and Wattenbarger 1996).
4A.9.1 Considerations Related to Liquid Recovery Comments. When estimating liquid

recovery from gas condensate reservoirs, note that the condensate yield changes as the reservoir
pressure is depleted. A typical condensate yield curve is shown in Fig. A-10. Estimating liquid
recovery using a gas EUR estimated from material balance and an initial condensate gas ratio may
lead to significant overestimation of liquid recovery.
Fig. A-10—Example of condensate yield curve (Dake 2001).
To properly provide an economic estimate of natural gas liquid reserves and/or resource
volumes, the following must be known:
• The reference point:
o As described in § 3.2.3.1 of the PRMS, the reference point location will define if
produced gas will be sold as wet gas or as dry gas and extracted liquids. The
102
economic value of those two products may vary significantly depending on
geographical region and available transport facilities.
• The equipment available at the facilities used to treat the gas:
o The processing equipment used will define the technical limits of liquid extraction
efficiency. In a smaller facility, only heavier hydrocarbons, such as pentane and
heavier hydrocarbons, might be easily extracted from the gas stream, whereas a full-
size gas plant facility may be able to remove virtually all liquids and even
components as light as ethane from a gas stream.
• The contractual arrangement related to the value of the liquids:
o Assuming liquids are removed prior to the reference point, there must be a contract
or arrangement in place that allows the producer entitlement to their value.
103
Chapter 5
Petrophysics
Luis Quintero (Chair)
Javier Miranda, Joshua Oletu, Cecilia Flores, George Dames, and Philip Gibbons
5.1 Introduction
Petrophysics (from the Greek πέτρα, petra, “rock” and φύσις, physis, “nature”) may be defined as
the study of rock properties (physical, electrical, chemical, and mechanical) and their interaction
with fluids (gases, liquid hydrocarbons, and aqueous solutions) (Archie 1950, 1967; Tiab and
Donaldson 2004; Chen and Pagan 2013). Petrophysical, or formation, evaluation (used
interchangeably; Archie 1967; Asquith and Krygowski 2004) is thus a practice that integrates
knowledge from several disciplines, including, but not limited to, geology, geochemistry,
geophysics, physics, chemistry, and reservoir and production engineering.
The discipline of petrophysics provides key parameters (e.g., net pay thickness, porosity, and
saturation) used in the volumetric estimation of petroleum initially in place (PIIP), helps to
determine reservoir fluid type, and characterizes the ability of the fluid to flow (permeability);
these properties are critical in assessing the potential and development of any petroleum resources
or reserves. The fluid type can be characterized with properties such as density, hydrogen index,
and viscosity, among others. All these properties are the subject of wireline electrical logs,
formation testing/fluid sampling devices, and core sampling.
Fig. 5.1 shows generalized sources of petrophysical data over different phases of exploration,
appraisal, and field development, aligned with the Petroleum Resources Management System
(PRMS) resources framework.
Fig. 5.1—Petrophysical support for resources and reserves classification and categorization (modified from PRMS § 1.1).
104
The range of petrophysical data acquisition will vary depending on basin, cost, environmental
issues, and/or local regulations. However, the overriding objective is to acquire as much quality
data on prospective reservoirs from exploration through appraisal phases to aid decision making
and move each project through the different resources classes. In practical terms, while some
petrophysical data are acquired throughout the life of the project (basic well logs, mud logs, etc.)
others, such as core and special well logs, are mostly focused toward the early stages. Additionally,
uncertainty analysis of subsurface petrophysical data assists in defining the appropriate resources
classes and categories as shown in Fig. 5.1.
Early in the life of a project, volumetrics and/or analogy are the foremost methods by which
PIIP and resources are assessed. As a project matures, performance and/or material balance
methods, including reservoir simulation, may supplant volumetric and/or analog methods.
However, the in-place volumes derived from these methods should remain consistent with PIIP
volumes from petrophysical evaluations (see volumetric Eq. 4.1b in Chapter 4—Assessment of
Petroleum Resources using Deterministic Procedures, herein).
Petrophysical parameters generally are the primary means of qualifying analogs used as the
basis of the range of potential recovery factors expected from a development project, the optimal
type of development project to be used, and the development spacing. The resources associated
with a project are dependent on the development plan and the expected recovery factor, both of
which may be supported by proper application of petrophysically based analogs. Analogous
resources founded, in part, on petrophysical parameters may also be used as the basis to declare a
discovery where a flow test is not performed, but it requires confidence in the presence of
hydrocarbons and evidence of producibility (PRMS § 2.1.1.1). See also Chapter 4—Assessment of
Petroleum Resources using Deterministic Procedures herein.
This new chapter of the Guidelines for Application of the Petroleum Resources Management
System emphasizes petrophysical evaluation methods used routinely in the estimation of in-place
and recoverable volumes of petroleum. This chapter provides a high-level review of relevant
petrophysical evaluation methods that are considered pertinent to the assessment of PIIP and
resources and the inherent, associated uncertainties. While all the petrophysical input parameters
are very important and critical, this chapter will focus on net pay assessment, the importance of
establishing reservoir continuity, and core analysis, including residual oil saturation assessment.
The term net pay, officially introduced in the PRMS (2018), is defined as “the portion (after
applying cutoffs) of the thickness of a reservoir from which petroleum can be produced or
extracted. Value is referenced to a true vertical thickness measured” (PRMS Appendix A). True
vertical thickness refers to the thickness of the bed measured perpendicular to the center of Earth,
regardless of bed orientation. Within the PRMS, it is referenced with respect to the analytical
procedures of volumetric analysis and analogy for the purpose of estimating recoverable quantities
of petroleum. A goal of this chapter, therefore, will be to address the identification and application
of net pay and pay cutoffs from the petrophysics perspective.
105
5.2 Volumetric Estimation of PIIP and Estimated Ultimate Recovery
A time-honored means for estimating the amount of PIIP is the volumetric method, defined by the
PRMS (§ 4.1.2) as a “procedure (that) uses reservoir rock and fluid properties to calculate PIIP
and then estimate that portion that will be recovered by a specific development project.” Details
regarding the volumetric method may be found in Chapter 4—Assessment of Petroleum Resources
using Deterministic Procedures, herein.
In Eq. 4.1b (Chapter 4—Assessment of Petroleum Resources using Deterministic Procedures),
petrophysical evaluation results comprise fundamental inputs to the assessment of PIIP: the
hydrocarbon-bearing area above fluid contacts or limits (A), net pay thickness (h), porosity (ϕ), and
initial water saturation (Swi). Other petrophysical inputs, including, but not limited to, permeability
(k), capillary pressure, relative permeability, and wettability, are used to assess the ability of fluids
to flow, informing the assessment of the recovery factor or recovery efficiency (RE) in Eq. 4.1a
(Chapter 4—Assessment of Petroleum Resources using Deterministic Procedures). These inputs
may be incorporated into dynamic models to assess potentially recoverable volumes under various
development scenarios. The PIIP and recoverable resources associated with a project are dependent
on the development plan and the expected recovery efficiency.
5.3 Petrophysical Evaluation

Petrophysical evaluation involves the integrated analysis of wellbore information, including
cuttings and data from core samples, well logs, fluid samples, and pressure tests in both a static
(geological) and dynamic context. Physical rock sample (cores and cuttings) data may include
lithologic and mineralogical data, laboratory-derived measurements of porosity, permeability, and
fluid saturation, and pore and pore throat size characterization data (e.g., from capillary pressure
or nuclear magnetic resonance measurements). Fluid samples may provide both water and
hydrocarbon property data, such as water salinity and fluid viscosity. Well logs, collected via tools
run into the borehole while drilling, after drilling, and/or during production testing, provide
additional information, including, but not limited to, electrical potential, electrical resistivity,
natural and/or induced radioactivity, electron density, acoustic velocity, nuclear magnetism,
borehole configuration and orientation, pressure differential, temperature, and fluid flow rates.
Pressure test data may include point measurements of reservoir pressure and apparent mobility
from well logs and/or pressure transient data collected during production tests. These data are
reviewed and analyzed using a variety of established industry practices, relations, equations, and
trends to determine the rock and fluid properties, forming the basis for estimates of PIIP and
resources in hydrocarbon- and non-hydrocarbon-bearing formations.
It is noteworthy that most industry-accepted equations or algorithms are empirical in nature and
vary in their applicability and uncertainty, depending on the characteristics of the formation. According
to Worthington (2011), quantitative petrophysical evaluation is mostly data driven, and interpretive
algorithms change from reservoir to reservoir. Hence, the uniqueness of the reservoir, the quantity and
quality of data and information available for analysis, the applicability of various industry algorithms,
and associated uncertainties should be considered in every petrophysical evaluation.
5.4 What is “Net Pay”?

The concept of “net pay” originated in the early days of petroleum engineering when producing
companies, and their investors, were interested in knowing the portion of a reservoir that is capable
of providing a return on their investment, and therefore the portion that would “pay” for
development (in the context of other commerciality considerations).
106
Within the industry, net pay assessments may include a lithologic/mineralogical, minimum
porosity, minimum permeability (where such information is available), and maximum water
saturation cutoff. A related but different industry term, known as net reservoir, excludes a water
saturation and/or permeability cutoff. In other words, net reservoir is the portion of the gross rock
that contains mobile, producible fluids including water and/or hydrocarbons, whereas net pay is
that portion of the reservoir from which petroleum may be produced or extracted for exploitation
and/or development. A schematic illustration of the application of cutoffs was described by
Worthington and Cosentino (2003) and is presented here in Fig. 5.2. Moving from left to right,
this illustration offers proper distinctions among “gross rock,” “net reservoir,” and “net pay.”
Fig. 5.2—Net reservoir and net pay description (modified from Worthington and Cosentino 2003).
5.4.1 Pay Cutoffs. In order to differentiate the pay section from the gross volume, petrophysicists
rely on pay cutoffs. Although the term is used in the PRMS, it was not defined. There are several
methods to establish cutoffs, out of which three are fairly common: the statistical approach,
whereby the operators’ cumulative experience in the area dictates cutoff values; the no-harm
approach, whereby no cutoffs are applied, and a dynamic reservoir simulation defines those
volumes of fluids that will flow; and the petrophysical approach, in which the rock and fluid
properties dictate the moveable hydrocarbon (under the proposed displacement mechanism). This
chapter will elaborate only on the petrophysical approach.
A typical process for selecting cutoffs to delineate the pay section involves:
• The selection of petrophysical parameters that relate to hydrocarbon presence and flow
potential.
• The selection of a cutoff value for each petrophysical parameter.
Pay cutoffs are used to delineate the pay sections because:
• Not all rocks (or layers of rocks) in a target formation host moveable fluids (water and/or
hydrocarbon).
107
• Not all rocks host potentially moveable hydrocarbon (and/or water) that may be
recoverable based on the extraction technology employed.
• Not all fluids in the rock significantly contribute to the energy of the reservoir system.
Historically, pay cutoffs are assigned to well log characteristics based on past performance of
other reservoir analogs until actual performance in the subject reservoir supports a revision of one
or more such cutoffs. It is not unusual for a minimum log porosity cutoff to be further reduced
after dynamic data establish fluid flow to a lower threshold. This uncertainty may constitute the
difference between resource category assignments for the evaluator.
The selection of petrophysical parameters depends on the quality and quantity of available
data and the formation complexity. Worthington and Cosentino (2003) compiled some of the
industry classification schemes and pay cutoff parameters as shown in Table 5.1a and Table 5.1b.
Table 5.1b shows various combinations of petrophysical parameters that have been used for
assessing net pay in the literature, including shale volume, porosity, permeability, water saturation,
resistivity, and mobile fluid index. Applying cutoffs to the said properties serves to delineate the
portion of the reservoir in which hydrocarbon is stored and able to flow under the anticipated
production mechanism.
In addition to typical shale volume, porosity, permeability and water saturation cutoffs,
parameters such as mobility from formation pressure data and fluid entry data from production
logging tools in conventional plays or geochemical (e.g., total organic carbon) and geomechanical
(e.g., brittleness) properties in unconventional plays may be used as cutoff parameters.
In the early exploration and appraisal phase, when well and/or core data are absent,
insufficient, or not fully representative (as discussed later), the selection of a cutoff (or range of
cutoffs) is often based on experience. As relevant log and core data and/or production tests become
available, cutoff parameters may be refined. In the development phase, when significant
petrophysical data exist, it is advised that cutoffs be data-driven, through the integration of
available well log, core, fluid, pressure, and, especially, production data.
Table 5.1a—Some net pay classification schemes (from Worthington and Cosentino 2003; see original publication for
references cited in table).
108
Table 5.1b—Selected examples of cutoffs application in reservoir studies (Worthington and Cosentino 2003; see original
publication for references cited in table).
The complexity of the formation should be considered when selecting the relevant pay cutoff
parameters and values with which to reasonably assess net pay. In certain clastic reservoirs, as
shown in Example A (at the end of this chapter), commonly acquired well logs may be sufficient
to define net pay intervals. However, in more complex formations, additional data sources such as
nuclear magnetic resonance logs or core data may be utilized, as shown in Examples B and C,
respectively. Other information that may be utilized in complex formations includes, but is not
limited to, dielectric data, mineralogy, and production test data. Examples of reservoirs requiring
additional logs and data include, but are not limited to, shaly sands, radioactive sands, freshwater
sands, reservoirs dominated by microporosity, complex/mixed lithology reservoirs with varying
grain density, thinly laminated and low-reservoir-quality formations, fractured and/or tight
formations, and unconventional reservoirs.
It is important to keep in mind that one of the criteria for “Recoverable Resources” is that the
petroleum should be producible (PRMS Glossary). Most of the petrophysical techniques described
above are “static” in nature, and any pay cutoffs derived from these techniques may not reflect the
“dynamic” aspect of producibility. Consequently, a production logging tool may be incorporated
to identify, under flowing (as well as static) conditions, intervals of a reservoir that will contribute
109
to production (whether petroleum or water). Example D shows the application of a production
logging tool in the identification of pay zones. Furthermore, the application of cutoffs should be
reviewed as a recovery mechanism progresses from primary depletion to secondary or enhanced
oil recovery.
For unconventional petroleum accumulations, in which advanced extraction techniques
typically define the recovery, the determination of net pay does not necessarily follow the
traditional petrophysical workflow developed for conventional accumulations. For instance, in
tight oil and gas plays, net pay is back-calculated by combining data from the stimulated reservoir
volume and the production performance response from the well. In other words, the net pay is
estimated after the well has been on production.
Stimulated reservoir volume data refer to parameters that describe the producing volume such
as fracture density, fracture conductivity, cluster spacing or completion design, parameters
affected by geologic conditions, reservoir permeability, or fracture design conditions. In practice,
the stimulated reservoir volume data and the production response, and not the petrophysical
parameters, are used to estimate net pay in unconventional plays. The reader is referred to Chapter
10—Unconventional Resources Estimation, herein, for more details.
5.5 Core Analysis

Core data are integral in the assessment of resources. Actual samples provide hard data with which
to characterize the reservoir(s) in question, by means as diverse as defining the rock wettability,
quantifying the relative permeability of the rocks to fluid flow, assessing the continuity of the
reservoir, modeling the saturation profile(s) of the reservoir, identifying transition zones, etc. An
adequate treatment of the value of core data analysis in the evaluation of resources is beyond the
scope of this current work; however, some key topics are reviewed that will be useful.
5.5.1 How Representative Is the Core Sample? As with any other data acquired from the
subsurface, core samples (such as whole core, sidewall cores, even cuttings) are subject to
uncertainties that may invalidate their use in proper characterization and mislead the analyst(s).
For example, it is not uncommon in thick reservoirs to recover whole core only from the poorer-
quality sections when higher-quality sections may be too friable or fractured to be recovered.
Utilizing only this data set will penalize the ensuing geological (and, consequently, simulation)
model when the higher-quality rock cannot be characterized. A qualified reserves evaluator, if they
recognize this situation, may consider the models to be conservative and categorize the resources
accordingly. Similarly, intervals selected for sidewall cores (plugs) may favor the better-quality
rock, and if the analytical results are applied without adjustment to lower-quality intervals, then
this would tend to exaggerate the PIIP and recoverable resources.
The core itself may be characterized by its state of preservation as native-state, cleaned, or
restored-state. Native-state core represents unaltered wettability conditions, while cleaned core has
been swept with solvents to remove all organic material (thereby altering its state to water-wet or
strongly water-wet). Restored-state core has also been cleaned but then resaturated with brine
(preferably matching the reservoir salinity) and then flushed with reservoir oil before being aged
at reservoir temperature, commonly for over a month. Core may be cleaned for porosity and
permeability measurements and then restored for relative permeability studies; if not restored, the
core cannot be used for relative permeability studies.
Core prepared for waterflood studies has the potential to provide misleading outcomes.
Coreflood tests must be certain to use water of the same composition, as well as salinity, as the
110
proposed injected water. For example, switching from a freshwater to a brine-water source
normally would be advised if the sandstone proposed for waterflooding contains clays or salts
(such as halite). However, if the reservoir has a high feldspathic content, a sodium chloride (NaCl)–
based water could cause ionic dissociation, whereby the sodium and potassium ions are chemically
exchanged in a process causing salt dissolution, whereas the original intention by using “brine”
was to avoid this process. Likewise, coreflood tests are typically set up in the laboratory for
forward (one direction) throughput of water volumes. Reversal of flow direction has a vital purpose
in identifying any predisposition toward “deflocculation,” in which the reversal causes migrated
(but stabilized) fines to dislodge within the core and cause plugging as flow ensues in the opposite
direction. In waterflooding operations, this has an impact on recovery when pressure gradients
induced by injection wells interfere and can result in unintended flow baffles.
For reservoir simulation of large areas or full-field models, the core coverage is also a concern.
For example, core recovered from wells situated at the crest of a structure are not expected to be
representative of the reservoir at its structural flanks. Likewise, core recovery only from wells
from, for example, the north end of a full-field model may not be representative of the reservoir
quality at the southern extreme.
Aspects such as these, and others not mentioned for lack of space, highlight uncertainties that
the qualified reserves evaluator should be aware of when categorizing the results of recoveries
estimated through the utilization of core data.
5.5.2 Capillary Pressure. Core data analysis frequently depicts the rock fabric as being composed
of a bundle of tubes of different diameters. This is a rudimentary but easy-to-grasp description of
the capillary system in a reservoir. The pressure exerted within each tube is one of the forces
influencing fluid flow and is referred to as capillary pressure. This pressure must be overcome by
a fluid in order to permit its entry into the pore space of the capillary tube; the pressure at which
fluid may enter the system is known as the displacement pressure. The displacement pressure
required for a capillary system of large pores is much less than that for a system of small pores.
This force (along with gravity forces) influences the rise of fluids within each capillary tube. To
properly characterize a reservoir and its ability to permit the flow of fluids, especially when
multiple fluids are in contact with each other, capillary pressure measurements must be taken in
the core laboratory. The capillary pressure data can be used to describe the fluid saturation
distribution versus height relative to the depth at which the particular capillary system is 100%
water saturated. (This depth may be called the free water level, or the depth corresponding to zero
capillary pressure in a water-wet system.)
Capillary pressure data may be generated in the laboratory most commonly by two methods:
mercury injection or centrifuge. Mercury injection capillary pressure data must be converted from
a mercury-air system (under which the data are obtained) to an oil-water system; there are several
caveats inherent in this approach (see, for example, Masalmeh and Jing 2007). The centrifuge
method spins the core at various speeds and measures the stabilized saturation at each speed, from
which a pressure-saturation relationship may be derived. Notably, this method is capable of
reaching a lower residual oil saturation than conventional coreflooding because it incorporates a
gravity force effect; the qualified reserves evaluator must recognize this effect and consider how
gravity forces are likely to influence fluid displacement in the reservoir in question when
estimating recovery efficiency (see also Chapter 4—Assessment of Petroleum Resources using
Deterministic Procedures herein).
111
Saturation-height relationships should then be developed for each representative capillary
system. To complement this, relative permeability data should also be prepared for each capillary
system. This is particularly important in transition zones, i.e., zones in which oil and water both
exist and from which both may flow (the vertical distance over which the water saturation ranges
from 100% to the irreducible saturation). An example of a capillary pressure relationship converted
to a saturation-height profile (top), with a relative permeability relationship (bottom) as measured
with a conventional coreflood, is shown in Fig. 5.3 (Fanchi et al. 2002).
Fig. 5.3—Capillary pressure (top) and relative permeability (bottom) relationship (Fanchi et al. 2002). [Note: capillary
pressure, relative permeability to oil (kro) and relative permeability to water (krw) plotted as functions of water saturation
(Sw) as a fraction of the pore volume, PV.]
The validity of a model prepared with saturation-height functions tied to relative

permeabilities necessitates acquiring core data representative of the productive or potentially
productive reservoir. The use of analog data (see Chapter 4—Assessment of Petroleum Resources
using Deterministic Procedures herein) introduces uncertainties that the qualified reserves
evaluator needs to be aware of when classifying and categorizing model-predicted recoveries.
Further, saturation-height functions should be compared against calculated log interpreted
saturations to explain any discrepancies.
Other means of populating the saturation within a reservoir column, such as the Leverett-J
function, exist but are beyond the scope of this chapter (Leverett 1941; Hassker et al. 1944; Tiab
and Donaldson 2004).
5.5.3 Core Wettability. Wettability is a key concern in planning the development of an oil
reservoir. Wettability is “the term used to describe the relative adhesion of two fluids to a solid
surface” (Tiab and Donaldson 2004). These fluids are treated as immiscible, such that their
characteristics may be measured separately. A widely used rule of thumb in defining rock
wettability, in terms of the reservoir being oil-wet, water-wet, or mixed-wet, hinges on relative
permeability data. One of the most easily identifiable characteristics that may be used to identify
wettability is the water saturation cross-over point of the oil and water relative permeability curves.
Craig (1971) stated that if the oil and water relative permeability curves cross (i.e., have the
same value) beyond a 50% water saturation, then the rock from which the core sample was taken
112
is water-wet (Fig. 5.4a). If the curves cross below 50% water saturation, then the rock is oil-wet
(Fig. 5.4b). How strongly wet the rock would be in terms of water-wet or oil-wet would depend
on the distance of the cross-over from the 50% mark. Granted, this is a rule of thumb, and the
mixed-wet possibility is not satisfactorily addressed; however, Craig also suggested other rules of
thumb to help in this situation (Table 5.2). The reader is referred to Craig (1971) for further detail.
Warner (2015) cautioned, however, that most oil reservoirs should not be described as being water
wet or oil wet, but rather mixed wet with an appropriate scale, and Craig’s rules of thumb do not
consider rock heterogeneity.
Fig. 5.4—Relative permeability assessment of wettability (Craig 1971).
Table 5.2—Rules of thumb for wettability assessment (Craig 1971).
Wettability may also be determined in the laboratory through the use of the Amott wettability
index, the US Bureau of Mines (USBM) method, measurement of the contact angle of the fluid-
solid interface, and other techniques (Tiab and Donaldson 2004). The scale to which Warner
(2015) refers for wettability characterization is the Amott-Harvey scale (+1 for water-wet, −1 for
oil wet) as described in Boneau and Clampitt (1977).
It is crucial to determine the wetting phase from a wettability analysis. If an oil reservoir is
water-wet, then the relative permeability measured when the core is flushed with water represents
an imbibition process (i.e., an increasing wetting phase saturation). Conversely, if the oil reservoir
is oil-wet, then the resulting relative permeability measured during core flushing with water
113
reflects a drainage process (i.e., a decreasing wetting phase saturation). If an oil reservoir is subject
to capillary hysteresis, in which there is a departure in the oil relative permeability relationship
with increasing water saturation between drainage and imbibition, then proper modeling of the
oil/water flow characteristics (and projection of the oil recovery with the least uncertainty)
necessitates application of appropriate imbibition and drainage relative permeability (and capillary
pressure) data in the simulation. [It is not uncommon for an oil reservoir to undergo multiple cycles
of drainage and imbibition, resulting in conditions including transition zones and residual (paleo)
oil zones, further complicating modeling. This area is beyond the scope of this chapter.]
Oil recovery is affected by wettability. Generally, under conditions of oil displacement by
water injection, oil recovery is lower in water-wet rock. In this instance, water adheres to the pore
surface while oil occupies the larger pore space and pore connections. As water moves through,
oil may become trapped in larger pores (a process known as “snap-off”), and the continuous phase
of oil is broken. On the other hand, oil-wet formations have oil adhering to the pore surface. The
movement of water through the system is less likely to break the oil continuity. The net result of
these situations, as demonstrated in core analysis, is that the residual oil saturation in oil/mixed-
wet reservoirs is typically lower than that for water-wet reservoirs; i.e., there is a higher oil
recovery in oil/mixed-wet reservoirs (all else being equal).
5.5.4 Rock Typing. “Rock typing” refers to the process of identifying reservoir rocks with a
characteristic set of petrophysical and flow properties. Generally, there are two ways to develop
rock typing for a reservoir: depositional rock typing and petrophysical rock typing. Depositional
rock typing deals with sedimentary deposition and the associated lithofacies, and it is dependent
on identified depositional environments (such as deltaic, alluvial, glacial, tidal, etc.). This approach
does not typically consider postdepositional activities, such as diagenesis or fracturing, or complex
pore systems, particularly in carbonates. Generally, depositional rock typing is a larger-scale
assessment, while petrophysical rock typing is focused at the log and core scales. As such,
petrophysical rock typing looks at the pore types and structures (e.g., pore throat sizes) as well as
the relationship between porosity and permeability. Petrophysical rock typing may be utilized for
dynamic (flow-based) analysis, such as through the use of “flow zone indicators” (Amaefule et al.
1993) and the Winland R35 method (Kolodzie 1980), which is dependent upon mercury injection
capillary pressure data to characterize the pore systems.
The population of rock types or facies has a major impact on the results and their associated
uncertainty analysis from dynamic simulation. Petrophysical rock typing is instrumental in
building geological models that are typically imported into dynamic (reservoir) simulations, as the
facies described in the process represent flow characteristics with their own dedicated models (e.g.,
saturation-height, porosity, relative permeability, fluid properties, etc.). Again, this topic cannot
be adequately covered in the scope of this chapter, but the characterization of the various facies is
one of the most important contributions of core analysis.
5.5.5 Reservoir Continuity. Even when a field has had significant exploitation, i.e., there have
been many wells drilled, a reservoir of interest may appear to be correlated by log analysis, but
production performance data may suggest otherwise. For example, structurally high wells may
produce water while downdip wells produce water-free resources from the “same” reservoir. The
integration of core and log data with other analyses (such as fluid compositions) can shed light on
why a reservoir that can be correlated between wells may be disconnected depositionally.
Sequence stratigraphy is a tool that can be used to explain this apparent disconnect.
114
Sequence stratigraphy may be described as “a branch of sedimentary stratigraphy … which
deals with the order, or sequence, in which depositionally related stratal successions … were laid
down in the available space” (Society for Sedimentary Geology 2021). It is a tool used to “interpret
the depositional origin and predict the heterogeneity, extent and character of the lithofacies”
(Society for Sedimentary Geology 2021). Sequence stratigraphy seeks to identify reservoir zones
that are stratigraphically connected through a combination of such means as time of deposition,
chemical composition, magnetic properties, fossil evidence, etc. Physical core data are, therefore,
crucial in establishing the connectedness of the reservoir(s) and subsequently extending the
individual well responses to seismic signatures.
Fig. 5.5a shows an example of the traditional (lithostratigraphic) log correlation (from Van
Wagoner et al. 1990), constructed using the top of the shallow marine sandstone in each of the
four wells. The top of the reservoir was picked based on either the spontaneous potential or gamma
ray data and the resistivity deflections. This interpretation shows the sand to be continuous across
the four wells.
Using sequence stratigraphic (chronostratigraphic) analysis (Fig. 5.5b), however, the
sandstone is not continuous at all. Van Wagoner et al. (1990, p. 22) noted that, if only the
lithostratigraphic correlation had been performed, “the continuity of the reservoir is exaggerated,
genetically different sandstones are linked together, and shallow-marine sandstone reservoirs
change facies up dip into marine shales and mudstones.”
Reservoir continuity is always a major concern for the qualified reserves evaluator due to its
influence on, among other things, project development costs, volumetric assessment of PIIP,
recovery efficiency estimates, etc. These factors represent uncertainties that will enter into the
resource classification and categorization.
Fig. 5.5—Sequence stratigraphy application (Van Wagoner et al. 1990).
115
5.5.6 Quantifying Residual Oil Saturation. Knowing the hydrocarbon target facilitates the
proper assessment of project development plans and accompanying economic requirements. For
example, the well-known equation for calculating maximum possible waterflood displacement
efficiency is:
(1−𝑆𝑆𝑤𝑤𝑤𝑤 )− 𝑆𝑆𝑜𝑜𝑜𝑜
𝑅𝑅𝑅𝑅 = , ............................................................................................. (5.1)
(1− 𝑆𝑆𝑤𝑤𝑤𝑤 )
where
RE = displacement efficiency, fraction of PIIP
Swc = connate water saturation, fraction
Sor = residual oil saturation, fraction.
As expressed, this is the oil saturation that is residual to a water-displacement mechanism.

(Recovery efficiencies are discussed in more detail in Chapter 4—Assessment of Petroleum
Resources using Deterministic Procedures, herein.) Further, the recovery efficiency calculated in
this manner is inclusive of the primary recovery efficiency; i.e., if the primary recovery is estimated
separately (perhaps by decline curve analysis) to be 15% of PIIP, and the recovery efficiency
calculated using Eq. 5.1 above is 25% of PIIP, then the incremental technical recovery due to the
secondary process of waterflooding is 10%. On the other hand, if the primary recovery is already
20%, will the incremental 5% of PIIP be economically justifiable for the installation and operation
of a waterflood?
The residual oil saturation (Sor) is dependent on the displacement process. If miscible carbon
dioxide flooding is implemented, the improved oil recovery process is expected to reduce the Sor below
that of a waterflooding process. (In this case, the calculated recovery efficiency includes all recovery
through the miscible flooding process.) Provided that relative permeability data are available to identify
the residual oil saturation under both waterflooding and miscible displacement processes, the evaluator
may address the upside potential in different categories within the resource class; however, the
development scope and associated expenses will generally result in separate projects.
Residual oil saturation typically is determined in the laboratory by flushing a number of pore
volumes of water (preferably of the same chemical composition as the planned injection water)
through a representative core sample. This procedure is continued until no further oil is recovered
at the core outlet. Higher core waterflooding rates are usually implemented near the end of the
coreflood to overcome the capillary end effects while determining Sor. (Other methods of
estimating Sor include log-inject-log techniques, such as the thermal decay time log and nuclear
magnetic resonance imaging logs. Sponge coring, on the other hand, may give an estimate of the
remaining—rather than the residual, oil saturation, as discussed below.)
The remaining oil saturation reflects the oil saturation left behind when water displacement
has not been optimized. In actual field operations, a well may be drilled in an area believed to have
been swept by an existing waterflooding operation, and a sponge core may have been recovered
for the subject formation. The oil saturation determined from the sponge core will not necessarily
be the residual oil saturation but rather the remaining oil saturation Sorem, which is normally taken
to be higher than the Sor to water displacement. The difference is due to the sweep efficiency of
the waterflooding process.
Example E in the following section illustrates the re-evaluation of Sor for reserves booking
revision (assuming all other criteria for reserves classification are satisfied) using log interpretation
116
and tracer testing. An alternative approach is to take the oil-saturated core and centrifuge it to
displace the oil. This technique usually results in the lowest Sor and, in the process, mimics the
added effect of gravity drainage. For reservoirs in which gravity drainage is expected to be
negligible, the centrifuge Sor will be lower than the value expected to be achieved in actual practice.
5.6 Examples
5.6.1 Example A: Basic Logging Suite in Conventional Clastic Reservoirs. Since the mid-
1980s, gamma ray (GR), resistivity, neutron porosity (NPHI), and bulk density (RHOB) logs have
comprised the basic set of tools used to quantify rock and fluid properties and estimate net pay.
Fig. 5.6 shows an example of these data from a high-porosity sandstone reservoir in deepwater
Nigeria. In this example, net gas pay intervals are delineated using neutron/density cross-over and
the operator’s pay cutoff criteria. In the presence of gas or light hydrocarbons, the low electron
density of the pore fluid results in a higher apparent density porosity. Conversely, the apparent
neutron porosity is low due to the lower concentration of hydrogen in the pore fluid. When
presented on consistent lithology scales, the apparent density porosity and the apparent neutron
porosity curves “cross over” (third track from left) in intervals with relatively low clay content.
The magnitude of the cross-over would depend also on shale content, and so the effect should
be viewed qualitatively, and the evaluator should avoid application of absolute porosity unit cross-
over rules of thumb. In contrast, oil- and water-bearing intervals do not exhibit “cross-over.” The
water-bearing interval is delineated from the gas- and oil-bearing intervals based on lower
resistivity values.
Fig. 5.6—Lithofacies-based corrections to density-neutron porosity showing

hydrocarbon intervals from a turbidite sandstone (modified from Spears 2006).
117
5.6.2 Example B: Coupling Basic Logs with Nuclear Magnetic Resonance Data in Carbonate
and Complex Reservoirs. In complex/mixed lithology and/or carbonate reservoirs, nuclear
magnetic resonance (NMR) logs may be coupled with basic log suites to identify pay zones. NMR
logs can provide lithology-independent estimates of porosity and inform fluid characterization.
NMR data provide information about the pore size distribution and quantities and properties of the
contained fluids, providing a means to identify mobile water in a prospective reservoir interval.
Fig. 5.7 shows a Middle East heavy oil carbonate example where NMR is used to delineate
pay intervals that will produce only oil, pay intervals that will produce oil and water, and non-pay
intervals that will produce only water. The technique described in the paper (Liu et al. 2013)
utilizes a variable, as opposed to fixed, set of pay cutoff criteria relying on NMR discrimination
of bound water and movable water volumes. (Note the integration of log interpretation with
wireline sampling.)
Fig. 5.7—Movable water controlling the hydrocarbon production (modified from Liu et al. 2013), where SP is spontaneous
potential, PEF is photoelectric factor, and BOPD is barrels of oil per day.
5.6.3 Example C: Thinly Laminated, Low-Resistivity Reservoirs. Thinly laminated, low-

resistivity reservoirs comprise sequences of thin interbeds, or laminations, of sandstone and shale
(or mudstone). Because lamination thicknesses are less than the vertical resolution of most
conventional logging tools, the resistivity of hydrocarbon-bearing sandstones in such settings is
118
suppressed by the interbedded shale. This may result in water saturation overestimation and
porosity underestimation in porous intervals due to tool resolution and measurement averaging in
such reservoirs, particularly when pay cutoffs are applied.
Multiple techniques have been published to address this reservoir type, including, but not
limited to, high-resolution core- or log-based image analysis methods and lower-resolution,
conventional log or NMR-based assessments of bulk volume hydrocarbon. One such approach
(Thomas-Steiber method) is used in Fig. 5.8 (Stromberg et al. 2007). The Thomas-Steiber method
identifies laminated and dispersed clay components in shaly sands to estimate the net-to-gross
(NTG) ratio. The box in the figure highlights a section that was not perforated but was
subsequently interpreted as containing net pay. In this example, it is evident that relying solely on
ultraviolet fluorescence (image track in center) to estimate net pay in thinly laminated intervals is
complicated by the presence of residual hydrocarbons in the low-resistivity, non-pay interval at
the bottom of the core. Dynamic data from formation tester tools, production logging tools, or
production tests may be required to confirm moveable hydrocarbons in such intervals. The reader
is encouraged to read Stromberg et al. (2007) for further details.
Fig. 5.8—Identifying net pay zones in laminated reservoirs in Oman (Stromberg et al. 2007).
5.6.4 Example D: Integrating Production Logging Test to Determine Net Pay in Laminated
Sands. Laminated sand-shale sequences are characterized by suppressed resistivity, elevated
gamma-ray values, and increased separation in neutron-density logs that are below the vertical
119
resolution of the log data, which make net pay easy to overlook during well log interpretation.
Combining log and core data with production contribution from the intervals is the key to
quantifying the net pay thickness in this type of reservoir.
In this example (Fig. 5.9), the contribution from laminated sands in the well was quantified
by integrating production logging tool data. The thin beds of sand in the laminated section were
difficult to interpret from basic logs because of limitations in vertical resolution but were clearly
identified from core ultraviolet light photos (showing high oil saturations) and mud log shows,
among other measurements. Though oil is visible, dynamic data were necessary to confirm
producible hydrocarbons across the laminated sequence. Perforated intervals are shown by the
tan markers.
Fig. 5.9—Wireline logging tools response integrated with production logging tool (PLT) data in a laminated sand interval
(perfs indicates perforated intervals).
Production logging tool data indicated three sharp cooling effects (∆ temperature track),
identifying inflow. Gradiomanometer (or density) data were used to identify different fluid phases.
A density curve can be created from the pressure data by comparing flowing and shut-in passes to
indicate fluid entry. The density and capacitance data also show a decrease moving up the wellbore
from brine to oil to light oil/gas condensate. Four spinner passes (i.e., spinner run at four speeds)
highlight flow starting at the base of the lower perforated intervals. By integrating these results,
we estimate that most of the oil inflow (approximately 65%) is coming from the good-quality
reservoir section at the base of the upper perforated interval. However, we also observe flow
contribution (approximately 15% of the total inflow) from the laminated section at the top of the
upper perforated interval. As a result, the production logging tool data confirm production (pay)
from the laminated sections, which may have been ambiguous based only on the open hole logs.
5.6.5 Example E: Reevaluation of Residual Oil Saturation. In a mature oil field with 15 years
of production, studies were performed to reevaluate the residual oil saturation (Sor), which could
120
have a material impact on the estimated recovery under the planned development program.
Initially, a value of 38% was obtained from relative permeability curves in laboratory tests and
used for 2P estimates. However, this Sor may be high, since the tests were performed on plugs of
core with permeability lower than the average reservoir permeability, due to the difficulty in
recovering representative rock samples from the unconsolidated rock.
Three data sources were considered for the studies: wells with open hole logs, in an effort to
define the fluid saturation in a washed-out area; cased-hole logs from a recent well (using a carbon-
oxygen log); and recent tests with a chemical tracer [single well tracer test (SWTT)]. Table 5.3
presents the values of Sor developed from each source:
Sor initially Log Sor values (open/cased hole studies) SWTT Sor
used Average Minimum
38% 33% 17% 23%
Table 5.3—Sor values considered from each source.
The Sor value obtained from the SWTT was considered to be more representative compared
to the values from other studies, so it was decided to adopt this result in the 2P (best estimate)
volume. Furthermore, the minimum of 17% was selected for the high estimate (3P) volume; results
of all three cases as determined through simulation are presented in Fig. 5.10.
Fig. 5.10—Estimated ultimate oil recovery prior to and after tests with chemical tracers.
5.7 Conclusion
Resource assessment has evolved immensely over the last few decades, and numerous
petrophysical analytical processes have been at the forefront of this evolution. These processes
rely fundamentally on two types of parameters: those which describe the static properties of the
rock, fluid(s), and their interaction, such as grain size, porosity, fluid density, capillary pressure,
fluid saturations, wettability, etc., and those which are determined by the dynamic phenomena of
121
hydrocarbon or water production. Examples of such parameters are net pay, irreducible fluid
saturations, mobility, gas-oil ratio, and oil rate.
Net pay is obtained by applying the dynamic process of hydrocarbon extraction to a rock that
contains such hydrocarbon and has the potential to produce. Since the extraction typically is the
consequence of a necessary pressure differential, net pay can only be estimated under given
pressure regimes.
Conventional and advanced wireline logging, mud cuttings, sidewall and whole coring,
production logging tools and pressure tests, and similarity to producing formations are some of the
tools used by petrophysicists to determine net pay.
5.8 Acknowledgments
Important feedback and editorial effort were provided by Charles Vanorsdale and Danilo
Bandiziol.
5.9 References
Amaefule, J. O., Altunbay, M., Tiab, D. et al. 1993. Enhanced Reservoir Description: Using
Core and Log Data to Identify Hydraulic (Flow) Units and Predict Permeability in Uncored
Intervals/Wells. Paper presented at SPE 68th Annual Technical Conference and Exhibition,
Houston, Texas, USA, 3–6 October. SPE-26436-MS.
Archie, G. E. 1950. Introduction to Petrophysics of Reservoir Rocks. Am Assoc Pet Geol Bull 34
(5): 943–961.
Archie, G. E. 1967. Formation Evaluation. In Impact of New Technology on the U.S. Petroleum
Industry 1946–1965, 150. Washington, DC, USA: National Petroleum Council.
Asquith, G. and Krygowski, D. 2004. Basic Well Log Analysis. AAPG Methods in Exploration
Series, No. 16.
Boneau, D. F. and Clampitt, R. L. 1977. A Surfactant System for the Oil-Wet Sandstone
of the North Burbank Unit. J Pet Tech 29 (5): 501–506. SPE-5820-PA.
http://dx.doi.org/10.2118/5820-PA.
Chen, A. and Pagan, R. 2013. Discover a Career: Petrophysics. The Way Ahead 9 (2): 19–21.
SPE-0213-019-TWA.
Craig, F. F. 1971. The Reservoir Engineering Aspects of Waterflooding, Vol. 3. Richardson,
Texas, USA: Monograph Series, Society of Petroleum Engineers.
Fanchi, J. R., Christiansen, R. L. and Heymans, M. J. 2002. Estimating Oil Reserves of Fields
with Oil/Water Transition Zones. SPE Res Eval & Eng 5 (4):311–316. SPE-79210-PA.
Hassker, G. L., Brunner, E., and Deahl, T. J. 1944. The Role of Capillarity in Oil Production.
Trans., AIME 155 (1): 155–174. SPE-944155-G.
Kolodzie, S. 1980. Analysis of Pore Throat Size and Use of the Waxman–Smits Equation to
Determine OOIP in Spindle Field, Colorado. Paper presented at the SPE Annual Technical
Conference and Exhibition, Dallas, Texas, USA, 21–24 September. SPE-9382-MS.
Leverett, M. C. 1941. Capillary Behavior in Porous Solids. Trans., AIME 142 (1): 152–169.
SPE-941152-G.
Liu, C., Kelsch, K., Yepes, O. et al. 2013. A Method to Identify Net Pay Zones in Clastics and
Carbonate Using Movable Water and Uncertainty Level Concepts rather than Fixed Cut-
offs. Paper presented at SPWLA 54th Annual Logging Symposium, New Orleans,
Louisiana, 22–26 June. SPWLA-2013-OO.
122
Masalmeh, S. K. and Jing, X. D. 2007. Improved Characterisation and Modeling of Carbonate
Reservoirs for Predicting Waterflood Performance. Paper presented at the International
Petroleum Technology Conference, Dubai, UAE, 4–6 December. IPTC-11722-MS.
Petroleum Resources Management System (PRMS), Version 1.01. 2018. Richardson, Texas,
USA: Society of Petroleum Engineers. https://www.spe.org/en/industry/reserves/.
Society for Sedimentary Geology (SEPM, formerly Society of Economic Paleontologists and
Mineralogists) 2021. Intro Sequence Stratigraphy. http://www.sepmstrata.org/page.aspx?pageid=15
(accessed 9 September 2021).
Spears, R. W. 2006. Lithofacies-Based Corrections to Density-Neutron Porosity in a High-
Porosity Gas- and Oil-Bearing Turbidite Sandstone Reservoir, Erha Field, OPL 209,
Deepwater Nigeria. Paper presented at the SPWLA 47th Annual Logging Symposium,
Veracruz, Mexico, 4–7 June. SPWLA-2006-O.
Stromberg, S., Nieuwenhuijs, R., Blumhagen, C. et al. 2007. Reservoir Quality, Net-to-Gross,
and Fluid Identification in Laminated Reservoirs from a New Generation of NMR Logging
Tools. Examples from the Gharif Formation, Southern Oman. Paper presented at the
SPWLA 1st Annual Middle East Regional Symposium, Abu Dhabi, UAE, 15–19 April.
SPWLA-MERS-2007-R.
Tiab, D. and Donaldson, E. C. 2004. Petrophysics: Theory and Practice of Measuring Reservoir
Rock and Fluid Transport Properties, second edition. Houston, Texas, USA: Gulf
Professional Publishing.
Van Wagoner, J. C., Mitchum, R. M., Campion, K. M. et al. 1990. Siliciclastic Sequence
Stratigraphy in Well Logs, Cores, and Outcrops, Vol. 7. Tulsa, Oklahoma, USA: Methods
in Exploration Series, American Association of Petroleum Geologists.
Warner, Jr., H. R. The Reservoir Engineering Aspects of Waterflooding, second edition, Vol. 3.
Richardson, Texas, USA: Monograph Series, Society of Petroleum Engineers.
Worthington, P. F. 2011. The Petrophysics of Problematic Reservoirs. J Pet Tech 63 (12): 88–97.
SPE-144688-JPT.
Worthington, P. F. and Cosentino, L. 2003. The Role of Cutoffs in Integrated Reservoir Studies.
Paper presented at the SPE Annual Technical Conference and Exhibition, Denver, Colorado,
USA, 5–8 October. SPE-84387-PA.
123
Chapter 6
Reservoir Simulation
Miles Palke (Chair)
Avi Chakravarty, Ali Albinali, and Charles Vanorsdale
6.1 Introduction
The practice of dynamic reservoir simulation dates back to the first analytical models of the late
1940s and early 1950s, but widespread use did not occur until the advent of commercial simulation
packages in the late 1980s. Detractors of the use of reservoir simulation refer to the process as a
“black box,” but, in the hands of an experienced petroleum engineer working with an experienced
geoscience team, reservoir simulation can materially help to evaluate the uncertainties associated
with investment decisions and consider development planning scenarios. It is a tool that can be
used to improve the understanding of reservoirs and their performance data, facilitate the decision-
making process, aid in the selection (in conjunction with economic analysis) of the optimal
development scenario, assess the range of resource potential, and so on. Nevertheless, there are
limitations to reservoir simulation and the reliability of its results that need to be clearly conveyed
to end users and decision makers.
The objective of this chapter is to focus on the application of reservoir simulation for the
estimation of reserves and resources in alignment with Petroleum Resources Management System
guidelines. The process and methodology of model construction, history matching, and running
predictive cases are touched on, but the detailed techniques that should be applied are better
addressed in other sources in the petroleum engineering literature. It is assumed for the purposes
of reserves and resources estimation that both existing and future development projects have been
established, and reserves and resources are being estimated for those distinct development plans.
6.2 What is Reservoir Simulation

Reservoir modeling or simulation has been described by the Petroleum Resources Management
System (§ 4.1.3.2) as:
“a more rigorous form of material balance analysis. While such modeling can be a reliable
predictor of reservoir behavior under a defined development program, the reliability of
input rock properties, reservoir geometry, relative permeability functions, fluid
properties, and constraints (e.g., wells, facilities, and export) are critical. Predictive
models are most reliable in estimating recoverable quantities when there is sufficient
production history to validate the model through history matching.”
Reservoir simulation is a computational modeling process that attempts to model the behavior
of fluids moving through porous media. Reservoir simulation can be used to emulate flow behavior
in one to three dimensions, and with a wide variety of fluid behaviors usually represented by the
results of pressure-volume-temperature (PVT) experiments. Reservoir simulation models (or
simply simulation models) render a representation of an actual reservoir honoring the laws of
material balance and multiphase flow in porous media and incorporating the effects of changing
reservoir conditions over time and throughout the reservoir. While the central formulations of
reservoir simulation address the flow of multiple phases given the impact of convective and
124
gravitational forces, assuming relatively simple tabular black-oil fluid treatments, more
sophisticated formulations include any number of additional reservoir phenomena, such as
compositional, thermal, or chemical effects. In addition, the utility of reservoir simulation has been
expanded significantly through the addition of options to interpret pressure loss between the
subsurface and surface in tubing, to obtain the simultaneous solution of the reservoir model with
representations of gathering systems, or even to incorporate sophisticated production plant
processes (also known as subsurface-surface reservoir coupling models).
The majority of commercial reservoir simulation code relies on a finite-difference
discretization scheme, but commercial code relying on other formulations such as stream-line
simulation is also available. In general, the guidance provided by this chapter does not depend on
the form of simulation that is used, as we are accepting that these tools provide a reasonable
representation of the physics of flow in a porous media, given the confidence of the input data for
the reservoir and the project.
Reservoir simulation is an integrated form of analysis, not a stand-alone or isolated form of
assessment. Integral components are discussed within several other chapters in this document,
most notably in Chapter 4—Assessment of Petroleum Resources Using Deterministic Procedures
and Chapter 5—Petrophysics. It is a powerful tool because it allows the inclusion and integration
of all relevant reservoir information with details of the development plan. However, as a form of
analysis, it is very dependent on the input data and their certainty, and small changes in the
underlying input can yield substantial differences in the output.
While this chapter focuses on the traditional use of reservoir simulation as a deterministic tool
(see Chapter 4—Assessment of Petroleum Resources using Deterministic Procedures), reservoir
simulation can also be applied in a probabilistic fashion (see Chapter 7—Probabilistic Resources
Estimation). While this extension of the use of reservoir simulation to the probabilistic methodology
is not discussed in detail herein, most of the guidance presented here applies equally well to
probabilistic use of reservoir simulation. Careful attention should be paid to maintain that selected
Proved cases conform to the strict limitations imposed on deterministic Proved Reserves model.
6.3 Use of Simulation in Resource Estimation

A simulation model used to develop an estimate of recoverable resource quantities should
generally include the entire petroleum initially in place (PIIP) that is the target of the development
in question. If an estimate is for the entire development plan for a whole field, then the modeled
reservoir(s) must include the entire PIIP in order to serve as a means to assist in the management
of resources, including all the PIIP from the free-water level up, inclusive of the transition zone,
the main pay zone, and any “tar” or residual oil zone. One of the key challenges to the application
of reservoir simulation to resource estimation is modelling, for the appropriate level of uncertainty,
reservoir connectivity (and hence effective drainage area) and drive mechanisms. To the extent
that identified geological features such as faults or changes in stratigraphy are understood or
anticipated to impact reservoir connectivity, they must be included in the description. However,
often and especially for newer fields, such features and their dynamic impact may not yet be
understood. Therefore, it may be appropriate to include a more conservative description for a “low
case” (1P or 1C) estimation, and a less restrictive, more widely connected description for a “high
case” (3P or 3C) estimation.
As will be discussed in greater detail later, simulation models are calibrated to historical
performance data through the history matching process. When considering the use of a simulation
model in resource estimation, attention must be paid to the maturity of the development of the
125
reservoir or field, with consideration given to the principal development process used to produce
the remaining recoverable resources. For instance, it is clear that a new field in the process of being
developed, where the only dynamic data may be limited to well tests or even repeat formation
tests, must be recognized as immature for reservoir simulation purposes. There will be a significant
amount of uncertainty in the outcome of development scenarios at such an early stage of
development. On the other hand, a field that has been waterflooded to a high water cut can be
treated as mature, unless a new physical process (for instance, some form of tertiary recovery) is
anticipated to be introduced for the purpose of extracting a significant fraction of the remaining
in-place resources.
6.3.1 Immature Reservoirs. As would be expected, immature reservoirs in particular have a

large range of uncertainty. The project may have little or no production history, few wells to
provide data coverage across the field, insufficient samples of rock and fluid property analyses
with which to characterize the reservoir(s) and its fluids, or other limitations on data available
to describe the reservoir(s).
How would we qualify a reservoir as “immature?” When reservoir simulation is viewed in the
context of a material balance approach (the reader is again referred to Chapter 4—Assessment of
Petroleum Resources using Deterministic Procedures), a generally accepted industry rule of
thumb states that a reservoir is considered immature for such an analysis if it has not yet produced
about 10% of its PIIP or if its reservoir pressure has not yet declined by approximately 10%
(assuming volumetric behavior). We likewise may categorize a reservoir as immature for
simulation purposes by applying this material balance analogy.
Another instance of immaturity would be associated with a significant redevelopment or the
introduction of a production mechanism that has previously been untested or has just commenced
in the reservoir, for example, the introduction of a secondary recovery project in the case of a
reservoir at the end of primary depletion, or the introduction of a tertiary project at the end of
secondary recovery. Judgement must be used in such situations to arrive at a determination of the
degree of maturity of the process. Issues to consider when appraising the maturity of such a process
would need to include whether there are pilot projects to include in history matching, or whether
there are analogous accumulations where the new recovery mechanism is more mature. For
instance, if a simulation model of a field with two fault blocks establishes a good history match of
the mature waterflood operations in the reservoir in Block A, although the operator has yet to
initiate waterflooding in Block B, the results from Block A very likely establish the analog for
Block B modeling.
Because an immature reservoir lacks the requisite amount of performance data with which to
properly calibrate the dynamic model, the industry widely recognizes the greater degree of
uncertainty in the predictions of such a dynamic model. In the absence of performance data with
which to adjust and constrain the input geological description, along with other input information
such as relative permeability properties and possibly PVT properties, the output is highly
dependent on the assumptions made during model construction.
Because of these issues, the key use of reservoir simulation for resource estimation in
immature reservoirs should focus on sensitivity analysis, where the ranges of potential outcomes
are assessed by varying the values of key input parameters. Key parameters may be identified
using, for example, tornado diagrams (see Chapter 7—Probabilistic Resources Estimation herein).
Projecting Proved Reserves (as compared to “Proved + Probable” or 2P reserves) for immature
reservoirs requires special attention in most cases. This topic is discussed further in Section 6.6.
126
A special caution in history matching is the reliance on reservoir conditions in history that
may not be representative of the forecast conditions. For example, the history match of wells that
have produced during the above-bubblepoint period in the reservoir may not reflect how the wells
will perform once reservoir pressure falls below the bubblepoint. Data such as three-phase relative
permeabilities and critical gas saturation must be validated (not just measured in the laboratory)
before a reliable forecast can be developed.
6.3.2 Mature Reservoirs. Mature reservoirs are those where a lengthy history of production has
occurred due to the application of the recovery mechanism upon which most of the remaining
recoverable quantities also depend. A mature reservoir has a significant history of performance
data that can be used to calibrate the input of a simulation model. At a minimum, such a history
will include rate (all hydrocarbons and other streams) and pressure data, but it may also include
formation pressure tests, open hole well logs, and cased hole logs taken over the course of
production of the reservoir. The available data may even include time-lapsed seismic data capable
of imaging the movement of fluids over the course of production.
In most cases, the integration of these data into the dynamic model through the history matching
process leads to modifications of the underlying static model description. If diligently incorporated,
such changes lead to a dynamic model that can, within a reasonable tolerance, explain the historical
pressures and fluid production from the reservoir. This does not ensure that the resulting model is
the only dynamic model that could explain the production history, or even that it is a likely
explanation for the dynamic history. Therefore, even in the case of mature reservoirs, it is important
to understand that there is still uncertainty in the future outcomes for the reservoir, and any
simulation-based prediction is only one predicted outcome within a range of possible outcomes. For
this reason, it is always useful to compare simulation results with other approaches.
Although it may be argued that at an advanced state of depletion, reserves may be more easily
quantified using decline curve analysis, there are still many reasons to include simulation models
in the tools used for resource estimation for such a mature reservoir. First, there are many mature
reservoirs that are not in production decline, due to aquifer or gas cap pressure support, capacity
limitations on facilities, sales contracts, export options, or market conditions. Reservoir simulation
is usually an excellent tool for such a scenario.
Furthermore, reservoir simulation is often employed to provide estimates of the recovery for
ongoing development or redevelopment options for which other performance-based
methodologies such as decline curve analysis or material balance analysis may be difficult to adapt,
or where simulation models combined with tools such as nodal analysis are likely to render a more
accurate and physically realistic portrait of future reservoir performance. Examples of such
situations include flood pattern realignment or infill drilling. Another similar situation is a change
in prevailing production constraints such as increasing water handling from a field or increasing
the gas available for injection. Reservoir simulation models provide the option to link wellhead
pressures to gathering and plant inlet points. This renders straightforward the process of predicting
the effects of modifications to facility capacity or the gathering system layout or constraints.
The following is a partial list of scenarios where simulation models of mature reservoirs may
provide more reliable predictions of future performance than other methods:
• Changes in facilities rate constraints
• Changes in constraining surface pressures
• Changes in injection rates or pattern realignment
• Redevelopment drilling campaigns
• Changes in well and completion design
127
• Introduction or alteration of artificial lift
• Production of an undersaturated oil as the reservoir goes below the bubblepoint
6.4 Fundamentals of Simulation Quality Assessment

As noted earlier, there has been a “black box” stigma associated with reservoir simulation due, in
many instances, to unfamiliarity with the inner workings of simulations on the part of the end user
of the simulation results. In contrast, there has been a tendency to accept the results of reservoir
simulation without first achieving a measure of confidence in the underlying geological model,
understanding the level of reliability of dynamic data, or assessing whether an acceptable match
of the model results to that of actual reservoir performance history has been achieved. In many
cases, once a simulation model has been constructed and accepted, users may begin to depend on
a model without appreciation that the accuracy of predictions from the same model varies
depending on what is being predicted and when. For instance, a well history-matched model will,
in general, provide more accurate predictions for the near future than for the distant future.
It is not the intent of this chapter to provide instructional material for simulating fluid flow or
reservoir properties. There are, however, elementary techniques used to assess the quality of a
reservoir model and its fitness for predicting hydrocarbon recovery. While evaluators are likely to
differ in the techniques they apply, there are four primary elements of simulation modeling that
should be examined when considering the reliability of reservoir simulation results used for any
purpose, including resource estimation.
These include:
(1) the construction of the underlying static model;
(2) the integration of dynamic data;
(3) the history matching process and results; and
(4) the construction and validity of predictions.
These four areas will be treated individually in the following sections.
6.4.1 Static Model Construction. The building blocks of a good reservoir model are the
underlying geological characterizations. It is important, however, that the geoscientists involved
have a clear understanding of the required output from the simulation model. For example, the
ways in which geological rock types and flow units are modeled will impact results regarding
assessment of recovery efficiency. If a reservoir is modeled using flow units, and a flow unit
consists of several rock types, while certain rock types are deemed hydrocarbon-bearing but
unproductive, it may be difficult or impossible to remove those rock types from the flow units in
the model to properly calculate a recovery factor using the PIIP.
The geological description consists of two primary constituent parts. The first of these is the
structural framework, which describes the physical container holding the PIIP, the physical limits
of the petroleum accumulation, compartmentalization, and (where appropriate) the underlying
aquifer. The second of these parts is the description of the rock within the structure, the
characterization of the formation, ultimately resulting in the distribution of attributes of reservoir
rock that are directly utilized by the reservoir simulation software, most notably distributions of
facies or rock type, porosity, permeability, and (depending on the preferences of the team
constructing the model) the net-to-gross ratio.
Structural frameworks should adhere to, and honor, several important pieces of data, starting
with the correlation scheme adopted by the geoscientists. The tops and bases of key zones and
subzones should be included, and vertical zonation should be adequate to capture the heterogeneity
of the reservoir, particularly heterogeneity impacting fluid flow and recovery. This likely depends
128
upon the recovery mechanism responsible for petroleum recovery from the subsurface. For
instance, the same heterogeneity could have a significant impact on the results of the waterflood
of an oil reservoir but only minimal impact on the results of primary depletion of a gas reservoir.
The structural framework also includes the areal gridding of the reservoir description. Grids
should have adequately fine definition to allow for the capture of lateral variations in properties
between wells. The grid block size is a compromise between accuracy and the computational
resources and time available to conduct a study. It is essential to capture the heterogeneity of the
geology in addition to the rock and fluid description in the model while allowing manageable
computation. The selection of fine grids requires detailed information and sufficient understanding
of the reservoir to guide the process of populating the grid properties and combining various data
from different scales.
Descriptions lacking sufficient definition, either vertically or laterally, will tend to allow
unjustifiable communication, as the averaging of the formation characterization through upscaling
will tend to increase connectivity. This is due to averaging finer-scale data values into coarser grid
blocks inappropriately, which is likely to occur for properties where averaging should be
associated with transmissivity (mostly permeability) rather than properties where the average
should be more strongly tied to storage capacity (porosity). Averaging the properties of porous
media can lead to severe errors if not practiced with insight on the process and the different
mathematical approaches.
In practical terms, most modeling practitioners are going to be limited to those upscaling
options that are supported by the geomodeling or simulation software suite(s) used by their
organization. These software suites usually provide detailed explanations of the methodologies,
along with recommendations about which methods to apply in which circumstances, and how to
review the results for reasonableness and reliability, complete with references to the original
source materials in the upscaling literature.
The areal spacing of the grid should be adequate to address the nature of the development plans
under consideration. Grid cells that are 100 m by 100 m may be adequate for some models of wells
on very large spacing but will likely prove inadequate for a model of a heavy oil reservoir developed
on very tight well spacing. Additionally, the structural model should capture, to the extent practical,
structural offset due to faulting, particularly where faults are of significant magnitude and are known
or may prove to meaningfully impact communication across the reservoir.
The petroleum engineering literature historically has recognized (Mattax and Dalton 1990)
that the utility of the gridding scheme can be enhanced through orientation of the grid with the
direction of the principal permeability tensor. Anisotropy in the permeability influences flow rates
and pressure depletion and propagation in the porous media. Having this information analyzed and
understood before construction of the geological model grid can improve the quality of the static
model and induce better estimation of pressure gradients and fluid saturations.
A grid may undergo modification to be used for specific studies or to address operational
conditions. For example, the size of the grid blocks might be reduced or increased. A very common
reservoir simulation technique is local grid refinement, which is widely supported in commercially
available simulation software. Local grid refinement can be used to replicate induced hydraulic
fractures, capture the transient effects of a low-permeability environment, or study the impact of
condensate drop-out near the wellbore for gas-condensate reservoirs. On the other hand, it is
usually possible to coarsen the grid blocks by increasing their dimensions in order to reduce the
computational load. This might be done in areas of low dynamic activity, for example, away from
129
the fluid front or outside the area of interest (such as in water-filled grid cells in an aquifer, but
relatively distant from the petroleum accumulation).
The second constituent part of the static model is the distribution of rock properties imported
into the simulation model. At a minimum, reservoir simulation software requires the input of
porosity, permeability, and saturation arrays. Most modern models also include the import of other
parameters such as rock type or facies. As noted earlier, vertical and areal grid resolution should
be of adequate refinement to capture important heterogeneities in input arrays that may affect flow.
Two important aspects of reservoir characterization that should always be reviewed when
considering the distribution of properties in the static model are (1) the degree to which the
properties in grid cells match the accepted petrophysical interpretation in the grid cells penetrated
by wells, and (2) the method by which those properties are populated between the wells.
In general, while the simulation model will not have the same vertical resolution as a well log,
the statistics of the important parameters such as porosity should be matched within a reasonable
tolerance by the grid cells along the paths of the wells in the static model. The grid cell properties
in zones tested or completed should be closely scrutinized. Zones indicating high flow intervals,
barriers, tar, etc., should be accounted for and reflected in the properties of these zones. Further,
data collected from vertical, deviated, and horizontal wells should be incorporated when feasible
to be used as control points in order to enhance the modeling results. In an age of widely available
three-dimensional geological description tools, proper workflows should result in upscaled grids
that agree with petrophysical interpretations within a reasonable tolerance with appropriate
upscaling. Likewise, the statistics and variograms for the different three-dimensional grid
properties must be examined to ensure reasonable consistency between the log (and/or core) data,
the fine-scale static geological model, and the final dynamic model.
When populating properties between (or beyond) the existing well control points, the
distribution of properties must appear to be geologically defendable. If the distribution of values
of petrophysical properties away from well control is significantly different from the distribution
from grid cells intersected by existing wells (control points) and averaged into those cells,
explanations of this behavior must be sought.
It is not uncommon to review the distribution of porosity or permeability along a vertical plane
of grid cells and find that the values of these properties become distinctly higher or distinctly lower
between (or beyond) the cells where actual well control data are available. This phenomenon may
be due to the misapplication of a methodology for the spatial distributions of properties.
Fig. 6.1 illustrates a scenario where a subzone in a dynamic model has been populated with rock
capable of producing petroleum, despite the lack of pay in the same subzone observed in the well
log. In this case, the statistics for the distribution of porosity for this subzone in the dynamic model
are not representative of the porosity distribution observed in the well logs, likely indicative of an
improperly implemented geostatistical distribution methodology.
The static properties must reasonably honor the acquired data and laboratory measurements.
Grid cell porosity or permeability values should not exceed observed data unless justified. If a
model is set up to estimate 1P or Proved Reserves, as discussed in Section 6.6.1, any improvement
in porosity, permeability, or hydrocarbon saturation between or beyond the control points will have
to be supported by the reasonable certainty standard.
130
A – Static model property distribution
results with pay immediately surrounding
an existing well with no pay.
B – Static model property

distribution results with pay
immediately surrounding an
existing well with no pay.
Fig. 6.1—A common issue with rock properties between wells not matching observations at wells illustrated with a display of
porosity in dynamic model grid cells. Clear cells at the well are not being assigned pay (and have no pore space).
Fluid contacts and surfaces are also generated in the process of constructing the static model.
Oil-water contacts, gas-water contacts, lowest-known hydrocarbon, free-water level, and transition
zones control the volumetrics of the model. Different contact depths can be set in different areas of
a field, for instance, in different zones or different fault blocks. In general, some degree of pressure
isolation between regions of varying contact depths is required for the regions to be in a state of
quiescence or pressure equilibrium. A good practice is to run a simulation model through a
quiescence checking case. The model is run from initial conditions with no production, and all wells
are shut-in for a prolonged period of time, typically years. (The wells should also not allow crossflow
between grid cells in this static condition.) If the model initialization is appropriate, fluid saturations
and pressures should not change by more than trivial amounts. If this is not the case, and either
pressures are changing significantly or fluids move around with no production, it indicates that there
131
is likely an issue with the model initialization. Models with multiple fault blocks and multiple
intervals with varying initial contacts are prone to have unexpected communication due to
juxtaposed grid cells from different intervals across fault planes. This can generally be dealt with
through sealing faults when clearly required to maintain initial pressure equilibrium.
6.4.2 Dynamic Data Integration. Aside from the initial static geological model, the simulation
model must also incorporate various types of dynamic data. Within dynamic data in this chapter,
we are including the description of fluid properties (PVT), the descriptions of rock-fluid
interactions (through special core analysis), pressure transient information, and the locations of
wells and description of their completions, as well as the historical production or injection
associated with those wells.
PVT properties are a fundamental aspect of any simulation model. Most simulation models
rely on the traditional “black oil” tabular PVT treatment, but advances in computer processing
power have led to an increase in the prevalence of the more sophisticated compositional PVT
treatment, where equations of state are used to calculate fluid properties considering variations in
composition, pressure, and temperature. In general, the traditional black oil treatment can address
most reservoir problems, but compositional treatment should be employed if there are processes
in play that may require it, such as volatile oil or retrograde condensate reservoirs, gas cycling in
condensate reservoirs, miscible gas injection, chemical enhanced oil recovery, or injected gas
where the composition is significantly different from that of native gas.
Regardless of treatment, the input should be checked to make sure that it matches
experimental data (laboratory reports) or that differences are explained. It should be confirmed
that an equation of state can reasonably reproduce laboratory experiment results. Simulation output
to the grid cells should be examined to confirm that the calculations of the simulation code arrive
at the expected results. For instance, outputting an array of fluid formation volume factors at initial
conditions can help to diagnose a mistake in the implementation of the PVT treatment in the model.
Furthermore, the movement and mobility of the different fluid phases are governed by the relative
permeability input combined with the PVT input.
The user (whether the modeler or a qualified reserves evaluator) of a simulation model should
review the treatment of capillary pressure data in the model to determine if they appear sound, and
if they are a reasonable match to laboratory measurements. Distribution of the water saturation by
a saturation-height function and interaction of the different fluids are controlled by the method of
implementing the capillary pressure and petrophysical data. The output saturations in the grid cells
along the path of wells with logs should be reviewed and agree reasonably with those logs.
Pressure transient analyses that indicate changes in completion efficiency, effective
permeability, and/or skin factor with producing time should be used as input when appropriate and
of useful quality.
Once these types of data have been incorporated, it is important to check that the dynamic
model PIIP and the hydrocarbon pore volume match the same values from the initial static model
within a reasonably narrow tolerance. Variance between these numbers beyond a reasonable
tolerance should be explained, and its potential influence on resource estimates should be given
consideration. “Reasonable tolerance” may be defined on a company-by-company basis (discussed
further in Section 6.4.3 below), but it is not unreasonable to expect the hydrocarbon pore volume
from the two models to be within ±5%. The variance between the total pore volume of the dynamic
and static models should generally be very small (assuming consistent application or
132
nonapplication of pay cutoffs), as it is not impacted by differences in PVT or saturation-height
function modeling between the static and initial dynamic models.
Additional dynamic data that require integration within a simulation model include well data
specifying both the trajectory of wells and a description of their completions, as well as the
available history of production from and injection into the wells. Important aspects of this data set
include the diameter of the well, the range of depths of perforations, the presence or absence of
stimulation treatments, and the use of pertinent tubing tables to capture pressure drop from the
subsurface to the surface. It should be recognized that wells change over time and are frequently
recompleted, plugged back, and/or stimulated over the course of their life cycles. Such changes in
wells should be incorporated into the model at the appropriate times.
6.4.3 History Matching. The robustness of any predictions from a simulation model is directly
related to the quality of the history match. On most occasions, a newly constructed simulation
model will not adequately explain the historical performance of a well or reservoir without
further adjustment.
The process of history matching is one of modifying reservoir simulation inputs in an attempt
to mimic the historical pressure depletion and the physics of fluid flow within the model. For
example, for an oil reservoir, the produced oil, water, and gas rates from the wells are input into
the simulation model as historical rates. Reservoir simulators cannot generally be forced to
withdraw the exact observed flow rate of more than one phase (in this case, phases generally
include oil, water, gas, total liquid) for any timestep. When one such phase is specified, the other
phases are the result of producing the observed rate of the specified phase. The simulator output
will be (usually monthly) oil, water, and gas rates, bottomhole pressure, and, if tubing tables are
provided, wellhead pressure. These outputs are then compared to actual measurements.
Normally, for a newly constructed model, these quantities will not match historical
observations satisfactorily or impart confidence to the evaluator at the initial stages of dynamic
modeling. History matching starts with the desire to improve the quality of the match to historical
data. As described below, it is vital to constrain changes made to the initial input values so that
they are supportable within available geological and engineering data. The resulting model should
be considered capable of reasonably explaining the actual performance of the reservoir.
There are several limitations to this process. First, any single history match is nonunique.
There are usually numerous alternative sets of input parameters that would arrive at equally good
matches of the actual observations. While the matches of historical data might be equivalent,
this does not ensure that predictions from the different history matched cases will be equivalent;
indeed, the predictions could be quite different. Second, it is usually possible to make history
match modifications to the underlying description that are unrealistic, and in some cases even
physically impossible. While unrealistic input may still explain historical performance, it is not
likely to produce reliable estimates of future recovery. Third, reservoir simulation modeling, as
compared to a tool such as decline curve analysis, tends to be very data intensive, requiring a
great deal of quality control. It is easy to make mistakes in the data import process, which may
go undiagnosed.
Companies may have internal guidelines constituting “good,” “acceptable,” and “poor”
matches to actual conditions. Time-variable parameters typically used to assess the quality of a
match at the field and well levels with time include:
• Static bottomhole pressure
• Fluid production rates (oil, nonassociated gas, water)
133
• Cumulative fluid production by phase
• Secondary ratios (gas-oil ratio, condensate-gas ratio, water-oil ratio, or water cut)
• Fluid injection rates (if applicable) and cumulative injection volumes
• Fluid breakthrough (water, free gas, injected CO2, etc.) time and trends
It is important to repeat that a history match is not usually unique, as there are normally
different combinations of reservoir parameters that result in a match to historical data equally well,
but the underlying fundamentals could be materially different.
Matching of the major producing phase in a model is not a reliable indicator of a good overall
match because the major phase production history will usually be a primary input to the model. It
is important to recognize the production parameters that have been input to the simulator before
making a judgment call on the quality of the match. Likewise, history matching of cumulative
quantities of a produced phase are inadequate because trends in production tend to be obscured in
cumulative data. Therefore, it is important to attempt to match all fluid rates and not just the rates
used to constrain the wells during matching. In Fig. 6.2, the history match of rate vs. time was
performed using total liquid rate, and it shows a good match, as expected. However, when we look
at the oil rate, the match usually shows some discrepancy with actual data. Further, the gas and
water rates should also be checked against actual rates.
Fig. 6.2 History match of model total liquid and oil rates to actuals, where BPD is barrels per day.
Secondary phase ratios and well pressures (bottomhole or surface) are also important
parameters to attempt to match. If good data from new wells is available after the start of
production, such as repeat formation test data or well logs capable of monitoring fluid sweep or
the movement of contacts, then this data can also provide insight into flow patterns to incorporate
in the matching process. When applicable, a good match of the breakthrough timing of free gas or
water in the wells helps to validate the accuracy of the dynamic model. Fig. 6.3 shows a common
suite of routine history matching plots (actual data are represented by dots).
134
Fig. 6.3—Common suite of history matching plots, where bopd indicates barrels of oil per day, and bwipd indicates barrels
of water injected per day.
A factor of equal importance to the quality of the history match is whether the modifications
used to create the match are sensible from both geological and engineering perspectives. If
possible, it is preferred to rely on parameter values derived from observation rather than history
matching modification (Lee and Sidle 2011). As mentioned earlier, commercial simulation code
allows great latitude in the values applied as input, and it is relatively easy to apply values that do
not make physical sense. Good history matching practice depends on the following:
• Varying global or “big picture” input before introducing more localized changes
• Varying parameters with uncertain values before varying parameters with more certain
values
• Varying parameters where changes within the reasonable range of values lead to a
meaningful change to simulation output, e.g., a small and otherwise reasonable change
in porosity might change simulation results by the same amount as a large and difficult-
to-defend change in capillary pressure
• Not implementing changes that are not defendable based on the best technical
understanding of the reservoir rock and fluid characterization
A careful review of the match must identify whether any “convenient features” have been
included to obtain the match. Examples of these features include zones or layers of very high
permeability (“super-k” zones), unrealistic vertical to horizontal permeability (kv/kh) ratios, strings
of cells acting as “pipelines” to aquifers or injection wells, “leaky” tar mats, frequent
compartmentalization, disappearing faults, and so on.
When history matching, it is also important to bear in mind uncertainty in the historical
observations. Oilfield data may be questionable, and in many cases is subject to a significant
degree of adjustment before being treated as factual, with the best example of this perhaps being
135
production allocated to individual wells. It is important to question the reliability of historical
observations, and to not overemphasize exact matching of data that is far from exact, especially
when doing so requires the introduction of unrealistic input. Produced phases that are not sold will
usually be measured less accurately than produced fluids that are sold.
As mentioned above, history matching is typically carried out by using the measured fluid
production rates to calculate corresponding pressure and saturation distributions in the reservoir.
During predictions, however, the simulator calculates fluid rates using Darcy’s law for flow to the
wellbore, and tubing performance curves for flow to the surface. The calculated flow rates are
controlled by the wells’ productivity indices and the backpressures imposed on the wells. These
backpressures can be flowing bottomhole pressures or flowing wellhead pressures, which may, in
some cases, be a function of the line pressure. A calibration step is usually required to ensure that
the model, when used in a predictive mode (constrained by pressure rather than flow rates, which
is standard in history matching), reasonably reproduces the recent historical rates.
Reservoir history matching should include the matching of static reservoir pressures observed at
the wells, whenever possible. In some instances, there are few or no static reservoir pressures available
to match, as is frequently the case for unconventional reservoirs. In these cases, other data related to
formation pressures should be used as available. This should always include flowing pressure, whether
surface or bottomhole, measured at wells, but could also include increases or decreases of the gas-oil
ratio or condensate yield. These types of data are related to the pressure out in the formation, but they
are also influenced by wellbore factors such as skin or completion efficiency. While history matching
without static reservoir pressures is possible, it should be recognized that the results are more uncertain
than those provided by matches of static reservoir pressures.
Before using a simulation model for predictions, the reliability of a history matched model is
often checked using a “blind test.” In this process, the most recent (6 months to a year) historical
data is omitted from the history matching process. A well can be run under pressure control (using
average flowing bottomhole pressures or flowing wellhead pressures over the period) or maximum
rate control (oil production rate or total liquid rate) during the blind test. The simulation output in
terms of production rates or pressure is compared with the actual recent historical data and should
reasonably match the data from the blind-test period. A variation of this process is to test the model
using wells that were not part of the history match and evaluate the model response. Generally,
these “blind wells” come on production after the input historical production data cutoff date,
although they may also be existing wells deliberately omitted from history matching (in this case,
the well must remain in the model, and it must produce its associated constraining phase rate in
order to ensure appropriate material balance is achieved). A successful “blind test” should increase
confidence in the reliability of the model for making predictions of future well performance.
6.4.3.1 Material Balance Error. An especially important indicator of the overall accuracy of
computational results is the material balance error. This value indicates deviation of the solved
numerical approximation from the exact solution. Material balance error is calculated per timestep
for each cell of the simulation model. Typically, it is reported as a mass accumulation at the end
of the simulation output. The solution method (fully implicit, adaptive implicit, etc.) along with
the numerical tolerance parameters and the nature of the reservoir control these values. Since
reservoir simulation is a numerical approximation, scenarios with high material balance error
values should be examined closely. Material balance error should generally be far less than 1% of
the initial volume for each phase, and if error values exceeding that magnitude are encountered,
then further analysis of the predictions from the model should be undertaken to be comfortable
they are reliable.
136
6.4.3.2 Three-Dimensional Visualization. Visualization software is a very effective tool for
assessing the simulation output and ensuring consistency. Initial pressure and water saturation can
be displayed using vertical cross sections to verify assignment of the datum pressure and initial
saturations and compositions. Time-lapse maps showing output pressure, fluid saturations, and
component concentrations are useful to understand reservoir changes. Furthermore, water cusping,
tracking of liquid fronts, and gas coning can be analyzed from the simulation results and used as a
verification of the model quality. Details about the concentrations of different components in the
phases can be traced to study fluid changes and phase interaction during the life of the reservoir
and are a very practical way to confirm validity and reliability of the computations.
6.4.4 Validity of the Predictions. Once the model reasonably duplicates the physics of reservoir
fluid flow based on historical performance, the next step is to check the reasonableness and
reliability of predictions.
There are several exercises that serve to help check the reliability of simulation predictions.
A common practice is the “productivity calibration” exercise, which is usually performed prior to
generating prediction scenarios. The task is to run predictions with the wells constrained to
representative recent rates for some duration in the near future and analyze the output flowing well
pressures (with bottomhole or surface pressures, depending on data availability). If the model is
showing higher or lower pressure than observed values, adjustment to the productivity of the wells
might be applied. However, as is the case with all such adjustments, these must be recorded and
justified to model reviewers. The same exercise can be conducted by constraining the wells to
representative recent flowing pressures and then comparing the resulting rates to recent rates.
Typically, in the predictive mode, the simulation model is best held in check by continuing the
actual constraints imposed on the wells by the day-to-day operations of the field. For instance, a
gas field might be constrained by the pressure of the pipeline it produces into, or by the total gas
rate that field compression provides, while a mature oil field may be constrained to fluid rates set
by artificial lift.
First, the results of a status quo/no further action (NFA)/“do nothing” predictive case should
be examined. This is a simulation case where the model is changed from being constrained by
historical rate limits to the prediction stage but without any operational changes to the field such
as adding new wells, introducing pressure support, etc. In such a case, the prior performance trends
should continue. For instance, if gas wells were in rate decline before transitioning to predictive
mode, but production flattens as the wells enter predictive mode, then this likely indicates an issue
with the history match, or with the setup of the status quo case, and it calls into question the
reliability of predictions from the model. Fig. 6.4 illustrates a comparison of the actual history and
the predicted NFA performance under two scenarios, only one of which appears to be reasonable.
Such a status quo/NFA case must itself be carefully constructed. The decline trend from a
large field that has benefited from an ongoing program of drilling and well work will not be
maintained if that activity is simply halted in the NFA case. In an instance such as this, it would
be incorrect to expect the prior trends (which includes activity) to continue while knowing the
investments stopped for the NFA case.
137
Fig. 6.4—No further action case with two forecasts.
Predictive case constraints for the status quo case, along with cases featuring further
development options, should be carefully reviewed. The constraints imposed on the model in
predictive mode should be representations of how wells or facilities will be operated in the future.
However, it is relatively easy to generate results that are not realistic. For instance, wells placed
on a constant liquid rate as their future constraint will produce at that rate until their flowing
bottomhole pressure reaches an extremely low value (typically near 1 atmosphere). Such an
outcome is usually unrealistic. Therefore, it is generally better to constrain wells with realistic
surface pressure limitations during the predictive model runs, when such constraints are
appropriate. Poorly implemented constraints, in particular, constant liquid or phase rate
constraints, are likely to provide unrealistically high estimates of future recovery under many
circumstances.
Fig. 6.5 shows one of the diagnostic plots frequently used in validating predictions. This
prediction of oil rate and water cut (at the reservoir level) shows a “bump” in the oil rate occurring
in year 2020. The source of this uptick should be investigated, and, if due to drilling, it should be
determined whether these wells should be drilled earlier if it would benefit the project net present
value. The water-cut signature should be smooth, and, if there are step changes or aberrations, they
should also be explained.
138
Fig. 6.5—Diagnostic plot of oil rate and water cut vs. time.
Further, Fig. 6.5 shows that, with the oil rate increase in the year 2020, there is a decrease in
the water cut, suggesting that new wells or workovers in 2020 did not result in an incidental
increase in water production. In other words, the new wells came on production at water cuts less
than the average of the other producing wells. On the other hand, if the oil rate had been maintained
but the water cut diminished, one possibility could be the shutting-in of high water-cut wells.
Fig. 6.6 shows the voidage replacement ratio (VRR), also called the injection-withdrawal
ratio, from the output. Typically, the VRR rises as fill-up conditions are approached in the
reservoir. Interpretation of such a graph assists the evaluator in waterflood monitoring and
potentially improving both sweep efficiency and the project economics, by addressing questions
such as:
• Are we injecting too much or withdrawing too little (VRR > 1)?
• Are we injecting too little or withdrawing too much (VRR < 1)?
Similarly, for a waterflood project, the VRR can provide an indication of whether the reservoir
is processing too many (or too few) injected hydrocarbon pore volumes. With too few injected
hydrocarbon pore volumes, the reservoir may not be swept efficiently; with too many, the operator
could be damaging the economics of the project and the recovery. Fig. 6.7 shows an example of
the diagnostic plot with an ending injected hydrocarbon pore volume of 1.575, which suggests
more could be injected than currently modeled. However, the flattening slope of the curve at 1.575
does indicate that the point of diminishing returns is being approached, and the economics of
continuing the flood deserve careful attention.
When an oil reservoir is under active waterdrive or waterflooding operations, a key diagnostic
plot is a semilog oil cut vs. cumulative oil production graph, as shown in Fig. 6.8. A similar plot
of water cut vs. cumulative oil production becomes more difficult to check and extrapolate at high
water-cut values, i.e., in the later stages of the operating life, as the curve is rising, the semilog
scale narrows, and a limiting condition (e.g., water cut) cannot be readily identified. This situation
can be improved by using oil cut or water-oil ratio plots instead. Shifts in the curve (such as
shutting in high water-cut wells, drilling in bypassed oil areas, etc.) must be understood.
139
Fig. 6.6—Diagnostic plot of voidage replacement ratio.
Fig. 6.7—Diagnostic plot of cumulative hydrocarbon pore volumes (HCPV) injected vs. cumulative oil production, where
MSTB is million stock-tank barrels.
140
Fig. 6.8—Diagnostic oil cut vs. cumulative oil production.
6.5 Proved (1P) Reserves Cases

Evaluating Proved Reserves (as compared to 2P or 3P Reserves) for immature reservoirs requires
special attention, regardless of the assessment method used (Palke and Rietz 2001). Notwithstanding
progress made in the application of probabilistic approaches to reservoir simulation, in many cases
only a single, well-calibrated, deterministic simulation case is developed. Such a model is much
more likely to be consistent with a 2P scenario rather than a 1P scenario. This is the typical situation
since simulation models serve many purposes other than reserves or resources estimation.
Simulation model cases can, of course, be developed that could be used directly to estimate
Proved Reserves. However, constructing models that conform to limitations imposed on Proved
Reserves estimates may require a significant amount of effort. For strict adherence to the reasonable
certainty requirement, restrictions are placed on the geological modeling of Proved Reserves. For an
immature reservoir, the entire geostatic model would need to be constructed honoring all constraints
imposed for Proved Reserves volumetric estimates, such as not including any porosity-thickness (φh)
greater than that observed in wells. Generally, this would imply an improvement of the reservoir
quality beyond the actual data; improvement of porosity should not be allowed for Proved Reserves
due to the reasonable certainty requirement. There are exceptions that allow for higher φh away from
wells, such as when the subject reservoir is demonstrated by seismic interpretation to be structurally
higher (e.g., above the oil-water contact), and so net thickness may increase. Provided the porosity
is less than or equal to the porosity found in nearby wells, the thickness (hence φh) increase may be
justifiable. Defendable seismic attribute mapping (see Chapter 3—Seismic Applications herein) may
be used to justify improvements in reservoir quality provided the mapping has been ground-truthed
in the area. Care must be exercised regarding hydrocarbon producibility and reservoir quality beyond
well control such as the volume between flank wells and the structural limit. Other constraints
include allowing no volumes in undrilled, potentially sealing fault blocks, and limiting the depth of
141
contacts to lowest known hydrocarbon depths. Furthermore, the dynamic model would need to be
carefully constructed to not overestimate the reasonably certain strength of the drive mechanism to
improve the recovery process, to make sure that well productivity indices in the simulation model
meet the required level of certainty, and that operating limitations are consistent with those likely to
be achievable in the field.
In many cases, a model conforming to all the constraints described above cannot be constructed.
In these instances, decisions should be made as to whether a model that is more consistent with a 2P
outcome can be adapted for Proved Reserves estimation as discussed in the following sections, or
whether another methodology would be better used for Proved Reserves estimation.
Proved Reserves estimates arising from simulation, whether from a model constructed
specifically for that purpose, or following one of the approaches taken below, need to be examined
carefully to ensure that they satisfy the requirement that Proved Reserves are reasonably certain to
be recovered. For instance, simulation results that indicate a recovery factor for a reservoir that
significantly exceeds recovery factors for analogous reservoirs should be investigated to review
whether they are reasonably certain.
6.5.1 Adjusting 2P Results for 1P Reporting. It is widely recognized that, in many cases, a
deterministic modeling approach will yield prediction results that are a “most likely” or “best
estimate” case, thereby approximating 2P reserves (assuming compliance with the commerciality
criteria). However, companies may need to report only 1P outcomes, in which case, the 2P output
must be adjusted.
Adjustment of the deterministic model results may be performed in basically two ways:
• Modification of the input data to generate proved output
• Modification of the 2P output data to develop proved output
Each will be discussed in turn in the following sections.
6.5.1.1 Modification of Input Data. In this situation, the data provided to the model adhere to
the reasonable certainty requirement for 1P reserves reporting. For example, when no water
contact is otherwise identified, the 1P reservoir booking depth is predicated on the lowest-known
hydrocarbon from the well logs within a structural trap. If a downdip well encountered the reservoir
as wet, and no pressure gradient data exist, no water contact or free-water level can be interpolated
between these wells. Consequently, a 1P model could be built containing no hydrocarbon
accumulation below the lowest-known hydrocarbon.
Another possible means by which to constrain the dynamic model is to establish pay cutoffs
and impose them in the model, thereby making the hydrocarbons in certain cells immobile. While
disagreement exists in the industry about this practice, it is still widely applied in the industry.
Although modification of the input data serves to strictly honor the guidelines for Proved
Reserves, it actually may introduce more uncertainty. For instance, by limiting the accumulation
to a lowest-known hydrocarbon, water breakthrough may become difficult to match, possibly to
the point of requiring other modifications that are not physically reasonable. Further, in history
matching, the degree of pressure depletion measured in the field will either necessitate additional
revision of the model PIIP and/or changes in the underlying aquifer size and strength. Imposing
pay cutoffs in the dynamic model may create flow barriers or baffles and alter fluid migration paths
and interwell pressure differentials and cause conformance problems. In some cases, these barriers
to flow may improve the ability of the model to match performance, and in other cases, they may
restrict the flow too much. Barriers and baffles should not be imposed merely to achieve a match
unless there is supportive geoscience or engineering data for their existence.
142
6.5.1.2 Modification of Output Data. If a deterministic model (static and dynamic
components) is created without 1P constraints, then the resulting output will usually reflect a “most
likely” or “best estimate” scenario as noted previously. There are several ways by which we can
adjust the output to honor a 1P evaluation.
The first method assumes that the model serves as its own analog. In this scenario, the model
output enables the calculation of a recovery efficiency (model recovery divided by model PIIP
reflecting relevant petrophysical cutoffs). The resulting recovery efficiency or recovery factor (RF)
should be compared against other analogs or analytical methods to ensure it is within the expected
range, and it should be adjusted if it is not within that range. (Further adjustment of the RF may be
necessary if there are concerns about, for example, the reservoir continuity or quality away from
the current well control points.) A 1P estimate of PIIP (likely from a volumetric assessment) can
then be assumed, and the modeled RF may be applied to arrive at the 1P ultimate recovery. The
output rate forecast will have to be adjusted to honor the 1P ultimate recovery.
Such rescaling requires careful consideration. It is usually inappropriate to simply rescale the
entire production curves for each phase up or down. Further, care needs to be taken that if the
model is run out to a technical limit, then the RF achieved is likely larger than the effective RF at
economic limits. Therefore, it would be expected that the ultimate recovery from the adjusted
production curves would end up being diminished by the application of the economic limit.
A second method also requires review of the model production forecast. In the Petroleum
Resources Management System, Table 3 spells out how to characterize undeveloped locations
(wells) that would be considered Proved relative to those incremental locations considered as
Probable or Possible. Similarly, if the model contains multiple reservoirs, some of the reservoirs
themselves may be categorized as Probable or Possible, incremental to the Proved reservoir. In
this scenario, any predicted recovery from undeveloped locations or reservoirs that would be
categorized as anything other than Proved would be subtracted from the total model production
profiles and ultimate recovery to develop a Proved production profile. Fortunately, most
commercial simulation code can be configured to quantify and report exactly those wells and
intervals that produced hydrocarbons during the forecast, making it relatively easy to estimate the
amount of production that is arising from non-Proved sources, and allowing it to be subtracted
from the otherwise 2P forecast to provide a 1P forecast.
As a reminder, care needs to be taken to differentiate quantities that are technically recoverable
according to the simulation model from those that honor economic or contractual limitations.
6.5.2 Further Comment on Adjusting 2P Results for 1P Reporting. Both of these methods have
their disadvantages. In the first (modification of input data), the RF that is back-calculated may
not be limited to the quantities considered Proved; recovery from the model may include volumes
from lower-quality, higher-water-saturation content rock that—without being isolated from the
Proved volume—likely would result in a lower RF than the higher-quality, lower-water-saturation
content rock alone. For example, in a structural trap, the development wells may be positioned at
the crest, while the flanks represent reservoir rock increasingly deeper within the transition zone.
Overall, the RF from deeper pore volume should be less than the RF associated only with the
better-quality rock higher in the transition zone (or even in the dry oil zone only). Consequently,
the back-calculated RF from the model is expected to be more representative of the 2P or 3P RF,
which could be lower than the 1P RF. Under some circumstances, application of this first method
will yield conservative 1P quantities.
143
In the second approach (modification of the output data), simply removing the production
from non-Proved wells does not remove their effect on the flow streams within the model. Probable
and/or Possible producers nonetheless create pressure sinks, which alter the path, magnitude, and
even composition of the production from the Proved wells; this is a major concern when treating
the recovery on an incremental as opposed to a cumulative basis. Special consideration should also
be given as to whether a model that includes an element of nodal analysis, such as a shared surface
gathering network, is appropriately modeled when the probable volumes or wells are removed
from the production streams.
6.6 Other Output from Simulation

A frequently requested result from a simulation model is the average formation volume factor. For
ease of discussion, we will refer to an undersaturated oil reservoir and the calculation of an average
initial single-phase Boi in reservoir barrels per stock tank barrel (RB/STB). To support and be
consistent with a volumetric estimate, the hydrocarbon pore volume and PIIP both must be
calculated using pay cutoffs. Provided that the values for these quantities match within
approximately 5% between the static and dynamic models (judged without pay cutoffs), either
source could be used for the calculation, but it is generally easier to impose pay cutoffs in a static
model. Provided the values with cutoffs can be obtained, dividing the hydrocarbon pore volume
by the PIIP will yield a reasonable Boi (in RB/STB) that may be used in volumetric documentation.
Similarly, the recovery efficiency must be based on the volume of hydrocarbon that initially is
located within rock of high enough quality to flow (rock volume considered pay after the application of
petrophysical cutoffs) under the planned development operation. Simulation output does not usually
yield “reserves,” as there has been no economic or commerciality evaluation performed at that stage. The
forecast production from simulation provides only the Technically Recoverable Resources. When the
simulation forecast is incorporated into a suitable economics program, and the economic limit is
estimated, the reserves will be determined (subject to other constraints, such as technical or contractual).
This volume will then be divided by the PIIP (using pay criteria) from the model to arrive at the recovery
factor to be used in volumetric calculations. Note that this RF represents a project area RF and is not
necessarily reflective of a per-well RF, the difference between which is more pronounced in
unconventional reservoirs. As pointed out earlier, division of the production forecast into 1P/2P/3P
categories will likewise influence the associated recovery factors.
6.7 Acknowledgments
Thanks are owed to our employers in availing the time and resources to produce this chapter.
6.8 References
Lee, W. J., Sidle, R. E., and McVay, D. A. 2011. Reservoir Simulation: A Reliable Technology?.
Paper presented at the SPE Annual Technical Conference and Exhibition, Denver, Colorado,
USA, 30 October–2 November. SPE-146524-MS. https://doi.org/10.2118/146524-MS.
Mattax, C. C. and Dalton, R. L. 1990. Reservoir Simulation, first edition, Vol. 13. Richardson,
Texas, USA: Monograph Series, Society of Petroleum Engineers.
Palke, M. R. and Rietz, D. C. 2001. The Adaptation of Reservoir Simulation Models for Use in
Reserves Certification Under Regulatory Guidelines or Reserves Definitions. Paper
presented at the 2001 SPE Annual Technical Conference and Exhibition, 30 September–3
October. SPE-71430-MS. https://doi.org/10.2118/71430-MS.
144
Chapter 7
Probabilistic Resources Estimation

Carolina Coll (Chair)
David Elliott, Enrique Morales, Karl Stephen, and Richard Wheaton
7.1 Introduction
The prediction of reservoir properties and quantities (such as petroleum initially in place and
Technically Recoverable Resources) is subject to uncertainty due to the limited reservoir sampling
provided by data control points (e.g., wells), yet these predictions are major determinants in
reservoir performance forecasting. These production forecasts are then used for decision making,
driving investment in future projects where, in many cases, large sums of capital are required.
Industry uses various methods to understand and quantify the impact of reservoir uncertainty
on production forecasts and estimates of resources. These vary from purely deterministic methods,
where single values of each parameter are considered, to multi-scenario methods (hybrid methods
considered an extension of deterministic methods), to fully probabilistic methodologies, where
input parameters (e.g., porosity) are defined by probability distributions that are then combined to
obtain cumulative distributions of the outcomes (e.g., in-place estimates).
The purpose of this chapter is to provide technical guidance on the use of probabilistic
methods for resources and reserves estimation compliant with the Petroleum Resources
Management System (PRMS 2018) guidelines. Deterministic reserves estimation is explained in
detail in Chapter 4—Assessment of Petroleum Resources using Deterministic Procedures herein
and will be referenced throughout this chapter.
Resource evaluators often use deterministic methods to produce a “best estimate” of reserves
and resources based on a defined “base case” reservoir model. The result of this deterministic base
case model is usually close to the P50 estimate of the output quantity (where 50% of the estimates
exceed the P50 estimate). When using deterministic methods, low, best, and high estimate cases
are derived to assess the influence that a downside or upside of the input parameters could have
on the best estimate outcome. As with the best estimate and the P50 case, the selection of these
cases does not necessarily correspond to a 90% (P90) or a 10% (P10) probability of occurrence
despite the terms often being used interchangeably. When a probabilistic model of the same project
is assessed and compared with the outcomes from the deterministic approach, the resultant P90,
P50, and P10 scenarios should, however, reconcile with the low, best, and high estimates,
respectively (PRMS § 4.2.3.3). For example, the review of contacts and areal extent are common
items where such reconciliation may need to occur. Section 7.7 of this chapter discusses
comparison of the results from the two methods.
Although the deterministic approach is preferred by many evaluators due to its relative ease
of use and transparency, some of the difficulties with this approach relate to the natural tendency
of practitioners to aggregate upsides and downsides of the inputs for the level of confidence of the
forecast referenced. Combining all the “low-end” or the “high-end” values of the input parameters
can result in low cases that are too pessimistic or high cases that are too optimistic (and would
have more than 90% probability for the downside and less than 10% probability for the upside).
146
This topic is discussed in more detail in Chapter 8—Aggregation of Reserves and Resources
herein. Differences between the low deterministic estimates (1P, 1C, 1U) and P90 estimates can
be substantial, with similar issues for the high estimates (3P, 3C, 3U) compared to the P10
estimates. In probabilistic methods, the outcome of the analysis is P90, P50, and P10 estimates
that can be used for the 1P/1C/1U-2P/2C/2U-3P/3C/3U reserves/resources ranges, ensuring that
values correspond to the level of confidence required by the PRMS guidelines.
There are different probabilistic methods, and the applicability of each is related to the phase
of the field development. Methods more appropriate for the exploration phase include pure Monte
Carlo simulation (MCS), multi-scenario, or multiple realizations approaches (such as a hybrid mix
containing deterministic and probabilistic aspects), and experimental design with global
optimization methods are more suited for fields that are in the development phase, where
production exists, as observed in Fig. 7.1.
Fig. 7.1—Generalized field life cycle with associated probabilistic methods. EOR/IOR = enhanced oil recovery/improved
oil recovery.
As more performance data become available, the deterministic method is more frequently
used; however, in several situations (as shown in Fig. 7.1), some of the probabilistic methods are
still applicable.
The selection of the appropriate probabilistic method with which to estimate reserves and
resources depends on multiple factors, including:
• The type, quantity, and quality of geoscience, engineering, and economics data available
and required for both technical and commercial analyses
• The reservoir-specific data, including, but not limited to, the geologic complexity, the
recovery mechanism, the stage of development, and the maturity or degree of depletion
of the reservoir
• The knowledge and judgment of experienced professional evaluators, which are not to be
underemphasized when relying on reserves and resources assessments for decision
making
147
Fig. 7.2 shows the most common probabilistic methods. The application of the methods is
related to the development phase of the field as shown in Fig. 7.1. Simpler approaches such as
Monte Carlo methods are often used during exploration and appraisal phases, while other
techniques such as experimental design (ED) methods are used from the appraisal to the
development phases (Section 7.5.2). There are more sophisticated methods such as stochastic
optimization or ensemble Kalman filter methods (Section 7.5.3) that should be used once
production data are available.
Fig. 7.2—Probabilistic methods (where RSM is response surface model).
It must be stressed that, in most cases, the outcomes of the probabilistic method are P90, P50,
and P10 estimates of Technically Recoverable Resources (TRR). This estimate of TRR must be
converted into a forecast (i.e., production rate vs. time) for each of the low, best, and high cases.
After satisfying the commercial conditions required in PRMS § 2.1.2 and § 3.1, these P90, P50,
and P10 forecasted quantities become the 1P, 2P, and 3P Reserves. The commercial considerations
are key factors and must not be overlooked when using deterministic or probabilistic methods to
classify recoverable quantities as reserves.
7.2 Resources Uncertainty

Recoverable resources are a function of multiple uncertain parameters (e.g., areal reservoir extent, pay
thickness, porosity, permeability, etc.). Uncertainty may be due to several sources, including
measurement error, modeling, or incomplete data sets. Uncertainty also can be related to lack of
knowledge, for instance, fault geometry and/or extent, which may be resolved (or diminished) by
acquiring more data (e.g., drilling another appraisal well). Uncertainties are also related to the
inherently random nature of some reservoir properties, such as permeability, with values affected (for
example) by grain size, stratigraphic units, cementation, or fracturing. In the case of variability due to
reservoir heterogeneity at different scales, data collection might not always help to reduce uncertainty.
148
Uncertainty in data acquisition/processing is often small compared to that of characterizations and data
population in reservoir models (see Chapter 6—Reservoir Simulation herein).
7.2.1 Uncertainty in Input Parameters. To evaluate the uncertainty in resources estimates, we

first should look at the sources of uncertainty and their influence. Table 7.1 outlines the major
parameters used in volumetric calculation of recoverable resources, their sources of uncertainty,
and their influence on the resources estimates.
Table 7.1—Sources of uncertainty for volumetric estimates.
149
The evaluator needs to define not only the key uncertainties, but also their reasonable ranges
and the type of distribution. Ranges should be defined based on the geoscience information or on
analogs if data are not available. Distributions should be selected to represent the data and/or
analog models, avoiding extrapolations beyond the understanding of the reservoir. Typically,
normal and log-normal distributions are used with truncation applied to avoid infinite tails.
Uniform distributions are used when data are limited, while triangular distributions are often used
if data are limited and ranges are narrow. Independent of the distribution used, it is important to
ensure that ranges are wide enough to consider the potential outcomes. A common mistake is to
limit the ranges based on data from the subject project, especially during exploration, ignoring data
ranges from analog reservoirs. Another aspect to consider is dependency; uncertainties among
parameters controlling reservoir performance may not be independent. In many cases, there are
dependencies between uncertainties (Carter and Morales 1998) that can cause unexpected and
significant deviations from expected project outcomes, if they are not taken into account (e.g.,
porosity and permeability).
Evaluators need to spend time trying to understand the influence of these subsurface reservoir
uncertainties, together with facilities and other constraints, on the estimation of resources quantities.
7.2.2 Project Uncertainties. Some authors (Acuña and Harrell 2000; Wheaton and Coll 2010)
have grouped project uncertainties that should be considered during the resources evaluation
process into three main groups: technical, project maturity, and economic.
7.2.2.1 Technical Uncertainty. The ranges for geological and engineering uncertainties
described above are typically very large before discovery, primarily impacting prospective
resources, and generally narrow through the appraisal and production phases.
Technical uncertainties mainly refer to geological and engineering parameters related to the
subsurface and surface aspects of the project. Geological uncertainties related to the volume of
hydrocarbons (e.g., gross rock volumes, net-to-gross ratio, porosity, hydrocarbon saturations) have
a large influence on the in-place estimates as described before. Engineering uncertainties (e.g.,
drive mechanism, relative permeability, capillary pressure, viscosity, development plans, surface
facilities, well spacing) influence the quantity of hydrocarbons that can be recovered through a
particular process and therefore will be key factors in the estimation of resources.
Project investment decision making is largely influenced by technical subsurface uncertainties
that can put at risk large sums of capital investment. The range of these technical uncertainties can
make the difference between project success or failure, stressing the importance of understanding
the uncertainties and associated risks for the project. (Note that this also applies to deterministic
methods.) Legal, contractual, and regulatory aspects as well as social and environmental aspects
should be considered (PRMS § 2.1.2.1).
7.2.2.2 Project Maturity Uncertainty. Project maturity concerns the classification and sub-
classification of resources based on contingencies and uncertainties. Uncertainty related to legal,
contractual, regulatory, governmental, working interest entities, market availability, or
transportation elements, along with the time required for full field appraisal, could, following
discovery, represent a significant uncertainty for potential value.
7.2.2.3 Economic Uncertainty. Future prices and costs (development and operating) represent
another type of uncertainty with a large influence on the economics of a project and therefore on
the commercial viability of a project for reserves estimation.
150
Economic uncertainties are mainly related to uncertainty in economic parameters used to
evaluate a project such as price and costs, and this may be addressed with sensitivity scenarios;
however, an evaluator needs to be careful because the PRMS does not allow split conditions (i.e.,
the same commercial conditions must be applied to the different categories within a resource
class). Companies take great effort to define ranges of forecast price scenarios that can be used for
the internal investment process and used for sensitivity cases. Economic evaluation is based on the
entity’s reasonable forecast of future conditions that will exist during the life of the project.
Economic evaluation is discussed further in Chapter 9—Evaluation of Petroleum Reserves and
Resources of this volume.
7.3 Deterministic Methods

The PRMS defines the deterministic approach as an “assessment method based on discrete
estimate(s) made based on available geoscience, engineering, and economic data and corresponds
to a given level of certainty” (PRMS Appendix A, Glossary). Deterministic methods are applied
by making discrete estimates of hydrocarbons initially in place or recoverable quantities, each
based on a single set of input parameters.
7.3.1 Deterministic Analysis. The results of a deterministic analysis can be a single best
estimate, or they can be sensitivities of the best case representing a range of resources outcomes
(i.e., the low and high cases). To understand how each uncertain input reservoir parameter
influences the outcome, the evaluator can use a sensitivity analysis (the deterministic scenario
method), where “the evaluator provides three deterministic estimates of the quantities to be
recovered from the project being applied to the accumulation. Estimates consider the full range of
values for each input parameter based on available engineering and geoscience data, but one set is
selected that is most appropriate for the corresponding resources confidence category. A single
outcome of recoverable quantities is derived for each scenario” (PRMS § 4.2.1.2). The selection
of the best estimate for each input parameter may be straightforward. The parameter selection
should be agreed upon by a multidisciplinary team with the objective of appropriately representing
resources uncertainty. A more difficult task is to select the combination of input reservoir
parameters to use for the low and high cases to produce realistic low and high resources estimates.
An example of the application of simple deterministic methods is shown in Fig. 7.3. As
tabulated, three values (low, best, and high estimates) are developed for each uncertain input
parameter, based, if possible, on distributions of reservoir properties derived from available
information and analogs.
Some of the advantages of making discrete deterministic estimates are that the results are
reproducible, fast to implement, auditable, and transparent in honoring resources guidelines
constraints (e.g., lowest known hydrocarbon, proved area), the results reflect a combination of
static and dynamic parameters that can physically exist, and they are relatively simple to prepare.
These methods allow input parameters to be selected that help to eliminate any inconsistent
(nonphysical) combinations of properties. Further, these methods are simple to interrogate and,
therefore, are popular among investors, companies, and regulatory agencies. However, to estimate
confidence levels associated with these reserves/resources estimates, hybrid deterministic or
probabilistic approaches should be used.
151
Fig. 7.3—Deterministic estimation for in-place volumes, where NTG indicates net-to-gross ratio and STOIIP indicates stock
tank oil initially in place.
7.3.2 Sensitivity Analysis and the Multiple-Realizations Method. Sensitivity analysis consists
of using deterministic runs to evaluate the potential influence of various parameter uncertainties.
Changes are performed to one parameter at a time on the best estimate parameters. These
sensitivities are typically run using the low and high values for each parameter, and the impact on
the response variable under evaluation is observed (e.g., recoverable volumes). Sensitivity analysis
is an integral part of probabilistic methods to estimate recoverable resources. This allows the
evaluation of the key uncertainties and their impact on the outcome variable. Plots such as tornado
diagrams are used for this purpose (as explained below) to determine the key uncertainties.
The multiple-realizations method (often called multi-deterministic scenarios method) is
essentially an extension of the deterministic scenario method described in Section 7.3.1, combining
the main advantages of the deterministic approach with those of a simple Monte Carlo approach
(see Section 7.5.1). The multiple-realizations method involves the identification of key parameters
(static and dynamic) from a sensitivity analysis, which are then ranked to identify those with the
greatest influence on the outcome (e.g., hydrocarbon recovery). The use of tornado diagrams in
identifying key parameters has been widely addressed (e.g., Twartz et al. 1998; Van Elk et al.
2000; Vahedi et al. 2005) and should be considered in sensitivity studies and uncertainty reduction,
thereby simplifying the overall estimation workflow. This analysis results in the generation of
multiple physically realistic reservoir descriptions (or realizations) that can be reconciled with
specific probabilities while meeting certain constraints (e.g., fluid distribution limited by fluid
contacts or lowest known hydrocarbon, or dealing with unpenetrated fault blocks). These
realizations can then be used to calculate a cumulative distribution function (CDF) as opposed to
three discrete scenarios generated from a deterministic scenario method. From the CDF, the
152
evaluator can select scenarios with at least a 90% probability for the low case, 50% for the best
case, or 10% for the high case in terms of quantities actually recovered.
The main advantage of this method is to help establish a sound and well-structured uncertainty
assessment and management process with a transparent audit trail while honoring the physical
understanding of the models and realizations. However, an exhaustive analysis and quantification
of the effect of key uncertainties is not feasible if there are large numbers of parameters and
combinations to consider, such that the time required for data preparation and computation
becomes unrealistic. During recent decades, the multiple-realizations/scenario method has evolved
into what is now widely used and referred to as the experimental design method or EDM, which
is a more practical option (see Section 7.5.2).
In the multiple-realizations method, the steps are:
(1) Identify the key uncertainties that significantly affect the outcomes through a sensitivity
analysis.
(2) Establish discrete ranges of outcomes for key parameters with their chance of occurrence.
(3) Build a realization tree.
(4) Derive a pseudo-CDF.
(5) Select P90, P50, and P10 outcomes.
Tornado diagrams are constructed using the results of a sensitivity analysis in which one
parameter is varied at a time from the base case while keeping all the other parameters unchanged.
Each parameter is set to a “low” and a “high” value, and the response variable is measured [e.g.,
oil in place, ultimate recovery, or net present value (NPV)]. The range of variation of the response
variable is captured on the x-axis, and parameters are ranked from largest variation down to
smallest. A threshold in the variation can be chosen to bracket the most important parameters.
Fig. 7.4 shows an example of a tornado diagram where, for this specific case, the gas initially
in place (GIIP) uncertainty range is being studied. In this example, the most impactful uncertainties
are the structure (e.g., due to time-depth conversion uncertainty) and the position of the gas-water
contact (GWC), as well as reservoir properties such as porosity combined with net-to-gross
thickness ratio. These key uncertainties significantly affect the value of the parameter under
evaluation (in this case GIIP) with more than ±25% impact. This assessment must ensure that the
key parameters with large impact are identified, so further analysis avoids spending time on
sensitivity studies of variables with negligible impact.
Fig. 7.4—Tornado diagram for parameters impacting gas initially in place (Twartz et al. 1998, simplified).
153
Similarly, Fig. 7.5 shows an example of a tornado diagram highlighting the key uncertainties
in the estimation of gas recoverable volumes. In this case, the key variables are GIIP, aquifer
support, and cross-fault transmissibility.
Fig. 7.5—Tornado diagram for parameters impacting gas recoverable volumes (Twartz et al. 1998, simplified).
This approach can be easily implemented and is most valuable when investigating the effect
of uncertain input parameters in the outcome (static or dynamic) in a multi-scenario approach.
One of the main limitations of this method is that each calculation only considers the effect of
one input uncertainty at a time and ignores all potential interdependencies and interactions. The
interdependencies between input reservoir parameters can be an important factor in the outcome
of the analysis. On the other hand, changing two parameters simultaneously from the base case
makes this method cumbersome as the number of simulations increases, but there is no set
systematic methodology with which to decide the number and the combinations that should be
attempted. It is recommended that interdependencies of parameters be carefully evaluated during
reservoir characterization so the selected workflow takes into account these key interactions. Such
interactions can have a more important effect than the individual uncertainties. An exhaustive
evaluation of the impact of these interdependencies is time-consuming and computationally
intensive, particularly in mature projects requiring numerical simulation and history matching.
However, such an evaluation can be accomplished using commercial software that combines all
parameters as part of a probabilistic workflow, minimizing the manual process of setting up all the
numerical simulation runs.
7.4 Hybrid Methods

7.4.1 Probability Tree Analysis. Probability tree analysis is similar to the multiple-realizations
method described previously, where the nodes in the decision tree represent realizations of key
uncertainties shown in the tornado diagram (i.e., the most important ones), but it has the advantage
of assigning a probability to each realization. Fig. 7.6 shows an example of this method used to
estimate recoverable volumes compared to a multiple-realization approach. Three different static
reservoir models were built based on three structural model realizations (low, middle, and high)
using different time-to-depth conversions and variable GWCs, two of the key uncertainties
154
impacting the GIIP in Fig. 7.4. Two GWCs were considered for each structural realization based
on a low and a high GWC value to generate six GIIP realizations. Ranges of recovery factors could
then be estimated for each static realization using tornado diagram analysis of key uncertainties as
explained in Fig. 7.5. This generated ranges of TRR for each static realization.
Fig. 7.6—Probability tree analysis for Technically Recoverable Resources (TRR), where RF is recovery factor.
The combined probability for a branch was estimated and plotted against the recoverable
volumes using a reverse CDF as shown in Fig. 7.7. Results show that the P50 estimate is close to
22 MMSTB, whereas the P10 and P90 estimates are 80 MMSTB and 4 MMSTB, respectively.
However, it must be pointed out that, while selecting an estimate from a probability tree analysis
(e.g., TRR from Fig. 7.7), values will not necessarily be linked to a subsurface realization but to
an extrapolation between two scenarios. For instance, the P10 TRR in Fig. 7.7 does not exist
physically but is an extrapolation between realizations, so the closest realizations available should
be identified to help define a P10 case taking into account assumptions that are consistent with the
guidelines, for instance, fluid contacts such as the lowest known hydrocarbon and GWC.
155
Fig. 7.7—Reverse cumulative distribution of TRR using decision tree analysis.
The application of probability tree analysis is probably more familiar through its use with
decision making under uncertainty. One such application occurs during field development by
looking at, for example, the expected monetary value because it allows assessment of alternative
options that may maximize the outcome (e.g., recoverable resources volumes or project value).
For example, Fig. 7.8 illustrates an example of a prospect with sparse two-dimensional seismic
data across the area of interest, where the seismic survey was run as part of an area survey around
another, existing, field. The accumulation in question has not yet been drilled. A decision tree is
developed to understand the economic impact of acquiring new three-dimensional seismic data vs.
drilling an exploration well without additional seismic surveys. The expected monetary value of
drilling an exploration well using three-dimensional seismic information is estimated at USD 18
million compared to an expected monetary value of USD 14 million for drilling the exploration
well without acquiring the three-dimensional seismic data (assuming a simple discount net back
of USD 1/bbl).
Probability tree analysis is a useful tool to provide a CDF of resources when multiple
geological scenarios are considered. This method has become popular during exploration,
appraisal, and early stages of development to evaluate the economic benefits (such as the potential
to identify larger quantities of petroleum initially in place) of data acquisition relative to the cost
required (such as acquiring three-dimensional seismic data or drilling a well) to obtain that data
for value of information decision analysis. The method offers user-friendly graphical
representations to help with the analysis and is intuitive to use. However, its accuracy is related to
the assignment of the probabilities for each branch, which are selected by the evaluator and
therefore subjective. A recommended approach is to conduct framework sessions with the
multidisciplinary team to agree on the probabilities to use.
156
Fig. 7.8—Example expected monetary value (EMV) probability tree for exploration well.
7.4.2 Multi-Scenario Method. The sensitivity analyses described in the previous section
focused on reservoir properties. However, there are also uncertainties related to the field
development plan, such as size of processing facilities, number of wells, delivery pressures, etc.
Different development scenarios can be selected that are technically feasible combined with
different subsurface realizations (Twartz et al. 1998; Van Elk et al. 2000). For instance, well
optimization can be a key uncertainty requiring evaluation because it will affect well count and
well spacing, which will determine sales volumes; other factors include delivery/arrival pressures,
minimum tubing head pressure, and compression facilities. All will affect the development plan
and production forecast.
An example of combined scenarios that are often considered during field development
planning is shown in Table 7.2. It is a common practice to run different development scenarios
using petroleum initially in-place (PIIP) ranges and then consider other variables that will affect
TRR (such as through the use of tornado diagrams). The number of scenarios to be considered will
depend on the project characteristics. As the number of parameters to be considered increases, so
do the number of simulation runs required to determine the low, best, and high case TRR scenarios.
157
Table 7.2—Example of multi-scenario method.
Multi-scenario methods allow for the explicit evaluation of different geological models and
development scenarios and therefore can be used to explore the wide range of outcomes required
during project screening. A simplifying assumption for these scenarios is that they be equally
probable, although different probabilities can be assumed if justified. A good understanding of
project and reservoir uncertainty effects on resources estimates can be achieved if enough cases
are considered to yield statistically representative results. Running a sufficient number of
combined scenarios can be time-consuming, so care should be taken in selecting the cases to run.
One of the main disadvantages of this method is that the quantitative treatment of probabilities is
more limited compared to the probability tree analysis method (Section 7.4.1).
The difference between the multi-scenario and the probability tree methods is that the former
considers development options along with certain combinations of reservoir uncertainties, and the
input parameters to use in each scenario are selected by the practitioner(s). These scenarios are
selected based on the multidisciplinary team views on key uncertainty and ranges. By comparison,
in ED, a statistical design of experiments is used to develop the scenarios and is more of a true
probabilistic approach.
The added value of using multi-scenario methods in the hydrocarbon maturation and
production process is that they:
• Identify the key parameters (static and dynamic) and their uncertainties, which play a role
in the project or field development and recovery, and which must be carefully addressed,
monitored, and managed
• Describe the full range of uncertainty and reveal upsides (opportunities) and downsides
(risks)
• Can reduce the range of uncertainty, especially when using performance data
• Can be used to estimate the value of information of potential activities aimed at
addressing upside or downside uncertainties
• Allow evaluation of the impact of interdependencies
• Provide a good interface with decision support and financial modeling methods
• Can easily be applied across the full life cycle from exploration to production activities
• Yield probability outcomes that can easily be inspected and quality controlled to ensure
they are related to a physically realistic discrete reservoir description (or realization),
which can be used to ensure regulatory or other constraints are accounted for (e.g., check
158
for volumes below the lowest-known hydrocarbon or in noncommunicating unpenetrated
blocks being included in the Proved Reserves estimation)
The chosen multiple realizations of the subsurface should be pragmatic and thus:
• Be based on a small number of key uncertainties (e.g., 3 to 5), obtained from ranking the
static and dynamic uncertainties using some quantitative approaches (e.g., sensitivity
analysis with tornado diagrams). (Some uncertainties are not significant and should be
set at their best estimate value to avoid an unmanageable number of combinations and
computational effort.)
• Have internally consistent realizations (i.e., a realization should consist of parameter
values or sets of conditions that can physically exist together).
• Be associated with a probability of occurrence (but not necessarily equally probable).
• Be related to technically sound development options.
The multi-scenario method also can be used with each branch representing an individual
simulation run (history matched, if production history exists). By assigning probabilities to these
branches, it is possible to identify appropriate P90, P50, or P10 realizations of recoverable
volumes. Because this is not strictly a probabilistic method, it is not necessary to select outcomes
at precisely the probability equivalents of these categories (e.g., at 90% probability).
Although the most probable value of the distribution is the mode, common industry practice
(as in the PRMS) is to use the median (P50) as the best estimate for a single entity (reservoir or
zone). Using the median implies that half the outcomes will be greater than this value, and half
will be smaller. Under the PRMS guidelines, the commerciality test for reserves determination is
applied to the best estimate (P50) forecast quantities (PRMS § 2.1.2.2).
7.5 Probabilistic Methods

The PRMS notes (§ 4.2.3.1) that in the probabilistic method, “the evaluator defines a distribution
representing the full range of possible values for each input parameter. This includes dependencies
between parameters that must also be defined and applied.” The PRMS (Appendix A, Glossary)
also clarifies that the “method of estimation of resources is called probabilistic when the known
geoscience, engineering, and economic data are used to generate a continuous range of estimates
and their associated probabilities.” Chapter 4 of Society of Petroleum Evaluation Engineers
Monograph 3 (Hall et al. 2010) describes a statistical methodology with which to estimate proved
undeveloped reserves for a resource play.
The objective of applying probabilistic volumetric methods is to estimate probability
distributions of TRR (Acuña and Harrell 2000). These are obtained by combining probability
distributions for all the parameters in the volumetric equation (see Eq. 4.1a in Chapter 4—
Assessment of Petroleum Resources using Deterministic Procedures herein) through random
sampling (e.g., using Monte Carlo methods or stochastic geological modeling). The probability
density functions for the input parameters may or may not be stochastically dependent.
As an example, when estimating in-place volumes, input uncertainties such as gross rock
volume, net-to-gross (NTG) ratio, porosity, and fluid properties need to be addressed (refer to
Table 7.1). In the case of gross rock volume, one of the key uncertainties for in-place estimates is
the locations of any faults that are present (for the purpose of resource categorization). This
example is addressed more fully in Chapter 8—Aggregation of Reserves and Resources herein.
The type of distribution to use should be based on the existing data, provided enough data
exist (El-Khatib 1999; Annan et al. 2020). The evaluator should consider the range and shape of
the input distributions based not only on the existing information from project area wells (which
159
often represents a limited sampling), but also on geoscience information or direct reservoir and
geoscience information from appropriate analogs. Recall that the range of values required is that
which represents the evaluator’s uncertainty in the value of the mean, rather than the distribution
of the data itself. Always keep in mind that the distribution function should describe the
distribution of the reservoir-averaged parameter value.
The most common distributions of reservoir properties are normal and log-normal
distributions. Average porosity is often found to exhibit a normal distribution, while average
permeability is typically modeled with a log-normal distribution. When there is not enough
information about the distribution (perhaps there are minimum, maximum, and mean values from
analogs), then a triangular distribution could be used to help avoid extrapolation to infinite values.
Triangular distributions are suited when extrapolation of the minimum and maximum needs to be
controlled but is not robust at the “tails,” e.g., P1 or P99.
Truncated distributions may prevent extrapolation to negative values but can affect the overall
shape of the probability density function, which is not recommended. Truncated curves also can
be used to account for cutoffs; for instance, a normal porosity distribution for a gas reservoir could
be truncated at a 5% porosity pay cutoff, if anything less than that was not expected to contribute
to flow.
For variables where all the values are considered to be equally probable, a uniform distribution
can be used. Examples of this could be formation volume factors as determined from the
laboratory. Results of the probabilistic analysis depend greatly on the type of probability
distribution used for the different reservoir parameters. Therefore, available data should be
carefully analyzed. At the exploration/appraisal stage, limited data exist, and thus confidence in
the derived distribution is low. At this stage, the use of analog data and prior geological knowledge
can help to define probability distributions that properly represent reservoir uncertainty.
A CDF of the outcome (e.g., recoverable quantities) will be obtained as part of this process,
along with the P90, P50, and P10 estimates (Fig. 7.9). If dynamic (reservoir simulation) models
are created, then the associated production forecasts generated must be screened for economic
viability with the applicable cost/investment profiles, economic limit, and other commercial
requirements as described in the PRMS in order to derive the corresponding low/best/high
estimates of the resource quantities (or 1P/2P/3P Reserves, if commercial). The relationship
between the P90, P50, and P10 estimates and the 1P, 2P, and 3P Reserves in terms of the reserves
guidelines requirements is explained in Section 7.7.
One of the main advantages of using probabilistic methods is that the confidence levels or
probabilities associated with the low, best, and high reserves can be calculated. This is important
information to consider for resources and reserves. For example, 1P Reserves or 1C Resources
should correspond to a low estimate (P90) of the quantity that actually will be recovered. The best
estimate (2P/2C) case will correspond to a P50 estimate, and likewise the high estimate (3P/3C)
case should correspond to the P10 estimate of the resources that actually will be recovered.
Fig. 7.10 shows the results of a probabilistic MCS using the example in Fig. 7.3. As
observed, the low case is lower than the P90 probabilistic estimate, whereas the high case is
higher than the P10 estimate due to the selection of all low input parameters for the low case
and all high-case input parameters for the high case. The best case is close to the probabilistic
P50 as expected (Carter and Morales 1998; Morales and Lee 2014). This demonstrates the
importance of selecting the correct input parameters to define low and high cases closer to 90%
and 10% probabilities, respectively.
160
Reservoir uncertainties Reserves estimations
Fig. 7.9—Probabilistic methods (in this case, commercial analysis has been performed and reserves categorized).
Fig. 7.10 also shows the effect of changing two of the input probability distributions (porosity
and Sw) from a normal to a uniform distribution on the P90, P50, and P10 petroleum initially in
place values (here denoted as stock-tank oil initially in place, STOIIP).
Fig. 7.10—Example probabilistic estimations for in-place volumes.
Fig. 7.11 shows the forward and reverse CDFs generated from the constant distribution in
Fig. 7.10. The left-hand side of Fig. 7.11 shows a forward cumulative distribution. To create this
chart, the frequencies are added cumulatively, starting from the lower end of the range, and then
plotted as a cumulative frequency curve.
To create the reverse cumulative distribution, on the right, the frequencies are added
cumulatively starting at the higher end of the range, and then plotted as a declining cumulative
frequency curve. In the PRMS, the low-magnitude value is associated with the 90% probability,
whereas the P10 estimate is associated with a high-magnitude value, which corresponds to a
161
reverse cumulative distribution. However, some companies prefer to use a forward cumulative
distribution where P10 relates to a low case and P90 corresponds to the high case.
Fig. 7.11—Example cumulative distributions (forward and reverse).
Probabilistic methods can vary from Monte Carlo techniques that use simple analytical
approaches to estimate recoverable volumes (Minin et al. 2011) to complex methods such as global
optimization workflows with algorithms that attempt to find the global minima or maxima of a
response function (e.g., recoverable volumes) using dynamic simulation models (Ogbeiwi et al.
2020). These algorithms are often described as minimization routines that are iterative and
computationally intensive using full-field simulation models. Some of these methods of stochastic
optimization include more sophisticated algorithms such as evolutionary strategy and genetic
algorithms (see Section 7.5.3 and Fig. 7.2) and ensemble-based methods.
The ensemble Kalman filter method, one of the global optimization methods, was introduced to
petroleum science by Lorentzen et al. (2005) to help solve complex history matching problems. This
method provides a workflow to incorporate diverse data types from seismic data to production. The
ensemble Kalman Filter method relies on a cross-covariance matrix (see chapter Glossary)
computed from an ensemble of reservoir models to relate reservoir properties to production data.
Many equally probable reservoir models can be created to honor the data/information (both static
and dynamic), making it possible to represent the uncertainty in the modeling.
The number of models used needs to be kept small for computational efficiency, but, if an
insufficient number of simulations are run (less than 100), then the results can include spurious
estimates (Arroyo-Negrete et al. 2008). One of the advantages of these methods is that petrophysical
properties between wells can be modified to improve the history match. Another advantage is that
correlations between uncertain parameters and the response can be investigated efficiently.
Selecting between the different probabilistic methods depends on the technology available to
(and the skillset of) the practitioner, the particular characteristics of the study, and the stage of field
development (as described in Fig. 7.1). A common and simple method used by practitioners for
resources estimation is decline curve analysis. Well-by-well production decline analysis is often
used to define a range of estimated TRR (see Chapter 10—Unconventional Resources Estimation
herein) and the parameters used for the decline equations (Coll and Elliott 2013). Probability
distributions of the different input parameters, such as the hyperbolic exponent factor “b,” can be
used with MCS to determine the P90, P50, and P10 ranges of the Estimated Ultimate Recovery.
162
7.5.1 Monte Carlo Methods. Monte Carlo methods are a class of computational algorithms that
rely on repeated random sampling to provide solutions to problems, particularly where there may
be significant uncertainty in reservoir description. MCS will provide ranges of solutions of the
outcome parameter. In a standard MCS, the probability distribution of the input parameters is
sampled at random from their specific distributions (Wadsley 2005; Komlosi and Komlosi 2009).
A mathematical model in which a dependent variable is a function of the independent
variable(s) is developed. The dependent variable is the response function, for instance, recoverable
resources or in-place volumes, whereas the independent variables could be uncertain parameters
such as gross rock volume, porosity, permeability, and/or net-to-gross ratio. Distributions are then
assumed for each independent variable based on the evaluator’s experience and the available data.
Once the mathematical model is built for the dependent variable (e.g., resources volumetric
estimate), random numbers will be generated for each independent variable based on user-defined
statistical distributions. Monte Carlo methods used for most forecasting methods can be easily
implemented in a spreadsheet formula (volumetric, analytical, or decline curve analysis) if the
uncertain parameters are described by a probability function. Other uncertain parameters related
to production can also be included in the analysis (such as processing plant availability), which
will influence yearly production.
For example, the in-place volumes and ultimate recoverable quantities are estimated by using
computationally simple volumetric and material balance equations, permitting many Monte Carlo
simulations (see Fig. 7.12) to be performed (typically >1,000 and <10,000), because the only
requirement is to solve simple algebraic equations.
Fig. 7.12—Monte Carlo approach to volumetric estimates, where GRV is gross rock volume, and NTG is net-to-gross ratio.
The CDF obtained with such a procedure contains the estimates of recoverable volume ranges,
but there is no information on the combination of input parameters (related to a subsurface
realization) that will produce a value for the resource estimate. For this reason, these methods are
mostly used during the exploration and early appraisal phases, when limited information exists
with which to build geological models.
163
7.5.2 Experimental Design Methods and Response Surface Models. ED methods are a group
of techniques that perform statistical design of experiments to maximize the information from a
minimum number of experiments (simulations). They have been used extensively in the oil
industry for uncertainty analysis since the 1990s (Elvind et al. 1992; Van Elk et al. 2000; Manceau
et al. 2001; Kabir et al. 2002; Cheong and Gupta 2005; Vahedi et al. 2005; Coll, 2019) and for
optimization (White and Royer 2003; Zhang et al. 2007; Fetel and Caumon 2008). ED methods
search the solution space by exploring the effect of reservoir uncertainty on forecasts using
reservoir models such as material balance (Vahedi et al. 2005) that can vary from basic reservoir
engineering analysis (e.g., production decline analysis) to more sophisticated workflows such as
multiphase simulation models.
The approaches can be thought of as an extension of the sensitivity analyses of the previous
section. The main difference here is that, instead of considering one parameter changing at a time,
multiple combinations are used. This allows interactions and nonlinear effects to be measured from
the models. If the response variable is thought to be a linear function of the parameters with no
interaction, then the Plackett-Burman method (see chapter Glossary) can be used. Combinations
of high and low values of parameters are used based on an algorithm. This requires slightly more
simulation runs than the number of parameters being investigated and can be more efficient than
one parameter at a time as typically used for tornado diagram analysis. If the response is thought
to vary in a quadratic sense with the parameters, then three values of each variable are needed so
that the number of models needed is three to the power of the number of parameters, if even
sampling is used. This reduces the need for every combination to be used.
Examples of alternative ED methods include Latin hypercube sampling and orthogonal
sampling (see chapter Glossary for definitions). Latin hypercube sampling is a statistical method
for generating a near-random sampling of parameter values from a multidimensional distribution
(Mishra and Datta-Gupta 2017). In statistical sampling, a square grid containing sample positions
of an uncertainty distribution is a Latin square if there is only one sample in each row and in each
column. In Fig. 7.13, two uncertainties are denoted as u1 and u2. The ranges of uncertainty are
divided into four bins. Latin hypercube sampling experiments are constructed in such a way that
each one of the dimensions (i.e., uncertainties) is divided into equal intervals or bins, and there is
only one point (or sample) in each interval (four intervals in Fig. 7.13).
In the orthogonal model, the sample space is divided into equally probable subspaces. All
sample points are then chosen simultaneously, making sure that the total set of sample points is a
Latin hypercube sample and that each subspace is sampled with the same density.
Fig. 7.13—Examples of experimental design (ED) for two uncertain variables.
164
Results of experimental designs are then used to generate a proxy model (which can be a more
accurate mathematical function including interaction terms between input variables) by using a
high-order polynomial function to generate a response surface model (RSM), as plotted in
Fig. 7.14. A higher-order function RSM is a quadratic polynomial function that is required when
interaction of parameters is needed to minimize the error (e) between the data and the simulation
model due to curvature of the solution space or interaction of uncertain parameters. An example
of a higher-order RSM is shown below, where ui are the uncertain parameters, ai are the unknown
coefficients we need to determine, and e is the error:
𝑅𝑅𝑅𝑅𝑅𝑅 = 𝑓𝑓(𝑢𝑢1, 𝑢𝑢2, . . . . 𝑢𝑢𝑢𝑢) = 𝑎𝑎0 + 𝑎𝑎1 𝑢𝑢1 + 𝑎𝑎2 𝑢𝑢2 + 𝑎𝑎3 𝑢𝑢3 + 𝑎𝑎11 𝑢𝑢1 2 + 𝑎𝑎22 𝑢𝑢2 2 + 𝑎𝑎33𝑢𝑢3 2 +. . . 𝑒𝑒.
Ultimate recoverable resources or recovery factors are examples of the RSM function obtained
using different uncertain input parameters (such as porosity and permeability). We call u1, …, un
explanatory or uncertainty variables. A reservoir simulator can be seen in this case as a complicated
function that relates the response RSM (e.g., Estimated Ultimate Recovery) to the explanatory
variables (input uncertainties). The RSM function in Fig. 7.14 assumes no interaction or
dependencies between the uncertain parameters (u1, …, un). This is a simplification, because, in
most cases, uncertain input parameters can have dependencies.
Fig. 7.14—Response surface model (RSM) along with simulation results (blue circles) selected using ED methods
(after Coll 2019).
The advantage of using a polynomial function as the proxy model is that it can be interrogated
to determine those parameters that are important. Simple EDs such as Plackett-Burman can be
used to eliminate unimportant parameters, and then higher-order EDs can be used to improve the
accuracy of the proxy model (e.g., Fig. 7.15).
MCS is then performed using the polynomial function (or whichever proxy model or function
is derived to approximate the full model) to estimate the cumulative distribution of the RSM function
(e.g., TRR), and, from there, the P90, P50, and P10 estimates of recoverable volumes are obtained.
If the interaction of input parameters is critical, then a different polynomial function should be used
that includes curvature terms instead of the simpler equation in Fig. 7.14. Fig. 7.15 summarizes the
recommended workflow using the two methods (ED and MCS) explained previously.
165
Fig. 7.15—Example workflow using ED and Monte Carlo analysis.
7.5.3 Global Optimization Methods. In fields with production data, a history matched model
can be used to better quantify ranges for in-place quantities and forecasts of production. Given the
nonuniqueness of history matching, it is necessary to find a set of dynamic models that more
closely approximate physically real combinations of input parameters and better understand the
uncertainty range. For this purpose, more sophisticated methods beyond ED and MCS, such as
global optimization methods, are necessary. Global optimization methods (also called direct search
methods) are used in different areas of engineering (Hooke and Jeeves 1961). These methods are
based on using an objective function to determine the next steps to search for the solution instead
of using gradient information (Schulze-Riegert et al. 2006, 2007). The objective function is the
difference between observed and modelled data. Several global optimization methods have been
developed for oil industry use, including an objective function defined, for instance, to minimize
the difference between measured and simulated bottomhole pressure during history matching
(Al-Shamma Teigland 2006; Schulze-Riegert et al. 2006, 2007; Choudhary et al. 2007).
Examples of these methods include evolutionary algorithms called genetic algorithms,
evolutionary strategies, and evolutionary programming. From these methods, the most common
are evolutionary strategies (Cheng et al. 2008; Abdollahzadeh et al. 2013; Ghamdi et al. 2020) and
genetic algorithms (Castellini et al. 2008; Sanghyun and Stephen 2018) based on biological
processes. In evolutionary strategies and genetic algorithms, the algorithm will search for the best
combination of the uncertain parameters in the simulation that will produce the best history
matches. In using these methods, many reservoir models are found that closely match the history
data. They can also find solutions that are quite different from each other in terms of the input
parameters. The statistics of the models can then be used to infer probabilities for uncertainty (or
risk) analysis (Stephen 2018).
As noted in Section 7.2, the ensemble Kalman filter is a method that is useful for modifying
reservoir properties away from the wells. The ensemble Kalman filter method is based on a
formulation of optimization where data are assimilated based on calculations of probabilities
(Aanonsen et al. 2009). The probabilities of existing models are used to update the model at the
166
next step. Once history matching is complete, the CDFs of model parameters can be evaluated by
resampling probabilities. Ensemble-based optimization is another popular method for history
matching that incorporates production history data into a set of model realizations.
The next section explains how some of these different methods can be used to develop a
probabilistic workflow that can be applied to either preproduction or on-production fields.
7.6 Probabilistic Workflows for Recoverable Resources Estimates

There are two main distinctive workflows that need to be considered for probabilistic recoverable
resources estimations as depicted in Fig. 7.2 (Coll 2019). The choice of workflow is related to the
stage of project development (see Fig. 7.1):
• Probabilistic estimation of recoverable volumes for fields at the preproduction
(exploration/appraisal) phase
• Probabilistic estimation of recoverable volumes for fields with production history
7.6.1 Preproduction Phase. There are several steps to be executed in a probabilistic workflow
that are common to both preproduction and on-production phases. This section describes an
example of the methodology that can be used as part of a probabilistic workflow for fields without
production, as detailed below and in Fig. 7.16. Alternatively, if there has been no production,
and/or a dynamic model is not available, simpler Monte Carlo methodologies and/or ED can be
used instead as described in Section 7.5. The workflow below is based on the different
methodologies explained in detail in the previous sections. Actual procedures used will vary from
company to company.
1. Define ranges for the uncertain parameters: Ranges for the uncertain parameters should
be defined for each input uncertainty as explained in Section 7.2 during asset framework
sessions as part of an integrated multidisciplinary effort. Members of the subsurface team
should agree on the reservoir uncertainties that ought to be considered, their ranges and
the shapes of probability distributions to be used.
2. Use sensitivity analysis and/or ED methods: These methods can be used with the base
case static and dynamic reservoir models to identify key reservoir parameters and their
influence on the outcome (e.g., in-place estimates, TRR) as described in Sections 7.3.2
and 7.5.2. Tornado diagrams are used for sensitivity analysis to identify key parameters
as described in Section 7.3.2. ED will be used to define the cases to run where the
uncertainties are combined to explore the impact of interdependencies. It is important to
note that a single development scenario (base case development plan) should be used at
this stage (see Fig. 7.16). The outcome of this step is to identify the key reservoir
uncertainties with a “major impact” on the outcome as well as the effects of their
dependencies/interactions.
3. Perform risk analysis: This can be performed after the sensitivity analysis. This step
consists of performing a higher-order ED (e.g., if the response variable is nonlinearly
related to the parameters) to improve the RSM (Fig. 7.15) to account for parameter
interaction as described in Section 7.5.2.
4. Conduct MCS: These are run using proxy models to define the cumulative probability
distributions for the outcomes (e.g., resources). Monte Carlo analysis uses the proxy
model to generate the RSM solution (Fig. 7.14) as described in Section 7.5.2.
167
Fig. 7.16—General probabilistic workflow (after Wheaton and Coll 2010), where EA and GA are evolutionary and
genetic algorithms, respectively, and HM is history matched.
5. Estimate the P90, P50, and P10 recoverable resources: The probability distributions
derived in step 4 can be used to determine the P90, P50, and P10 estimates of resources.
Production forecasts can be derived for the resources ranges by running the simulation
models. The models and profiles are fully auditable and transparent.
6. Construct corresponding deterministic low, best, and high estimate models:
Corresponding deterministic scenarios can be derived based on the P10, P50, and P90
resource estimates from the previous step, as obtained from the CDFs. In most cases, the
input parameters used to derive the P90, P50, and P10 probabilistic estimates might need
to be recombined to select a realistic set of input parameters consistent with the PRMS
requirements. These production forecasts will also need to be processed through
commercial analysis to establish, among others, the project limit (truncated as necessary
by technical, contractual, economic, or other limits), so the quantities can be classified as
resources and/or reserves per the PRMS guidelines.
7.6.2 Production Phase. The workflow described in the previous section should be adapted for
fields that have enough production data to ensure production history is honored during the probabilistic
analysis. To achieve this goal, different methods can be used as described in Section 7.5.
To evaluate the effect of uncertainty, a recommended approach will be to generate three
different geological realizations corresponding to low, best, and high cases that represent the static
uncertainty. These models can then be transferred to a dynamic model and history matched.
However, in many simulation studies, only a best estimate history match (HM) might be available.
If this is the case, reservoir uncertainties should be evaluated during history matching by following,
for example, the workflow described in Section 7.6.1 Steps 1 and 2.
168
Multiple reservoir models can be generated using ED methods by varying the input reservoir
parameters in the dynamic simulation model. Global optimization methods can be used in Step 2
to speed up the HM process of the different model realizations compared to a manual HM. During
Step 3 in Section 7.6.1, many dynamic realizations can be run that take into account the ranges of
uncertainty in the input reservoir parameters. These models can be run simultaneously, converging
to multiple plausible models that produce a reasonable HM of production and pressure data. These
workflows are time-effective, with hundreds of potential solutions to the HM problem, recognizing
the reservoir uncertainties, generally made available in a matter of weeks. These methods compare
favorably with the traditional manual HM workflow, which is a time-intensive process that
produces a few matched models.
Step 4 can be skipped if enough simulation models exist. Cumulative distributions of TRR
can be generated using these HM models generated in Step 3 to determine P90, P50, and P10
resources estimates as part of Step 5 in Section 7.6.1. The matched models can be considered as
being equally probable in the absence of any other reliable assumptions (e.g., prior probabilities
and data errors). Alternatively, the degree of mismatch among the HM models can be used to guide
assigned probabilities. A further step (Step 6) is required for estimating resources once the input
variables to the P90, P50, and P10 simulation HM models are reviewed to generate the
corresponding deterministic 1P, 2P, and 3P Reserves or 1C, 2C and 3C Contingent Resources. If
input parameters need to be recombined, the HM of the new simulation model will need to be
evaluated to test the validity of the recombined assumptions. The corresponding deterministic
models can be used to generate P90, P50, and P10 forecasts suitable to be screened for their
commercial maturity as per the PRMS guidelines.
An advantage of probabilistic methods, as described in this section, compared to simple Monte
Carlo methods is that the P90, P50, and P10 models derived from the cumulative distribution can
be easily interrogated to determine input assumptions in a manner similar to that of deterministic
workflows. These methods provide an unbiased and rigorous analysis of the influence of reservoir
uncertainties on resources estimates, helping to evaluate and potentially mitigate risks during a
field development plan. Another advantage of these methods is that the asset team can have a
quantitative tool for evaluating the key reservoir uncertainties for a value of information analysis
to justify future data acquisition campaigns with the objective of reducing field development risk.
These methods allow determination of P90, P50, and P10 resources estimates that can be used with
confidence to represent the low, best, and high estimated resources cases.
A further advantage of these methods is that the production forecasts can be used to optimize
facilities design and/or field development scenarios and to evaluate the economic impact of the
uncertainty in resources.
Fig. 7.17 shows how the profiles resulting from a probabilistic workflow of 27 example cases
were screened economically to determine the uncertainty in NPV and the return on investment
(ROI) compared to the internal ROI company threshold (in this example, ROI = 12%) for
project approval.
Examination of the economic results indicated that all cases had a positive net ROI, but for
some cases, the net ROI was below or marginally above the internal ROI (see cases 7, 8, 16, 17,
25, 26, and 27). The integrated team looked at these lower ROI cases and the combination of
uncertainties that generated these results to find opportunities to mitigate the downside through
further optimization of the development plan and/or by acquiring new information. Cases with the
highest ROI (e.g., cases 1, 2, 5, 10, 11, and 19) were also reviewed by the team to evaluate the
combination of uncertain parameters that contributed to generate these upsides. These results
169
allowed the team to evaluate the key development risks and opportunities that justified some
changes to the initial development plan. Another outcome from this analysis was a cumulative
distribution of the ROI to estimate the P90, P50, and P10 ROIs. The P10, P50, and P90 ROIs
shown in the table within Fig. 7.17 were communicated to management to convey the economic
uncertainty and investment risk during final investment decision (FID) project approval.
Fig. 7.17—Probabilistic reserves estimations (after economics) used for optimizing facility design.
A disadvantage of probabilistic methods is that these methods can be computationally

intensive, requiring a large number of numerical reservoir models to be run, which may not be
practical for some fields (for example, in multimillion grid cell simulation models with long
production histories). The evaluator should weigh the benefits that these methods can bring in
assessing risk during development for each field vs. the associated time and computing cost.
7.7 Consistency between Probabilistic and Deterministic Methods

Probabilistic and deterministic estimates results need to be analyzed for consistency (Purvis and
Strickland 2001). Once probabilistic recoverable volumes have been estimated, such as by
following the workflows explained in the previous section, the corresponding deterministic
recoverable volumes can be estimated following the guidance provided by Wheaton and Coll
(2010) and Coll (2019). The PRMS guidelines (PRMS § 4.2.3.3) state that P10, P50, and P90
resources estimates should reconcile with deterministically derived low, best, and high estimate
cases. If this is not the case, an evaluation of the reasons for discrepancies must be undertaken to
reconcile and update the results as needed.
The following discussion assumes, for simplicity, that all estimated quantities satisfy the
requirements for Reserves status; hence, the discussion will focus on 1P/2P/3P and P90/P50/P10
nomenclature. However, the same commentary and process apply for Resources in general.
Under the PRMS guidelines, a low estimate corresponding to Proved Reserves (1P) should
meet several criteria, including that the area considered as Proved involves (1) the area delineated
by drilling and defined by fluid contacts, if any, and (2) adjacent undrilled portions of the reservoir
that can reasonably be judged as continuous with it and commercially productive on the basis of
available geoscience and engineering data. Fluid contact assumptions must be scrutinized. It is a
170
requirement that, in the absence of encountering fluid contacts, Proved quantities should be limited
by the lowest known hydrocarbon as seen in a well penetration unless otherwise indicated by
definitive geoscience, engineering, or performance data. Another criterion to consider could be the
correspondence of the P50 case with a P50 in-place volume.
Fig. 7.18a illustrates the response function R (in this case, recoverable volumes) against two
input uncertain parameters (parameter 1, parameter 2). It also shows the surfaces and contour lines
indicating P90, P50, and P10 recoverable volumes estimated following the workflow depicted in
Fig. 7.16. If the R function were to be created by considering parameter interaction terms, then the
solution space R would be a more complicated surface (Fig. 7.14). Fig. 7.18 also shows the actual
simulation run results from Section 7.6.1 Step 2 (green dots). It could be possible that none of the
green dots at a P50 probability is consistent with the P50 in-place volume that could be expected
from the PRMS guidance, but rather, they could correspond to a P90 or P10 in-place volume. It is
possible that none of the cases at the P90 confidence level satisfies the requirements for
Proved Reserves.
If production data exist, the solution space R will be restricted to the orange area as shown in
Fig. 7.18b. Simulations outside the solution space of valid HM should be discarded. Cumulative
distributions of recoverable volumes should be estimated using only the valid HM cases.
Fig. 7.18—(a) R space solution (reserves) showing the corresponding P90, P50, and P10 recoverable volumes as surfaces
and contours, with their input parameter values, where symbols indicate simulation results, and (b) R space solution
constrained by considering only models that match history (orange shaded area, the constrained space solution) (after
Wheaton and Coll 2010).
Table 7.3 shows an example of the input parameters for three runs corresponding to P90, P50,
and P10 probabilities (green dots in Fig. 7.18a). As tabulated, the P90 case has a GWC contact
that is deeper than the P50 case. In this example, the GWC for the P90 and P10 cases can be
switched to ensure consistency with the guidelines. This change will require other parameters to
be adjusted to ensure original gas-in-place (OGIP) and TRR are not altered. As observed, the P90
OGIP value is as expected for the low case. Once the parameters are changed, the new runs should
produce realistic production profiles with the same ranges of recoverable volumes as the original
analysis (blue dots in Fig. 7.18a). The process of recombining input variables to comply with the
PRMS guidelines is an iterative process to generate reasonable 1P, 2P, and 3P profiles.
171
Table 7.3—Example input and output parameters for P90/P50/P10 recoverable volumes used for 1P/2P/3P Reserves
(after Wheaton and Coll 2010).
7.8 Commercial Considerations

The outcomes of the probabilistic workflows described herein are technical estimates of
recoverable resources that will need to undergo an economic evaluation. As stated in the PRMS (§
2.1.2.1), the commerciality requirements include that “the entity satisfied the internal decision
criteria (typically rate of return at or above the weighted average cost-of-capital or the hurdle
rate).” As a result of the economic and subsequent commercial assessments, sales products will be
estimated that take into account the PRMS commerciality requirements described in PRMS § 2.1.2
and § 3.1. It is the assessment including economic criteria (especially if the assessment includes
production sharing contract (PSC) terms and conditions, and a given set of commodity prices and
OPEX/CAPEX profile) that will result in key differences between the TRR volumes obtained from
probabilistic approaches and the resources scenarios that can be used for project valuation.
Probabilistic workflows will typically contain hundreds of reservoir realizations that represent
the influence of subsurface uncertainties on resources estimates and economic value. The amount
of information provided by probabilistic workflows allows entities to better assess economic risk
associated with investment in a field development plan (as discussed in Section 7.6) to guide
investment decisions. Another benefit of probabilistic workflows during commercial evaluation is
that results of the economic analysis can be used to justify the acquisition of new data, provided
the input for a value of information analysis is aimed to reduce not only the reservoir uncertainty
but also the investment risk.
The P90, P50, and P10 scenarios obtained from the probabilistic workflows after commercial
evaluation can be used to support public disclosures for project valuation. Corresponding
deterministic 1P, 2P, and 3P cases as described in Section 7.7 can be used after commercial
evaluation for regulatory reporting. One other benefit of probabilistic methods is that the
confidence levels associated with the corresponding deterministic scenarios are fully supported by
the probabilistic workflows, making possible a rigorous analysis of the rationale behind the
selection of the reserves and resources estimates at each confidence level.
172
7.9 Final Considerations
This chapter provides technical guidance on various probabilistic methods and workflows that can
be applied at different stages of field development. Several methodologies have been used in the
oil and gas industry with the objective of promoting integrated probabilistic workflows. These
methods provide powerful mathematical tools with which to assess the influences that reservoir
uncertainties have on resources estimates.
Advances in technology and computing capacity have allowed these methods to evolve,
addressing many of the concerns that evaluators have regarding lack of transparency on the
parameters used in the probabilistic models. There are significant potential variations in estimates
of P90, P50, and P10 TRR (and eventually the related reserves), which can be due to the selection
of unsuitable distributions of the uncertainty parameters that may lead to misleading conclusions.
A careful selection of the uncertainty ranges and probability distributions is important to the
successful application of any probabilistic method.
As explained in this chapter, the method selected should be dependent on the stage of the field
development and available information as summarized in Table 7.4. As observed at the
exploration stage, sensitivity analysis and MCS are widely used and are sufficient to evaluate
uncertainty, whereas ED is recommended for preproduction fields at the appraisal stage where
more data and information exist and dynamic reservoir models have been built, justifying the use
of more sophisticated workflows. If the field is at the production stage, and production data need
to be history matched, global optimization methods are recommended. Deterministic methods such
as scenario analysis can be used for all the pre- and on-production stages as well as at different
field phases.
Table 7.4—Preproduction and on-production recommended workflows (Coll 2019). EA and GA are evolutionary and genetic
algorithms, respectively.
The use of probabilistic estimates for recoverable resources without also considering the
corresponding deterministic cases as explained in Section 7.7 can compromise the validity of the
probabilistic results if the parameters selected for the P90, P50, and P10 cases are inconsistent with
PRMS requirements (e.g., fluid contacts). Also, P90, P50, and P10 estimates must be recognized
as TRR that need to satisfy economic and commercial criteria before they can be classified as
Reserves under the PRMS guidelines.
173
7.10 References
Aanonsen, S. I., Nævdal, G., Oliver, D. S. Et al. 2009. The Ensemble Kalman Filter in
Reservoir Engineering—A Review. SPE J. 14 (03): 393–412. SPE-117274-PA.
https://doi.org/10.2118/117274-PA.
Abdollahzadeh, A., Christie, M., Corne, D. et al. 2013. An Adaptive Evolutionary Algorithm for
History-Matching. Paper presented at the EAGE Annual Conference & Exhibition
incorporating SPE Europec, London, UK, 10–13 June. SPE-164824-MS.
Acuña, H. G. and Harrell, D. R. 2000. Adapting Probabilistic Methods to Conform to Regulatory
Guidelines. Paper presented at the SPE Annual Technical Conference and Exhibition,
Dallas, Texas, USA, 1–4 October. SPE-63202-MS. https://doi.org/10.2118/63202-MS.
Al-Shamma, B. and Teigland R. 2006. History Matching of the Valhall Field Using a Global
Optimization Method and Uncertainty Assessment. Paper presented at the SPE Annual
Technical Conference, San Antonio, Texas, USA, 24–27 September. SPE-100946-MS.
https://doi.org/10.2118/100946-MS.
Annan, E. B., Aidoo, A. B., Ejeh, C. J. et al. 2020. Mapping of Porosity, Permeability and
Thickness Distribution: Application of Geostatistical Modeling for the Jubilee Oilfield in
Ghana. Koforidua, Ghana: Department of Oil and Gas Engineering, All Nations
University College.
Arroyo-Negrete, E., Devegowda, D., Datta-Gupta, A. et al. 2008. Streamline-Assisted Ensemble
Kalman Filter for Rapid and Continuous Reservoir Model Updating. SPE Res Eval &
Eng 11 (06): 1046–1060. SPE-104255-PA. https://doi.org/10.2118/104255-PA.
Carter, P. J. and Morales, E. 1998. Probabilistic Addition of Gas Reserve within a Major Gas
Project. Paper presented at the SPE Asia Pacific Oil and Gas Conference and Exhibition,
Perth, Australia, 12–14 October. SPE-50113-MS. https://doi.org/10.2118/50113-MS.
Castellini, A., Vahedi, A., Singh, U. et al. 2008. Reconciling History Matching and Assessment
of Uncertainty in Production Forecasts: A Study Combining Experimental Design, Proxy
Models and Genetic Algorithms. Paper presented at the International Petroleum
Technology Conference, Kuala Lumpur, Malaysia, 3–5 December. IPTC-12745-
ABSTRACT. https://doi.org/10.2523/IPTC-12745-ABSTRACT.
Cheng, H., Dehghani, K., and Billiter, T. 2008. A Structured Approach for Probabilistic Assisted
History Matching Using Evolutionary Algorithms: Tengiz Field Applications. Paper
presented at the SPE Annual Technical Conference and Exhibition, Denver, Colorado,
USA, 21–24 September. SPE-116212-MS. https://doi.org/10.2118/116212-MS.
Cheong, Y. P. and Gupta, R. 2005. Experimental Design and Analysis Methods for Assessing
Volumetric Uncertainties. SPE J. 10 (03): 324–335. SPE-80537-PA.
Choudhary, M. K., Soon, S., and Ludvigsen, B. E. 2007. Application of Global Optimization
Methods in History Matching and Probabilistic Forecasting—Case Studies. Paper
presented at the SPE Middle East Oil & Gas Show and Conference, Bahrain 11–14
March. SPE-105208-MS. https://doi.org/10.2118/105208-MS.
Coll, C. 2019. Application of Probabilistic and Deterministic Methods for Consistent Reserves
and Resources Estimation and Reporting. Paper presented at the 81st EAGE Annual
Conference & Exhibition, London, UK, 3–6 June. SPE-195465-MS.
https://doi.org/10.2118/195465-MS.
Coll, C. and Elliott, S. 2013. Probabilistic and Deterministic Methods: Applicability in
Unconventional Reservoirs. Paper presented at the 75th EAGE Annual Conference &
174
Exhibition, London, UK, 10–13 June. SPE-164820-MS. https://doi.org/10.2118/164820-
MS.
El-Khatib, N. 1999. Waterflooding Performance of Communicating Stratified Reservoirs with
Log-Normal Permeability Distribution. SPE Res Eval & Eng 02 (06): 542–549. SPE-
59071-PA. https://doi.org/10.2118/59071-PA.
Elvind, D., Asmund, H., and Rolf, V. 1992. Maximum Information at Minimum Cost: A North
Sea Field Development Study with an Experimental Design. J Pet Technol 44 (12):
1350–1356. SPE-23139-PA.
Everitt, B. S. and Skrondal, A. 2010. Cambridge Dictionary of Statistics, fourth edition.
Cambridge, UK: Cambridge University Press.
Fetel, E. and Caumon, G. 2008. Reservoir Flow Uncertainty Assessment Using Response
Surface Constrained by Secondary Information. J Pet Sci Eng 60 (3–4): 170–182.
https://doi.org/10.1016/j.petrol.2007.06.003.
Ghamdi, A., Ganis, S., and Hammad, K. 2020. Evaluation of Ensemble Smoother with Multiple
Data Assimilation and Evolutionary Algorithm for History Matching Process
Optimization. Paper presented at the SPE Asia Pacific Oil & Gas Conference and
Exhibition, 17–19 November. SPE-202216-MS. https://doi.org/10.2118/202216-MS.
Hall, R., Bertram, R., Gonzenbach, G. et al. 2020. Guidelines for the Practical Evaluation of
Undeveloped Reserves in Resource Plays, Vol. 3. Houston, Texas, USA: Monograph
Series, Society of Petroleum Evaluation Engineers.
Hooke, R. and Jeeves, T. A. 1961. Direct Search Solution of Numerical and Statistical Problems.
J ACM 08 (02): 212–229. https://doi.org/10.1145/321062.321069.
Kabir, C. S., Chawathe, A., Jenkins, S. D. et al. 2002. Developing New Fields Using
Probabilistic Reservoir Forecasting. Paper presented at the SPE Annual Technical
Conference and Exhibition, San Antonio, Texas, USA, 29 September–2 October. SPE-
77564-MS. https://doi.org/10.2118/77564-MS.
Komlosi, Z. P. and Komlosi, J. 2009. Application of the Monte Carlo Simulation in
Calculating HC-Reserves. Paper presented at EUROPEC/EAGE Annual Conference
and Exhibition, Amsterdam, The Netherlands, 8–11 June. SPE-121256-MS.
https://doi.org/10.2118/121256-MS.
Lorentzen, R. J., Naevdal, G., Valles, B. et al. 2005. Analysis of the Ensemble Kalman Filter for
Estimation of Permeability and Porosity in Reservoir Models. Paper presented at the SPE
Annual Technical Conference and Exhibition, Dallas, Texas, USA, 9–12 October. SPE-
Manceau, E., Mezhani, M., Zabalza-Mezghani, I. et al. 2001. Combination of Experimental
Design and Joint Modeling Methods for Quantifying the Risk Associated with
Deterministic and Stochastic Uncertainties—An Integrated Test Study. Paper presented at
the SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana, USA, 30
September–3 October 3. SPE-71620-MS. https://doi.org/10.2118/71620-MS.
Minin, A., Guerra, L., and Colombo, I. 2011. Unconventional Reservoirs Probabilistic Reserve
Estimation Using Decline Curves. Paper presented at the International Petroleum
Technology Conference, Bangkok, Thailand, 15–17 November. IPTC-14801-MS.
https://doi.org/10.2523/IPTC-14801-MS.
Mishra, S. and Datta-Gupta, A. 2017. Applied Statistical Modeling and Data Analytics. A
Practical Guide for the Petroleum Geosciences. Amsterdam, The Netherlands: Elsevier.
175
Morales, E. and Lee, W. J. 2014. Probabilistic Reserves Addition Within a Project. SPE Econ &
Mgmt 07 (01): 16–21. SPE-166153-PA.
Ogbeiwi, P., Stephen, K. D. and Arinkoola, A. O. 2020. Robust Optimisation of Water
Flooding Using an Experimental Design-Based Surrogate Model: A Case Study of
a Niger-Delta Oil Reservoir. J Pet Sci Eng 195 (December): 107824.
https://doi.org/10.1016/j.petrol.2020.107824.
USA: Society of Petroleum Engineers.
Purvis, D. C. and Strickland, R. F. 2001. Problems Reconciling Probabilistic and Deterministic
Reserve Classifications and Evaluations. Paper presented at the SPE Hydrocarbon
Economics and Evaluation Symposium, Dallas, Texas, USA, 2–3 April. SPE-68591-MS.
https://doi.org/10.2118/68591-MS.
Sanghyun, L. and Stephen, K. D. 2018. Optimizing Automatic History Matching for Field
Application Using Genetic Algorithm and Particle Swarm Optimization. Paper presented
at the Offshore Technology Conference Asia, Kuala Lumpur, Malaysia, 20–23 March.
OTC-28401-MS. https://doi.org/10.4043/28401-MS.
Schulze-Riegert, R. W., Diab, A., and Griess, B. K. 2006. Application of Global Optimization
Techniques for Model Validation and Prediction Scenarios for a North African Oil Field.
Paper presented at the SPE Europec/EAGE Annual Conference and Exhibition, Vienna,
Austria, 12–15 June. SPE-100193-MS. https://doi.org/10.2118/100193-MS.
Schulze-Riegert, R. W., Krosche, M., Fahimuddin, A. et al. 2007. Multi-Objective Optimization
with Application to Model Validation and Uncertainty Quantification. Paper presented at
the SPE Middle East Oil and Gas Show and Conference, Bahrain, 11–14 March. SPE-
Stephen, K. D. 2018. Assisted Seismic History Matching of the Nelson Field: Managing Large
Numbers of Unknowns by Divide and Conquer. J Pet Sci Eng 171 (December): 1231–
1248. https://doi.org/10.1016/j.petrol.2018.07.055.
Twartz, S. K., Gorjy, F., and Milne, I. G. 1998. A Multiple Realisation Approach to Managing
Uncertainty in the North Rankin Gas Condensate Field, Western Australia. Paper
presented at the SPE Asia Pacific Oil and Gas Conference and Exhibition, Perth, Western
Australia, Australia, 12–14 October. SPE-50078-MS. https://doi.org/10.2118/50078-MS.
Vahedi, A., Gorgy, F., Scarr, K. D. et al. 2005. Generation of Probabilistic Reserves
Distributions from Material Balance Models Using an Experimental Design
Methodology. Paper presented at the International Petroleum Technology Conference,
Doha, Qatar, 21–23 November. IPTC-11009-MS. https://doi.org/10.2523/IPTC-11009-
MS.
Van Elk, J. F., Guerrera, L., Vijayan, K. et al. 2000. Improved Uncertainty Management in Field
Development Studies through the Application of the Experimental Design Method to the
Multiple Realisations Approach. Paper presented at the SPE Asia Pacific Oil and Gas
Conference and Exhibition, Brisbane, Australia, 16–18 October. SPE-64462-MS.
https://doi.org/10.2118/64462-MS.
Wadsley, A. W. 2005. Markov Chain Monte Carlos Methods for Reserve Estimation. Paper
presented at International Petroleum Technology Conference held in Doha, Qatar, 21–23
November. IPTC-10065-MS. https://doi.org/10.2523/IPTC-10065-MS.
Wheaton, R. and Coll, C. 2010. Reserves Estimations under New SEC 2009 Rules when
Using Probabilistic Methods. Paper presented at SPE EUROPEC/EAGE Annual
176
Technical Conference and Exhibition, Barcelona, Spain, 14–17 June. SPE-131241-MS.
https://doi.org/10.2118/131241-MS.
White, C. D. and Royer, S. A. 2003. Experimental Design as a Framework for Reservoir Studies.
Paper presented at the SPE Reservoir Simulation Symposium, Houston, Texas, USA, 3–5
February. SPE-79676-MS. https://doi.org/10.2118/79676-MS.
Wikipedia. 2022a. Expected Value (4 June 2022 revision).
https://en.wikipedia.org/wiki/Expected_value (accessed 6 June 2022).
Wikipedia. 2022b. Global Optimization (25 February 2022 revision).
https://en.wikipedia.org/wiki/Global_optimization (accessed 6 June 2022).
Zhang, J., Delshad, M., and Sepehrnoori, K. 2007. Development of a Framework for
Optimization of Reservoir Simulation Studies. J Pet Sci Eng 59 (1–2): 135–146.
https://doi.org/10.1016/j.petrol.2007.03.013.
177
7.11 Glossary and Definitions
In probability theory and statistics, a cross-covariance matrix is

a matrix in which the element in the (i, j) position is the covariance
Cross- between the ith element of a random vector (uncertainty) and the jth
Covariance element of another random vector (uncertainty). A random vector is
Matrix a random variable with multiple dimensions. Each element of the
vector is a scalar random variable. A covariance matrix measures the
degree to which two variables are linearly associated.
A cumulative distribution function (CDF) of a real-valued random
Cumulative variable X, or just distribution function of X, evaluated at x, is the
Distribution probability that X will take a value less than or equal to x. Cumulative
Function (CDF) distribution functions are also used to specify the distribution of
multivariate random variables.
Optimization procedures motivated by biological analogies. The
primary idea is to try to mimic the “survival of the fittest” rule of genetic
mutation in the development of optimization algorithms. The process
begins with a population of potential solutions to a problem and a way
Genetic of measuring the fitness or value of each solution. A new generation
Algorithms of solutions is then produced by allowing existing solutions to
“mutate” (change a little) or cross over (two solutions combine to
produce a new solution with aspects of both). The aim is to produce
new generations of solutions that have higher values (from Everitt
and Skrondal 2010).
A branch of applied mathematics and numerical analysis that
attempts to find the global minima or maxima of a function or a set
Global of functions on a given set. It is usually described as a minimization
Optimization problem because the maximization of the real-valued function g(x) is
equivalent to the minimization of the function f(x) := (-1)*g(x).
(Wikipedia 2022b)
A recursive procedure that provides an estimate of a signal when
only the “noisy signal” can be observed. The estimate is effectively
constructed by putting exponentially declining weights on the past
Kalman Filter
observations, with the rate of decline being calculated from various
variance terms. Used as an estimation technique in the analysis of
time-series data (from Everitt and Skrondal 2010).
The generalization of a Latin square to an arbitrary number of
dimensions, whereby each sample is the only one in each axis-
aligned hyperplane containing it. A stratified random sampling
technique in which a sample of size N from multiple (continuous)
Latin Hypercube variables is drawn such that for each individual variable, the sample
is (marginally) maximally stratified, where a sample is maximally
stratified when the number of strata equals the sample size N and
when the probability of falling in each of the strata equals N − 1 (from
Everitt and Skrondal 2010).
178
The expected value of a random variable X, denoted E(X) or E[X], is
a generalization of the weighted average, and it is intuitively the
arithmetic mean of a large number of independent realizations of X
(Wikipedia 2022a). The expected value is also known as the
expectation, mathematical expectation, mean, average, or first
Mean, Expected moment. It is the average value over the entire probability range,
Value or weighted by the probability of occurrence:
Expectation n
∑ ∫
Mean = xi ⋅ P( xi ) or x ⋅ P( x) ⋅ d ( x) ,
i =1
where x = variable (e.g., reserve value) and P(x) = probability of x.
The mean of statistical distributions can be added arithmetically in
aggregation.
The simple values that give us information about the distribution of
data such as arithmetic mean, median, mode. The three measures of
Measures of
centrality defined below coincide only when probability distribution
Centrality
functions are symmetrical. This is seldom the case for reserves. In
general, and for most practical purposes, they differ.
The median is the value (middle value) separating the higher half
from the lower half of a data sample, a population, or a probability
distribution; therefore, the probability that the outcome will be higher
Median (also is equal to the probability that it will be lower. The benefit of using the
known as P50) median compared to the mean ("average") is that it is not skewed by
a small proportion of extremely large or small values, and therefore it
provides a better representation of a "typical" value. The median is
often used for resources best estimates.
The mode is the value that appears most often in a set of data values
or most probable value. If X is a discrete random variable, then the
Mode, or Most mode is the value x at which the probability mass function takes its
Probable Value maximum value. In other words, it is the value that is most likely to be
sampled. It is the reserves quantity where the probability distribution
function has its maximum value.
Procedures for finding the maxima or minima of functions of,
generally, several variables. It is most often encountered in statistics
Optimization
in the context of maximum likelihood estimation, where such
Methods
methods are frequently needed to find the values of the parameters
that maximize the likelihood (from Everitt and Skrondal 2010).
A designed experiment is orthogonal if the effects of any factor
balance out (sum to zero) across the effects of the other
Orthogonal
factors. Orthogonality guarantees that the effect of one factor or
Design
interaction can be estimated separately from the effect of any other
factor or interaction in the model.
In orthogonal sampling, the sample space is divided into equally
Orthogonal
probable subspaces. All sample points are then chosen
Sampling
simultaneously, making sure that the total set of sample points is a
179
Latin hypercube sample and that each subspace is sampled with the
same density.
The 10th percentile (possible) is the highest figure; it means that 10%
of the calculated estimates will equal or exceed the P10 estimate. For
P10 reserves, it means the quantity for which there is a 10% probability
that the quantities actually recovered will equal or exceed the
estimate.
The 50th percentile (the median) is the score below which 50%
(exclusive) or at or below which (inclusive) 50% of the scores in the
P50, or Median distribution may be found. For reserves, it means the quantity for
which there is a 50% probability that the reserves actually recovered
will equal or exceed the estimate.
The 90% percentile (Proved) is the lowest figure; it means that 90%
of the calculated estimates will be lower. In contrast, for percentiles,
a percentage is given, and a corresponding score is determined,
which can be either exclusive or inclusive. The score for a specified
P90
percentage (e.g., 90th) indicates a score below which (exclusive
definition) or at or below which (inclusive definition) other scores in
the distribution fall. In reserves estimation, this is the number quoted
as the proven value.
A percentile (or a centile) is a score below which a given percentage
of scores in its frequency distribution fall or a score at or below which
a given percentage fall (inclusive definition). For example, the 50th
Percentiles
percentile (the median) is the score below which 50% (exclusive) or
at or below which (inclusive) 50% of the scores in the distribution may
be found.
Type of experimental design (ED) that helps to find the factors in an
experiment that are important. This design assumes that the
interactions between factors (uncertainties) can be completely
Plackett-Burman ignored, and the main effects can be calculated with a reduced
Design number of experiments (reservoir models with combinations of
uncertainties). This means large amounts of data do not need to be
collected in experiments (reservoir models with combinations of
uncertainties) on relatively unimportant factors.
A polynomial function is a function that can be expressed in the form
of a polynomial. The definition can be derived from the definition of a
Polynomial polynomial equation. A polynomial is generally represented as P(x).
Function The highest power of the variable of P(x) is known as its degree. The
degree of a polynomial function is very important because it tells
us about the behavior of the function P(x) when x becomes very large.
Probability is the branch of mathematics concerning numerical
descriptions of the likelihood of an event to occur, or the likelihood
Probability that a proposition is true. The extent to which an event is likely to
occur is measured by the ratio of the number of occurrences to the
whole number of cases possible. Note that the probability used in
180
reserves estimation is a subjective probability, quantifying the
likelihood of a predicted outcome.
A probability density function (PDF), or density of a continuous
Probability random variable, is a function for which the value at any given sample
Density (or point) in the sample space (the set of possible values taken by the
Function (PDF) random variable) can be interpreted as providing a relative likelihood
that the value of the random variable would equal that sample.
Set of all possible points of the choice variables of an optimization
Solution Space problem that satisfy the problem's constraints, potentially
including inequalities, equalities, and integer constraints.
The standard deviation is a measure of the amount of variation
or dispersion of a set of values around its mean. A low standard
Standard deviation indicates that the values tend to be close to the mean
Deviation (expected value), whereas a high standard deviation indicates that
the values are spread out over a wider range. It is calculated as the
square root of the variance.
Variance is a measure of how far a set of numbers is spread out from
their average and is the expectation of the squared deviation of
a random variable from its mean. The variance is calculated by
adding the square of the difference between values in the distribution
and the mean value and calculating the arithmetic average:
Variance
∑ ( x −=
µ)
n
b
∫ (x − µ)
2 i i 2
s
= i =1
f ( x)dx ,
n a
where x = variable (e.g., reserves), µ = mean, and f(x) = probability

density function. It is convenient to square the differences because
this avoids the cancelling of positive and negative values.
181
Chapter 8
Aggregation of Reserves
and Resources
William J. Haskett (Chair)
Tyler Schlosser
8.1 Introduction
In reserves and resources assessment, technically recoverable quantities are typically based on the
results of analog analyses, performance evaluations, and/or volumetric calculations for individual
wells, reservoirs, or portions of reservoirs. These estimates may be combined to arrive at
recoverable quantities for fields, properties, projects, and portfolios. The uncertainty of the
combined estimates at each of these levels may differ greatly, depending on such factors as
geological setting and maturity of the resource. This cumulative process, which takes into account
the uncertainty, is referred to as “aggregation.” Appropriate aggregation methods enable efficient
decision and valuation assessments for both conventional and unconventional assets.
Resource estimates may be stated as single deterministic values, range estimations (such as
those derived probabilistically), or aggregated collections of producing or nonproducing assets. The
aggregation method used should be associated with the purpose of the assessment. The assessment
of discovered resources is usually capable of being combined within the 1P, 2P, and 3P Reserves (or
1C, 2C, and 3C Contingent Resources) categories as outlined in Chapter 2—Petroleum Resources
Definitions, Classification, and Categorization Guidelines. Simple summation and range
aggregation (using probabilistic methods) are both appropriate paths to a valid resource estimate;
however, that validity and therefore the aggregation method used are contingent on the purpose of
the aggregation and the inherent uncertainty of the components being aggregated.
Oil and gas companies considering long-term development and performance of their assets
likely will use an assessment of volumes for investment, development path, infrastructure, and
competitive advantage decisions. The Petroleum Resources Management System (PRMS 2018, §
2.2.2.10) states that “for project justification, it is generally the best-estimate Reserves or
Resources quantity that passes qualification” and that (§2.1.3.7.2) the “best estimate (or P50)
production forecast is typically used for the economic evaluation for commerciality assessment of
the project.” Consequently, the PRMS equates “Best Estimate” to the P50 or median result. It is
recognized that this term is not used consistently across industry, and, in many locales, it is avoided
entirely. The reader is advised that when the term is used within the PRMS and this document, it
refers to the median or P50 result.
Of late, some companies have been moving to a more confidence-based approach to resources
estimation involving an aggregation process that takes into account the full range of potential
outcomes. They work on the assumption that the maximization of their confidence of making the
correct development and infrastructure decisions is the efficient path to resource assessment and
development, and ultimately the best value of their uncertainty-filled portfolios will be realized.
182
For multiple reservoir assessments, decision-based evaluation can seem to conflict with the
recommended levels of certainty associated with 1P, 2P, 3P and 1C, 2C, 3C categories. For
example, what we are at least 90% certain will be achieved from a collection of producing horizons
is never less than the arithmetic sum of the 1P components. When uncertain, ranged potential is
aggregated across elements of a project or portfolio, our 90% reserves confidence occurs at a
reserves tally higher than the simple sum of the 1P Reserves. We will illustrate this aspect several
times throughout this chapter and provide guidelines as to how to create, assess, and communicate
aggregated results.
Although there are several valid reasons to aggregate volumes, and different aggregation
methods have different suitability, there are a few overarching considerations specific to reserves
reporting. When considering aggregation of reserves for reporting purposes, reserves classes
should not be directly aggregated. The PRMS states (§ 2.2.2.5): “Quantities in different classes
and sub-classes cannot be aggregated without considering the varying degrees of technical
uncertainty and commercial likelihood involved with the classification(s) and without considering
the degree of dependency [relationship] between them.” The PRMS also states (§ 4.2.5.4): “It is
recommended that for reporting purposes, assessment results should not incorporate statistical
aggregation beyond the field, property, or project level. Results reported beyond this level should
use arithmetic summation by category but should caution that the aggregate Proved may be a very
conservative estimate and aggregate 3P may be very optimistic, depending on the number of items
in the aggregate.” This chapter will provide detailed discussion of the latter point.
While it is vital to retain the 1P confidence definition on a single assessed project, the link
from qualified deterministic calculation to resource confidence on a multiple-reservoir collection
is broken in aggregations due to the uncertainty within the contributing reservoirs. In the case of
multiple reservoirs, fields, project areas, or any sort of portfolio, the arithmetic tally and
assessment considerations of “what has been found” will differ from that of “what do we expect
from the portfolio.”
The resource assessment approach taken will depend on the intended use of the results. On
individual reservoirs, it is best practice to create and report reserves estimates as a range (1P/2P/3P)
or, in the case of Contingent Resources, 1C/2C/3C. Where assessments may be based on
deterministic methods, summations are typically (but not always) arithmetic and by category.
Where probabilistic assessments are available, companies may aggregate probabilistically to the
field/property/project level or, for internal use, portfolio level and then use a P10, P50, P90
characterization for the confidence levels within the ranged potential. It is a mistake to assume the
P90 aggregate result is equal to the sum of the individual 1P Reserves in multiple or single
reservoir evaluations. For internal portfolio analyses, companies may use fully probabilistic
methods given that they also include shared chance and correlation as components of
the assessment.
Investors, analysts, and utilities independent of development decisions will usually require a
high level of certainty and concentrate on the Proved (1P) volumes, or to a lesser extent, the
Proved-plus-Probable (2P) volumes. Development and infrastructure investors, wishing to ensure
efficient right-sizing of hard assets, will want to understand the aggregated reserves taking into
consideration uncertainty across the components. Long-term gas contracting is sometimes based
on Proved-plus-Probable Reserves, where there is a large gas resource that is most economically
developed over the life of the gas contract.
Accountants may use the ratio of production to Proved Developed Reserves or other reserves
categories as the basis for depreciating or depleting the cost of acquiring and developing reserves
183
over time as the reservoirs are produced. In some areas, the ratio of production to Proved-plus-
Probable Reserves (including any Undeveloped Reserves) is used as the basis for depreciation. For
these calculations, accountants require the reserves to be assessed at the level at which the
investments apply. Accounting looks at what has been done and what has been found. As such,
preserving 1P and 2P tallies is important across an aggregation, but it should be recognized that
simple arithmetic 1P sums will have a high probability of underestimating what will ultimately be
produced.
Section 8.2 addresses some general technical issues in resources aggregation. The discussion
on the aggregation of resources also addresses the issue that the uncertainty of the sum of volumes
will be less than the sum of the uncertainties of the individual volumes. In other words, the
uncertainty decreases with an increasing number of independent units available. The implications
of the resulting uncertainty reduction in a diverse portfolio, also called the portfolio effect, will be
discussed in Section 8.3.
Section 8.4 discusses regional aggregation over resources categories, and the use of scenario
methods for reserves aggregation is shown in Section 8.5, followed by Section 8.6, which
summarizes the chapter in a few simple guidelines.
8.1.1 Defining Reservoir Relationships. As we aggregate resource estimates from multiple

sources, we need to recognize their commonalities and shared assessment components. Collections
of reservoirs will have varying degrees of similarity in such elements as depositional character,
diagenetic history, economic thresholds, and evaluation bias, which will influence any aggregated
assessment.
As we start to work with concepts that relate to a range of possible outcomes, the general
naming guideline is to stick with definitive terms in reports and communication. Many companies
try to avoid the use of ambiguous terms and be explicit about what they mean. Terms are important,
but as long as there is clarity on the definitions, resources and reserves estimates may be calculated
and communicated effectively. However, we recognize definitions vary, so we need to be clear as
to their use within the PRMS.
Looking at the statistical terms, the Expected Reserves is the Mean reserves amount. While
statisticians may differ as to the pure definition of “Mean” vs. that of “average,” for our purposes,
Mean, Expected and average will be used interchangeably when describing a continuous
distribution. We shall use “P50” or “Median” to describe that point in an uncertainty distribution
that will have 50% of the possible outcomes above and 50% below that number.
There are two types of inter-reservoir relationships that must be considered in order to arrive
at a valid and reliable aggregation result: Correlation and Dependency. We need to be clear on
aggregated Correlation and Dependency because it is critical to keep the two methods separate for
valid aggregation. The two topics are discussed in Sections 8.2.3 and 8.2.4, respectively.
Nearby reservoirs that have not been drilled may have shared chance. “Chance” refers to the
collection of successful elements needed to be present in order to have an accumulation. While
specific chance elements may differ between organizations, they typically belong to one of two
categories, container (reservoir, vertical seal, horizontal seal, seal capacity) or contents (source,
migration, timing, diagenetic history). The presence or absence of a major chance element in one
location that changes the probability of success assessed for a second location indicates a
dependent relationship between the probabilities of success for the two potential reservoirs. In this
chapter, we will refer to this shared chance relationship as Dependency. We capitalize
“Dependency” to distinguish the shared chance source from the colloquial “dependency,” which
184
is used to denote a nonspecific relationship between two items or event. Dependency, as shared
chance, is causative. Similarly, we shall use uppercase to denote the names for statistical positions
along the uncertainty distribution (principally Mean, Median, and Mode).
In this chapter, we will refer to the similarity in outcomes of range-based elements of different
reservoirs as Correlation. We recognize reserves uncertainty in our 1P-2P-3P categorization, but
more typically in the predictive probabilistic notation of P90-P50-P10. While there may be some
root cause for similar ranged outcomes, positive Correlation should not be taken to imply
causation, but merely a similar outcome.
Correlation is usually found at a contributory level shared by correlated entities. For example,
two reservoirs on a common hydrocarbon migration path will have their ultimate resources
estimates linked to the amount and timing of the migration. The location across the range of
potential volumetric results related to trap fill extent will be linked, but not exact. Other factors
such as seal capacity, porosity, diagenetic enhancement, or destruction may be independent and
mitigate at least a portion of the correlation. As such, the typical correlation we see is not “full
correlation.” The correlated elements, be they wells in a reservoir, reservoirs in a field, or fields in
a portfolio, have a tendency but not a requirement to have a common degree of outcome.
8.2 Aggregating Over Reserves Levels (Wells, Reservoirs, Fields, Companies,

Countries)
8.2.1 Reservoir Performance. The most reliable estimate of the recoverable quantity, referred to
in the PRMS as the “Best Estimate” and usually equated to the P50 estimate when probabilistic
methods are used, is derived through extrapolation of well performance in mature fields [e.g., by
decline curve analysis (DCA)]. In applying DCA methods, industry practice is to work from the
lowest independent production level (e.g., wells or completions) upwards, comparing both
individual and reservoir- or field-level analysis. Unconventional reserves assessment is
particularly dependent upon well-level analysis because any unconventional field-level DCA
would be, at a minimum, misleading.
Performance extrapolation at the conventional reservoir level can lead to a higher Estimated
Ultimate Recovery than the sum of the extrapolated well decline curves for that reservoir for many
reasons, including catastrophic failures such as wellbore or completion damage. Also, the
comparison of individual-well DCAs to a field-level DCA will highlight small, systematic biases
that could otherwise be undetectable at a low level of analysis (where operational activities, such
as routine workovers or production downtime such as for maintenance, are typically handled at the
field level as opposed to well level for production decline analysis). Field-level economic
considerations that may shut in individual wells, thereby elevating the aggregate margin of the
remaining wells and extending the mean field producing life, result in a form of survivorship bias.
Another problem, which is specific to gas fields, is that the material balance p/z plot per well
often does not properly reflect the overall reservoir pressure decline. In these situations, it is
preferred to use an overall reservoir-level performance extrapolation.
8.2.2 Issues with Arithmetic Summation. The individual well aggregation vs. reservoir-level
aggregation difference, when numerous wells are present, is heightened if we use only 1P/1C
estimates for the well extrapolations (taking care that the 1P/1C forecast satisfies the reasonable
certainty requirements, as discussed in Chapter 4—Assessment of Petroleum Resources using
Deterministic Procedures herein). If we sum the individual well 1P/1C estimates to the reservoir
level, then we have assumed full correlation (i.e., that all wells will develop their low case
185
simultaneously). Arithmetically summing the 1P/1C results, while producing a valid 1P/1C total,
implies that all wells will be able to develop only their low (90% confidence) case resources. The
probability that all wells will only produce their 1P/1C estimate is extremely small and decreases
further as more wells are included. When we look at the arithmetic sum of 1P/1C values across a
larger area such as a pool or field, the likelihood that reserves will amount to at least a minimum
of the sum of those 1P/1C values is dramatically higher than 90%. As such, a simple arithmetic
sum of 1P/1C estimates across a reservoir will result in overly conservative predictive estimates at
the reservoir level, and most certainly across elements of a portfolio.
Understanding the disconnect between simple arithmetic sums and the aggregate probability
of a result is critical to the understanding of appropriate aggregation. Table 8.1 shows a simple
five-well example, where the individual reserves estimates (in Bscf here, with all Commerciality
criteria having been met) are identical, and the wells are uncorrelated (the result from one well
provides no information about the result from any other well).
Well 1P 2P 3P Cumulative 1P Sum
1 5 10 20 5
2 5 10 20 10
3 5 10 20 15
4 5 10 20 20
5 5 10 20 25
Table 8.1—Resource ranges and simple 1P sum for five wells in Bscf.
The confidence that each individual well would have at least 5 units of reserves is 90%. In the
reverse perspective, the probability of each individual well having its outcome ≤5 units is 10%. Is
it reasonable to state that the reserves across all five wells with 90% confidence is a total of only
25 units?
What is the probability that the aggregated result of the first two wells is 10 units or more? It
is far better than 90%. We have only a 1 in 10 chance that the result for Well 1 and Well 2
individually is ≤5. The probability that both wells will be at the small end of their range at the
same time is going to be even smaller. If we aggregate the distributions for the first two wells, we
see the probability of the combined result being less than 10 units is only 2.7%, or the confidence
we have that the aggregate result is greater than 10 units (the sum of the 1Ps) is 97.3%. The higher
is the number of elements being aggregated, the bigger is the deviation of the aggregate 90%
confidence from the simple cumulative sum.
We have discussed the underestimation error at the 1P end of the Proved Reserves uncertainty.
Similarly, an overestimation error occurs at the 3P end. Arithmetically summing the 3P results
simply provides the sum of the 3P numbers. As shown in Fig. 8.1, it does not produce a total that
in any way represents at least a 10% probability of occurring.
The results shown in Fig. 8.1 are from a stochastic aggregation of the five independent well
distributions shown in Table 8.1. The results were tracked over 10,000 realizations, and the
percentiles shown represent the likelihood of arriving at the arithmetic sum or less of the 3Ps
through the particular well aggregations (i.e., Well 3 results sum the results from Wells 1–3 and
then locate the probability of having ≤60 units of volume within the range of a probabilistic
aggregation).
186
Fig. 8.1—The percentile outcome of achieving at least the simple sum of the 3P Reserves across a portfolio of five
independent wells, given a distribution of 5-10-20 units for each individual independent well. The graph was created
through a stochastic simulation of the well results.
In our Table 8.1 and Fig. 8.1 example, the aggregated result of five wells that provides an
exceedance probability of 10% is 68 units, as opposed to the 100 units created by the simple 3P
arithmetic sum (5 × 20 units). In fact, the probability of achieving 100 units or better when we
aggregate the five wells is approximately 0.05% or 1 chance in 2,000. To rephrase, the outcome
that represents the 3P is 68 units, not the simple arithmetic sum of 100 units. Here, 100 units would
be exceedingly rare, and a statement that attributes the simple sum to be the 3P estimate would be
a gross overestimation of the true aggregated outcome. Again, while it may be permissible to add
the defined 1P, 2P, or 3P Reserves to indicate the specific category total, that sum is not a valid
representation of the reserves potential on an aggregated predictive basis.
Let’s take a look at this logic. The probability of one well’s outcome being in the top 10% of
its distribution is 10%. In fact, it should be obvious that the probability of any well being equal to
or greater than the 3P of the distribution is 10%. As we aggregate, we are adding together the
outcomes. Setting the 3P number equal to the arithmetic sum of the individual 3Ps means that all
the wells need to be at the higher end of their uncertainty range at the same time.
What is the probability of every independent well producing greater than or equal to its 3P
result? That is equal to the product of the individual probabilities, or 0.15, a chance of 1 in 100,000.
Clearly this is not 10%.
But wait, didn’t we just say (three paragraphs above) that the chance was about 1 in 2,000?
Why the 50× difference?
We said in our calculation that each independent well would be “at or greater than” its 3P (or
P10) result. It is important to remember as we aggregate distributions that all probabilities relate
to a threshold point and beyond. If we are just targeting the arithmetic sum of each well’s 3P, then
for every well that comes in over the 3P, it allows at least one of the other wells to come in under
the 3P threshold in order to have the aggregate sum be equal to the arithmetic sum of the 3Ps. It
should be obvious that we cannot aggregate independent well resource/reserve distributions by
187
arithmetically summing their 1Ps or 3Ps. Because of the skewness of the distributions, a similar,
though less consequential, error results from the addition of the 2Ps.
Table 8.2 shows the stochastic assessment results for the aggregated distributions
cumulatively as we add progressive wells. Note that as we add wells, the magnitude of the
arithmetic sum increases. Note also that as we add more population, the magnitude error increases.
When more elements are added together, as opposed to being properly aggregated, the calculation
will be the further off the valid answer.
Well Cumulative Cumulative Cumulative Cumulative Cumulative Cumulative

Arithmetic Sum Aggregate 1P Arithmetic Sum Aggregate 2P Arithmetic Sum Aggregate 3P
1P (P90) 2P (P50) 3P (P10)
1 5 5 10 10 20 20
2 10 13 20 21 40 35
3 15 22 30 33 60 50
4 20 31 40 44 80 64
5 25 40 50 56 100 78
Error Sum lower than Aggregate Sum closer but still lower than Sum higher than Aggregate
Aggregate
Table 8.2—Comparison of cumulative sums vs. cumulative aggregates.
Fig. 8.2 shows the percent error of our independent five-well example. The error is dependent
on the underlying uncertainty of the individual members of the collection. After only five wells,
the arithmetic sum of the 1P (P90) individual well outcomes is only 63% of the appropriate
aggregated number that provides at least a certainty of 90%. This error would continue to widen if
we were to increase the population with more independent members.
Fig. 8.2—Error in the arithmetic sum as the well population is increased.
Fig. 8.3 shows the 80% certainty range as we aggregate through the five independent wells.
It is important to note that as we add more wells, the percent difference decreases. Larger
populations of similar members have narrower result ranges. This narrowing can get to the point
188
of feeling too narrow, too confident in a particular group result. This is often because we typically
have relationships, links, or commonalities between the members of our aggregated entities.
Fig. 8.3—As population increases, the 80% certainty range decreases as a percentage of the Mean.
Well groupings, adjacent pools, or shared structural or depositional settings may provide
common causative factors that align or provide linked uncertainties in our reserves calculations.
These commonalities can cause a general agreement in uncertainty outcome. As such, the related
elements in the aggregation may show a degree of correlation.
As a guideline, anticipate the presence of some degree of correlation between wells in the
same reservoir because they have the same geological formation, drive mechanism, mode of
production, etc., but they will certainly not be fully correlated. There will be differences between
them resulting in a well distribution. Similarly, adjacent reservoirs will also have a degree of
correlation, which will provide a tendency to similar outcome. There will be differences well by
well and reservoir by reservoir.
1P Reserves are Proved and, by the PRMS definitions, have at least a 90% certainty (when
evaluated using probabilistic methods) that the quantities recovered will equal or exceed the
estimate. Nonetheless, the three designations 1P-2P-3P or P90-P50-P10 acknowledge uncertainty
in the final outcome. As a guideline reminder, we note the use of “at least” in the 1P definition; it
cannot be assumed automatically that 1P = P90, 2P = P50, and 3P = P10, especially for aggregation
purposes. These are the minimum acceptable confidence levels for these categories. Arithmetic
summing of 1P estimates is valid to provide a sum of 1P across a reservoir, across reservoirs, or
across a portfolio of projects. However, this simple sum should not be equated to a P90 value of
the aggregate (the amount of aggregated reserves having a 90% probability
of occurring).
Two approaches have been proposed to avoid the effect of arriving at too-low aggregates for
1P volumes when aggregating populations:
1. For reservoir estimates, apply decline analysis at the reservoir level.
2. Statistically aggregate Proved estimates from well level to reservoir level.
8.2.2.1 Method 1: Performance Extrapolation and DCA at the Reservoir Level. This
approach is an obvious and necessary supporting part of the performance analysis. In cases where
189
reliable production data at the well level are not available, DCA at a higher level where the
production is reliably measured (e.g., platform, plant, production station, or reservoir) may be the
only basis for the performance extrapolation. Another condition that calls for a higher-level DCA
is the occurrence of strong interference effects between neighboring wells.
Performance extrapolation at the reservoir level has a number of pitfalls:
• The performance will include the effects of ongoing drilling, development, and
maintenance activities. Time slicing the wells into these groups helps to differentiate
behavior related to activities.
• The aggregate may include wells at different stages of decline, with different gas/oil
ratios, etc. As such, be wary of survivorship bias.
• In multiwell aggregates, the decline may be dominated by the high-rate wells or
recent/well-specific activities, which may lead to incorrect estimation of the reserves.
• DCA early well behavior, pre-water breakthrough, enhanced recovery methods, or
market or infrastructure capped production may distort the reserves total.
• As parts of the reservoir are shut down due to negative local margin, the remaining wells
receive an economic benefit that may extend their anticipated economic life and reserves.
For example, offending high water-cut wells may be shut in, thereby reducing water
handling and disposal costs. This is a form of survivorship bias and should be taken into
account in the initial aggregation.
8.2.2.2 Method 2: Statistical Aggregation of Well-Level Proved Estimates. A different
approach to compensate for arithmetic addition of high-confidence estimates may be to apply a
form of statistical addition. This method also has elements that will require attention:
• Well-level Proved estimates are often correlated due to common aquifers, formation
heterogeneity, facilities, operational constraints, etc. If independence is assumed, it is up
to the reserves evaluator to justify this assumption. If full correlation is assumed, a high
level of justification is required.
• Probabilistic or statistical methods often rely on statistical simplifications, e.g., the
assumption of particular distribution shapes or probability density functions for the
reserves estimates. The upper end of such distributions must be appropriately managed
with limits (capping the upper ends at a set feasible magnitude as opposed to truncating
the upper ends).
It should be noted that many of the above problems may be avoided through the use of
stochastic simulation techniques. At the reservoir aggregation level, DCA is the method of choice
for 1P/1C and sometimes in cases of low uncertainty 2P/2C ranges (when 2P/2C is close to 1P/1C)
when correlation is minimal. Justification from the qualified reserves evaluator should be able to
support the aggregation level chosen.
8.2.3 Correlations Between Estimates. A brief introduction to Correlation vs. Dependency has
been covered in Section 8.1. We will now delve further into this critical part of an assessment.
Aggregation of resources must consider correlation between the items being aggregated.
Correlation is present whenever the volume outcome (known or presumptive) from one entity
indicates a likely volume tendency of another. It is described in a way that quantifies the degree of
similar outcome. Numerically, the degree of similarity of outcome is expressed in terms of a
correlation coefficient, which ranges from −1 to +1. A correlation coefficient of zero (0) indicates
no correlation; i.e., the knowledge of one outcome has no impact on expectations from another. A
correlation coefficient of +1 indicates an exact match (full positive correlation) of probabilistic
190
outcome between entities being aggregated; i.e., the exact same probability range result will be
found between wells (Well 1 meeting its P15 result means that Well 2 will also produce its P15
result). A correlation coefficient of −1 indicates a perfect, symmetrical, and certain counter-result
situation or full negative correlation (Well 1 meeting a P15 result would mean that Well 2 would
produce a P85 result). The end points of −1 and +1 would be extremely unlikely to ever occur in
nature. We must remind ourselves that the simple addition of 1P Reserves necessarily defaults to
a correlation coefficient of +1. Correlation exists if the outcome from one or more wells in a
collection affects the anticipated result from other wells. The degree to which it affects the
anticipated result is the correlation coefficient.
Note, in the previous paragraph, we did not mention the magnitude of any particular outcome.
We consistently described outcomes by their probability location. It is important to state that when
we discuss correlation (the similarity between two outcomes), the correlation to which we refer is
a Rank correlation. Rank is the probability position of an outcome across the full uncertainty range.
Rank correlation, therefore, is the similarity between outcomes on the probability range for the
uncertainties, not the actual magnitudes. Items being aggregated will have their own distribution
of potential results, but the tendency to have a particular outcome, be it low or high within the
respective uncertainty ranges, indicates correlation is present.
Correlation rests at the reservoir component level, and it is possible to have several different
levels of correlation within the many parameters in the resource calculation. Correlation originates
from common influencing factors such as depositional environment, diagenetic history, shared
spill point, and, on the human interaction side, similar evaluation bias and common completion
mechanisms. Correlation originating from similar evaluation bias, or modeling bias, is often
underestimated in the assessment of overall correlation. Examples of modeling bias include
velocity (time-depth) uncertainty or assumptions about decline curve parameters. Recovery factors
based on potentially insufficient or imperfect sampling also may not be reliable for estimating
future development outcomes.
As the aggregation level increases, correlation coefficients move closer to zero but remain
influential. Table 8.3 shows degrees of correlation for a selection of feasible causative situations.
A positive correlation between two estimates can be illustrated with the depth-area plot of a
field shown in Fig. 8.4, which consists of two reservoir sands divided by a shale layer within the
same enclosing structure. The sands have a common oil/water contact. Obviously, in this case, the
reserves for both sands will change in the same direction if an exploration well finds the oil/water
contact to be somewhat shallower, if a new seismic interpretation lifts the flank of the structure, or
if the time-depth relationship is found to be different. Summation of the low estimate values for
the two sands may be used to arrive at a low estimate for the field, but only if all other parameters
are also fully correlated, or their variance is very small.
This situation results in a strong correlation coefficient between the individual producing
horizons, often in the 0.7–0.8 range. The temptation is to assume a simple arithmetic sum of the
high (P10) and low (P90) respective assessments, but this ignores other reservoir and completion
uncertainties. A correlation coefficient of +1, which would allow simple arithmetic summing
across the predicted range of results, is extremely rare. It dictates that all relevant reservoir
characteristics found as reservoir 1 is drilled/completed would enable precise predictive capability
of characteristics to be found in reservoir 2. This is not to say that the uncertainties would resolve
to have equal magnitudes, but only that the position within the range of possible outcomes would
be known (i.e., the rank).
191
Degree of Correlation Examples
None – Correlation coefficient = 0
Local or independent pressure systems, structurally independent, significant vertical
No common factors in deposition,
separation, different depositional and diagenetic systems, and different completion
reservoir, method of completion,
paths.
or assessment.
Weak – Correlation coefficient

< 0.3
A shared uncertainty is not
Common clastic or carbonate depositional system classification.
considered to be an important
Common source of recovery factor estimates, tools (e.g., reservoir simulator), and ranges
indicator of results between the
Elements having a common saturation-calculation method (e.g., Archie, dual-water, etc.)
elements. Correlation does not
or saturation-height function (particularly those using a common reference data set).
materially affect the aggregated
results.
Medium – Correlation coefficient

0.3 to 0.7 Common depositional and migratory path context. The success of a low-pressure
The shared uncertainties and compression project in one field is a prerequisite of success in another, and hence the
evaluation bias could be real and recovery factor uncertainties are likely linked. However, the major components of the
significant. uncertainties in reserves of the two fields (structure, etc.) remain independent.
Strong – Correlation coefficient

The aquifer and pressure systems between two proximal fields are likely to be common,
d> 0.7 and < 1
and actions in one field will affect recovery in the other. Well behavior within an
The shared uncertainties are
unconventional part-play. Reservoirs with a common structurally defined spill point, leak.
known to be real and significant.
Aggregation of unconventional wells across a part-play.
Total – Correlation coefficient = 1 Two adjacent accumulations have commonality assumed in all uncertainties (e.g.,
The shared risks are absolute. reservoir unit, velocity model, aquifer drive); thus, their reserves estimates can be added
arithmetically.
An oil field is developed in a core area only. Additional upside in stock-tank oil initially in
Negative Ranges – Correlation
place (STOIIP) in flank areas will result in a reduction in the average recovery factor.
coefficient between 0 and −1
Uncertainty in fault location works in the opposite direction for gross rock volume (GRV)
The shared uncertainties are
in two adjacent blocks. This is an internal negative correlation within a single reservoir.
countervailing or reciprocal.
Correlations may be negative (opposing outcomes) but need not be absolute.
Table 8.3—Ranges of correlation between reserves estimates of fields, reservoirs, or wells

(adapted from Carter and Morales 1998).
Fig. 8.4—West Star field area vs. depth.
Negative correlation coefficients are possible. They represent a reciprocal behavior between
reservoirs (or wells). As an example, Fig. 8.5 shows two noncommunicating blocks having a
192
common structure separated by a fault of uncertain location. If the fault is found to occur closer to
the crest of Block A, then the volume associated with Block A decreases, and the volume for Block
B increases. The reverse would occur if the fault was further into the mapped Block B area. At the
gross rock volume component level of the reserve calculation, there would be a (−)1 correlation
coefficient. At the resource volume level, there would be a negative correlation apparent, but due
to all the other uncertainties in the assessment, it would be highly unlikely to be as strong as (−)1.
Fig. 8.5—Block A and Block B separated by a fault of uncertain location. As the fault location changes, the volume
associated with one side of the structure transfers to the other, thereby creating a negative correlation.
If we arithmetically sum the low estimate values in each of the two blocks, we arrive at the
low estimate for the structure. However, given that the individual low estimate assessments for
each side must necessarily reflect the amount each block is at least 90% certain to contain, this
would require holding the fault location at or near the reasonable minimum proximal location for
each individual block. This is not reasonable. The fault cannot be in two different locations at the
same time. The arithmetic sum of individual block low estimates will underestimate the aggregate
low estimate for the structure. This structure is best and most fairly evaluated at the aggregate level
by taking into account the negative correlation.
Another commonly encountered negative correlation is the situation in an oil reservoir with a
gas cap, where solution gas below the gas/oil contact is estimated separately. If there is an
uncertainty in the gas/oil contact depth, then there is a negative correlation between the gas
reserves that are carried above and the oil reserves below the gas/oil contact.
The depth-area plot method for volume assessment is particularly useful in this situation. Fig.
8.4 is an example of such a plot. Depth-area plots are created by taking the area of the reservoir as
it appears at a particular depth. Calculation of the resource volume in place becomes a simple
integration of the volume below the plotted line. When the reservoir is limited in height, the depth-
area plot of the base of the reservoir is needed in addition to the reservoir top. Fig. 8.4 shows two
reservoirs. Each reservoir has a top and a base plotted. This method is particularly useful in
nonuniformly shaped reservoirs.
Unless information is available, such as detailed fluid properties or verified/conformable
seismic amplitude anomalies to guide the placement of the gas/oil contact, it is usually appropriate
to assume that the volume above the highest-known oil is occupied by the lower-value product
(usually gas) as the less-impactful case.
193
Adding up P50 (or Best Estimate) values will show a more valid result; however, the only
truly valid arithmetic summation occurs (outside of full correlation across all uncertainties) when
the Means are summed. Obviously, if the low estimates are added, then the low case for free gas
will correspond with a high case for solution gas and vice versa. To handle this, a stochastic
procedure referencing a depth-area plot construct (using an Excel™ spreadsheet with stochastic
modelling techniques or with a commercial add-in) can be used to arrive at the resultant
distributions for predictive gas initially in place (GIIP) and potential reserves at the field level.
Finally, in rare situations, it is possible to have correlations that are more complex than these
examples. Resource volume correlations may not be monotonic in nature and can take on a “U” or
upside-down “U” shape. For example, there may be situations arising from varied uncertainty
reservoir connectivity models, where both a small-volume outcome or a large-volume outcome for
one reservoir could be associated with higher volume estimates for a second reservoir. In this
instance, outcomes for the first reservoir occurring closer to the distribution center would decrease
the reservoir volume of the second. Relationships between reservoirs may be simple or complex,
but they should always be taken into account for reserves or resource aggregation.
8.2.4 Dependency. When we examine Dependency, we require at least one member of the
“dependent pair” to still have probability of success (or Pg) remaining unresolved. As such, it
would be a rare circumstance indeed to see Dependency within a 1P-2P-3P context. However, we
would likely encounter it within or across the 1C-2C-3C categories. “Contingent” relates to a
condition being established or a threshold being passed. In either situation, the sought-after result
is binary: It is either there, or it is not; the threshold is either exceeded, or it is not.
Dependency is assessed between two potentially productive entities. One of the entities may
be thought of as independent and bearing its own Pg. The second entity’s Pg will be determined
dependent on the success or failure of the independent entity. Given resources are identified for
the independent entity, the shared chance elements with the dependent entity are moved to 100%.
The entities together may be referred to as a dependent pair.
The Pg is a product of its chance element probabilities. These chance elements differ company
to company as mentioned previously in Section 8.1.1, but they generally pertain to aspects of
container and contents: Source, horizontal seal, vertical seal, reservoir, and timely migration are
common chance elements. A Dependency exists when the probability of success for the dependent
element is altered by the knowledge of success or failure of the independent element. Dependency
must be considered in resource assessment when there is chance learning between elements being
assessed, in other words, if knowledge of one binary outcome alters the probability of a second
binary outcome.
Valid aggregations of Contingent Resources will require Pg to be kept separate from
uncertainty. It is recommended to aggregate the contingent component by sampling from the P90-
P50-P10 range Pg % of the time; otherwise, appropriate aggregation that produces a range of
confidence will not be possible. For example, as depicted in Fig. 8.6, our discovery well had a
40% chance of success. Upon discovery, our assessment of chance factors changes. We now know
that the Pg for Reservoir A is 100%, and Reservoir A has a P90-P50-P10 range of 4, 6, and
9 million barrels of oil, respectively. After Reservoir A success, Reservoir B has a 30% Pg, and
given success, it is estimated to have a P90-P50-P10 of 2, 4, and 8 million barrels of
oil, respectively.
Let’s examine how to aggregate these reservoirs when they are independent, and when they
share risk, before and after the initial discovery.
194
Fig. 8.6—Venn diagram of the prediscovery chance of success state (not to scale).
8.2.4.1 Independent Reservoirs. When reservoirs are independent, the success of one has
no effect on the assessment of chance of success for the other. In our independent situation,
PA = 40%, PB = 30%, and therefore
PAB = PA × PB = 12%,
and
PFailure = PA Fails × PB Fails = (100% – 40%)(100% – 30%) = 60% × 70% = 42%.
The aggregation needs to take each of the probabilities of success into account. Again, the
easiest way to aggregate the two reservoirs is to set up a quick simulation. Use two random number
generators, one for each reservoir. If the realization returns a random number less than the
probability of success for the reservoir, then sample from the distribution; otherwise, the result for
that reservoir in that particular realization (trial) is 0. Sum and track the result for each trial. This
will provide a probability of at least one of the reservoirs being successful and the distribution
given at least one successful reservoir.
It is incorrect to factor a reservoir’s distribution by the chance of success and assume it
has meaning.
Let’s take a small step to look at the postdiscovery assessment.
195
Given our discussion above, it should be obvious that all we have to do in the aggregation
model is to assign a chance of success to Reservoir A of 100%. In a Venn diagram representation,
it looks like Fig. 8.7.
Fig. 8.7—Postdiscovery Venn diagram (not to scale).
Reservoir B’s success looks smaller, but it isn’t. As they are independent, Reservoir B’s
success takes up 30% of whatever our “world view” is. Prediscovery Reservoir B took up 30% of
the world that included dry holes. Our world has changed to only show the cases in which A is
successful. PB is still 30%.
PAB = PA × PB = 100% × 30% = 30%.
If we aggregate the two reservoirs in this context, we sample every trial from the Reservoir A
distribution and only 30% of the data from Reservoir B’s distribution. As Reservoir A is
successful, we don’t have a probability of at least one success. We have Reservoir A’s distribution
that is bumped by the occasional addition of Reservoir B. This will have a greater effect on the
upper portion of the aggregated distribution. In our example, the aggregated distribution is
P90 = 2.2, P50 = 5.3, and P10 = 11.8 million barrels of oil.
Other methods, such as prorating the Reservoir B at 30% and allocating it across the
Reservoir A distribution, or sampling Reservoir B only for the P10 case of Reservoir A, do not
provide valid aggregation.
8.2.4.2 Dependent Reservoirs. What do we do if the two reservoirs share chance, such as a
common seal or migration path? This shared chance means that if we are successful at Reservoir
A, then the probability of success for Reservoir B has increased (some of its chance threat has been
196
solved). Given that Reservoir A is successful, we see the probability of success for Reservoir B
increasing, perhaps to 50%. The formulaic abbreviation for the probability of B given A success
is P(B|A). Since Reservoir A’s success doesn’t solve all of Reservoir B’s chance elements,
Reservoir B still can fail.
Where do these dependent numbers come from? If the reservoirs share seal and migration
risk, and Reservoir A is shown to be successful, then seal and timely migration must also be present
for Reservoir B. However, seal is only one of the risks inherent to Reservoir B. A proven seal and
timely migration alone do not guarantee success of Reservoir B, but the estimate of the probability
of Reservoir B being successful increases when the results of Reservoir A are known.
The rule for the maximum dependency of two dependent elements is: The maximum
probability of the dependent element given success of the independent element (the one you find
out about first) is equal to the ratio of the dependent element’s Pg to the independent
element’s Pg.
In our example, the highest possible probability of success (prior to drilling) for Reservoir B
is 30/40 or 75%. We must always preserve the initial independent probability of success because
dependency changes only when something occurs, not when its size changes, and we can never
create probability.
When we have a dependent set of reservoirs to aggregate, we have to modify the dependent
reservoir’s chance of success given the independent reservoir’s success or failure outcome.
In the prediscovery aggregation model, we have to account for the percentage of the time A
and B will occur together, the percentage of the time they will occur separately, and the percentage
of time that neither would occur. These percentages are a function of their individual independent
probabilities and their shared chance. Fig. 8.8 shows the configuration with a degree of partial
shared chance (“partial dependency”).
Fig. 8.8—The configuration of probabilities of success in a dependent situation where P(B|A) = 50% (not to scale).
197
When working with dependencies, it is important to remember that the individual probabilities
do not change (predrill) given any amount of positive or negative dependency. Dependency only
controls when success happens, not how often it happens. Before we drill Reservoir A, we have
assessed the common risk elements between the two reservoirs and have found that if Reservoir A
is determined to be productive, then the improvement in probability of success due to the shared risk
elements between the reservoirs changes the probability of Reservoir B given A success [P(B|A)] to
50%. Pg of Reservoir A is 40%, and so 50% of that is 20%; therefore, the probability of both
reservoirs being productive is 20%, an increase from 12% in the independent case.
The opposite action will also be present. If Reservoir B is more likely to occur when Reservoir
A is successful, then when Reservoir A is a failure, Reservoir B has a lower chance of success. In
our example, the probability of Reservoir B success given Reservoir A failure reduces from the
independent probability of 30% to only 10%. This is due to the risk shared between the reservoirs.
Note that the independent probabilities of 40% and 30% for A and B, respectively, are
preserved. The only element that has changed is the overlap of the probability circles. The B circle
has shifted into the A circle to the extent that it now covers 50% of the A area. In the shift, B
converts some of the time it is occurring without A into the time it occurs with A (delta caused by
the shared chance). In that process, the probability of both reservoirs failing increases by the same
amount as A and B occurring together increases. Dependency in the form of shared chance
decreases the probability of having at least one success but increases the distribution of resource
when successful. This balance means that risked resources do not change.
When we look at the postdiscovery situation, the dependent question regarding the likelihood
Reservoir B will be successful has been solved. Fig. 8.9 shows the postdiscovery case.
At this point, remember that reserves do not include “risked reserves.” Risked aggregation
normally occurs at the resource level, but the concepts can also be applied to the lateral extent of
a known continuous reservoir, especially when planning infrastructure, appraisal, or development
well locations.
Fig. 8.9—Venn diagram of the dependent (shared chance) situation (not to scale).
198
8.2.5 Higher-Level Aggregation. As discussed previously, aggregation of resource estimates in a
statistical way will usually result in different volumes than simple arithmetic summation.
Probabilistic aggregation is valid at the company portfolio or regional assessment level. Industry
in general is comfortable with the idea of aggregating probabilistically up to the field level for
development and infrastructure decision support, provided correlations and dependencies are
handled properly.
The PRMS (§ 4.2.5.4) recommends that reserves figures should not incorporate statistical
aggregation beyond the field, property, or project level to allow comparability and consistency.
This aggregation method has been followed by many if not most companies in industry for internal
use for plays, portfolio management, and broad regional assessment.
It is important to adhere to the specific definition for reserves qualified as 1P-2P-3P or resources
qualified as 1C-2C-3C. For example, maturation of reserves on single producing entities may be
legitimately tracked as they progress through the categories. However, as we aggregate multiple
producing and potentially productive assets, we recognize that their inherent uncertainties contribute
to a progressive misrepresentation of the minimum 90% certainty objective. How we phrase the
results must take certainty into account in order to provide a fair and reliable assessment of the
quantity that is expected to be produced. The rough affiliation between 1P-2P-3P and P90-P50-P10
no longer applies as we aggregate the uncertainty across several assets. Companies, government
agencies, and other evaluation entities have recognized for some time that the P90 of the aggregate
will be higher than the arithmetic sum of 1P Reserves, and the P10 aggregate will be lower than the
arithmetic sum of the 3P Reserves. 1P-2P-3P categories maintain the reserves definition for the
quantities that exist by asset. P90-P50-P10 maintains the realistic expectation or reserves certainty
for production, transport, and facility decisions.
A field containing different reservoir blocks (layers, pools, accumulations) may be fiscally
ring-fenced and developed as one unit. This is the “level of economic decision.” Alternatively, the
level of economic decision may be based on an area of infrastructure development within a larger
fiscal ring-fence. Regardless, stochastic aggregation of reserves to the economic decision level is
recommended prior to development/infrastructure decisions.
Resource aggregations used in policy and decision making should ideally be probabilistic in
context, reflecting the full spectrum of possibilities. Fiscal unit-of-production depreciation of the
assets may be defined at this level. Use of higher-level aggregation for lower levels of economic
decisions may lead to poor outcomes. It is best to match the level of aggregation to the level of
economic decisions being made.
At the lowest level, deterministic Low/Best/High Estimates, defined consistently using the
certainty-based definitions discussed in Chapter 2—Petroleum Resources Definitions,
Classification, and Categorization Guidelines, may be the starting point for a range-based
aggregation. The use of scenarios as a starting point may be appropriate given that the scenarios
are mutually exclusive and fully comprehensive, and there is a valid/reliable assessment of
individual scenario probability.
It should be noted that if only deterministic estimates are available, then the only option is to
use arithmetic summation. The discussion of statistical aggregation only applies for probabilistic
analysis (or to convert scenarios to quantitative probabilities).
199
8.3 Adding Proved Reserves
8.3.1 Pitfalls of Using Arithmetic Addition of Proved Reserves. When investors or companies
add 1P Reserves of several reservoirs arithmetically, they typically underestimate the aggregated
value of their assets. This is because the upside on most reserves estimates will more than
compensate for the downside on the underperforming assets in the portfolio. Portfolio
diversification provides downside protection. This will certainly happen if the estimates of the
volumes are independent of each other. For this reason, most companies will rely on the 2P
numbers rather than on the high-confidence 1P estimates for business planning purposes. The 2P
estimates are not additive to the valid aggregate P50, but their use will deliver an answer closer to
the aggregate P50 than the summed P90s or P10s will to their respective aggregate values (refer
to Table 8.2). Due to the skewed nature of the potential outcomes, the only potentially valid
additive point in a range is the Mean, but the elements must be independent and noncorrelated.
Addition of individual 2P assessments, even if they are taken to be representative of the P50, will
continue to underestimate the actual P50 of the aggregate reserves in right-skewed distributions.
If the resource estimates are independent, then the upsides in one field may offset a
disappointing outcome in others. In other words, the statistically aggregated P90 of the group is often
significantly higher than the (arithmetic) sum of the P90 volumes of the individual fields (see
Schuyler 1998). For the same reason, arithmetic addition of the 3P values of individual reservoirs
will likely overestimate the real upside of the combined asset. As reservoir linkage is increased due
to common uncertainties (increasing correlation), the total range possibility increases.
If we stick to arithmetic summation of Proved Reserves, we run the risk of systematically
underestimating the value of our combined assets. Technically, this can be avoided because tools
are readily available to account for diversification, i.e., the favorable condition of having a mix of
assets. Such tools enable the investing community (and some government agencies) to value a
combination of assets higher than the sum of the Proved volumes of the individual parts. Error at
the upper end due to simple summing of 3P estimates can cause commensurate issues, particularly
with large area portfolio policy decisions.
Organizations and government agencies often have a responsibility to provide a scope of
Proved, Probable, and Possible Reserves across a business unit, basin, region, or total portfolio.
Projecting future production, delineating potential regional infrastructure systems, or supporting a
nation’s strategic energy decisions all rely on the appropriate aggregation of reserves and resources.
Governments of some countries around the North Sea, including Norway and The Netherlands,
were early adopters of probabilistic or stochastic aggregation methods. The Dutch Ministry of
Economic Affairs has applied the method of probabilistic summation for Proved Reserves since the
mid-1980s. The confidence-based approach of P90-P50-P10 has been shown to be a valid and
reliable method with which to assess and communicate reserve and resource totals.
8.3.2 Arithmetic or Correlation Inclusive Summation. Arithmetic summation is the usual

straightforward way of adding volumes and thus of aggregating reserves. Let us look at two gas-
bearing reservoir blocks, A and B, with the parameters in Table 8.4.
With the range and probability density function of these parameters, we can construct a
probability distribution for each individual block as shown in Fig. 8.10, with the cumulative
probability of exceeding a given volume on the vertical axis.
200
Parameter Units
Block A Block B Total
Total gross rock volume (GRV) 109 m3 1.74 1.16 2.90
Porosity Fraction 0.22 0.22 0.22

Net-to-gross Fraction 0.85 0.85 0.85
Saturation Fraction 0.80 0.80 0.80
Gas expansion m3/ m3 205 205 205
Expectation of GIIP 109 m3 53.4 35.6 89.0

Proved GIIP 109 m3 43.3 28.5 71.8
Table 8.4—Example case: Gas reservoirs A and B
Fig. 8.10—Probability distribution for reservoir blocks A and B.
Note that for the sum of the Proved GIIP figures in Table 8.4, we have taken the arithmetic
sum of two Proved numbers, both of which have a 90% probability of being met or exceeded. In
fact, by adding the points along each distribution shown in Fig 8.10, we have erroneously assumed
complete correlation between the two blocks; i.e., we assume that if the low side of one case
materializes, then the same exact outcome will happen with the other block. Summing in this way,
we arrive at a pessimistic number for the Proved GIIP, representing the situation that both blocks
turn out to be relatively disappointing. While we may think this is possible if both blocks have a
common gas-water contact, or if their volumes are determined by the same seismic phenomena,
every single uncertainty would have to be fully correlated between the blocks. This is a virtual
impossibility. Even the bias introduced by the same subsurface team, applying the same methods,
working on two reservoir blocks may introduce a positive correlation.
Mean values are additive given no correlation between the independent entities. A common
saying in probabilistic circles is: “The sum of the means equals the mean of the sums.” Correlation
201
can alter this. * It may have an effect on the Mean depending on the skewness, number, and relative
magnitude of the aggregated elements. Means may be calculated through statistical methods or
stochastic sampling of a range, or by using Swanson’s mean approximation method, i.e., [0.4×P50
+ 0.3(P10 + P90)].
8.3.3 Probabilistic Aggregation and the Portfolio Effect. Each element in an aggregation has a
range of potential outcomes. If we aggregate the ranges instead of individual potential outcomes,
we are carrying out a probabilistic aggregation. When we derive the aggregated 1P/1C value from
the distribution of the sum (the magnitude at which we are ≥90% confident), we may have
situations (e.g., in a probabilistic simulation) where a low outcome of Block A may at times be
combined with a high outcome of Block B, or the other way around, and the individual outcomes
across the entities compensate for the end-member outcomes in the other cases. The upside and
downside aggregated outcomes become muted. It is still possible for them to exist, but they are
much rarer. This results in a cumulative distribution curve for the combined GIIP that is steeper
(i.e., has a smaller spread) than the curve for the arithmetically added volumes, as shown in
Fig. 8.11.
0.8
Cumulative Probability
0.6
0.4
Arithmetic Proved
72 x109 m3
0.2
Probabilistic
Proved
77 x109 m3
0
50 60 70 80 90 100 110 120 130 140
GIIP (10^9 m3)
Arithmetic addition Probabilistic addition
Fig. 8.11—Arithmetic and probabilistic addition, A and B.
This tendency of the uncertainty range to narrow is a statistical truth that will always be
observed if we aggregate quantities that have independent statistical distributions. As we increase
the number of items being aggregated, the ratios of P50:P90 and P10:P50 decrease, though the
magnitude difference between P90 and P10 increases. This “portfolio effect” increases with the
*
The best way to prove this point is to look at a negative correlation coefficient. Think of the situation shown in
Fig. 8.5, where a fault separates two reservoirs. Is it possible to achieve the mean outcome in each reservoir at the
same time? No. As one reservoir increases to the mean, the other has a compensating decrease, since the fault location
prevents simultaneous mean outcomes. Adding the means together would overestimate the aggregated value. Provided
the population is not correlated, the means are additive.
202
number of elements being aggregated. One may think of this in a slightly different way: When
higher numbers of entities are aggregated, the P10 and P90 (in percentage magnitude) will be
closer to the mean, and the addition of another similar entity will have less influence on
the distribution.
In our example, applying this approach and assuming complete noncorrelation, we can state
with 90% certainty that there is at least 77 × 109 m3 of gas in both reservoir blocks, as opposed to
72 × 109 m3 of gas using arithmetic summation. In situations where gas contracts are based on
Proved Reserves, this may have considerable business implications. This difference is significant,
and we are only aggregating two items. The effect increases with increased uncertainty in the
individual portfolio elements.
Methods to aggregate volumes independently (assuming no correlation between aggregated
items) are:
• Stochastic methods, using a spreadsheet or commercial tool sampling techniques.**
• Scenario trees, representing the possible outcomes as branches of a tree and calculating
the overall outcome. This method is treated in Section 8.5 (and Chapter 7—Probabilistic
Resources Estimation, Section 4).
• Compiling distributions statistically using Mean, Standard Deviation, and Variance.
8.3.4 Simple Probabilistic, Stochastic Simulation, and Correlation Matrices. Initial corporate
ventures into simulation as an aggregation method typically combine elements based on
submission of P90-P50-P10 values. Simple probabilistic sampling randomly chooses one of the
P90-P50-P10 values for each of the elements, and, within an element, the choice among P90-P50-
P10 defaults to the Swanson’s mean approximation method. This sampling method holds true to
a valid center-weighted approach, which, ultimately, over 1,000 or so realizations, provides a
valid sampled mean for each component. If moderate to high correlations exist between any of
the elements, the simple probabilistic method is not appropriate for aggregation. If the population
of elements being aggregated is greater than 20, and correlation is not present, it is acceptable to
use a simple probabilistic method. Corporate portfolios are often aggregated based on P90-P50-
P10 realizations of reserves and capital requirements.
Stochastic simulation is the most flexible and comprehensive method used to aggregate. It is
possible to construct a probability density function for each element based on resource assessment
for Low/Best/High Estimate definitions. The ranges are then sampled over a large number of
iterations. The number of iterations (or realizations) desired will be a function of the skewness,
population, and consistency of the elements being aggregated.
While most companies and government agencies use one of the several off-the-shelf
commercial products (typically add-ins to Excel), it is possible to design a custom simulation
spreadsheet completely within Excel. There are non-Excel based custom software products on the
market as well. Each has its advantages and disadvantages, but it is important to cross-check
dependency and correlation methods to ensure correct calculation.
Most of the time, the distribution shape for an element will be skewed, and a log-normal
probability density function will be appropriate. However, the upper end of the distribution must
be managed in the assessment, especially distributions having an infinite upper endpoint, such as
log-normal distributions. Consequently, a maximum low estimate is established for the
**
Stochastic simulation is often referred to, improperly, as “Monte Carlo” simulation, but Monte Carlo is simply a
sampling technique, and not the best one. Latin hypercube, where the distribution is binned, and all bins are sampled
before any repeat, gives a far more robust/stable result with fewer iterations.
203
“acceptable” sampling. The raw untruncated distribution is sampled, and when the sample
exceeds the limit, it is reset to the limit. This method preserves the number of samples created
above the P10 to 10% of the realizations. It also conforms to the Swanson’s mean approximation
and natural data set deviation from skewed probability distribution functions. This effect can be
seen in Fig. 8.12, which shows a log-probability plot of Estimated Ultimate Recovery (EUR) for
an unconventional play in Texas.
Fig. 8.12—Log-probability plot of well Estimated Ultimate Recovery for a liquid-rich Texas unconventional play (in
millions of barrels of oil equivalent).
Failure to account for the upper deviation from log-normal will corrupt the resource/reserve
aggregation, leading to an overestimation of potential.
Correlation may be handled within a commercial product using correlation matrices or by
simple pair-wise assignment of correlation coefficients. It is also possible to design a window-
sampling technique in Excel that equates to correlation coefficient–based sampling (the higher
the correlation coefficient, the narrower the allowed sample space).
Most practical situations will be in between the three simply handled correlation coefficient
endpoints of 1, 0, and −1. Some parameters of our estimates will be correlated, while others will
be completely unassociated. This results in different degrees of correlation.
Ignoring correlation produces invalid results. The rigorous solution in this situation is to
calculate probability distributions, specify the correlation between them, and generate the
resulting probability distribution for the aggregate. The aggregation may be combined using
advanced statistical methods, though those methods assume consistency in distribution shape and
work best when skewness is minimized. Stochastic simulation is the obvious and most expedient
method to achieve this.
Carter and Morales (1998) provided a real-life example of correlation in a portfolio of 25
fields sharing common production facilities. Each field had a range of gas reserves, expressed at
the P90 (Proved), P50, P10, and expectation (mean) levels. The Proved Reserves per field were
defined as the volume that had a 90% chance of being met or exceeded. Adding the P90 volumes
204
arithmetically results in a volume of Proved Reserves across the project that is 15% lower than
the stochastically combined P90. Because neither full correlation nor full noncorrelation can be
assumed, the authors then proceeded to analyze the areas of potential correlation between the
individual estimates by applying the following procedure:
1. The areas of correlation were tabulated for individual fields to identify common factors
between fields. These areas included technical, methodological, and natural subsurface
commonalities between the GIIP estimates of the fields. Commonality was classified as
weak, medium, or strong (examples of which have been previously shown in Table 8.4).
2. An estimate of correlation coefficients was made by assigning values for a weak,
medium, or strong correlation and combining them into an array suitable for use in a
Monte Carlo presentation.
3. The reserves distribution (for each field) as defined by the P90, P50, and P10 confidence
levels was expressed as a double-triangular probability distribution function.
4. A matrix of correlation coefficients was used to describe the shared risks between fields,
with a coefficient for each pair of fields varying from 0 to ±1 correlation coefficient.
5. The reserves distributions for each field were then stochastically sampled and
aggregated using the previously defined correlation matrices employing stochastic
simulation software stand-alone tools or spreadsheet add-ins.
The result of applying this method for the case described was that the gas reserves at the 90%
confidence level were some 9% greater than those resulting from arithmetic summation. Not
taking the dependencies between the fields into account, the increase would have been 15% over
the straightforward arithmetic summation. Correlation stretched the reserves range.
The linked risks resulting from shared surface facilities and constraints were also excluded
from the analysis. They are considered to be common (project) risks, and problems with facilities
are considered surmountable if they materialize. This type of shared risk can be included in the
analysis, if required.
Use of correlation matrices as described above is similar to other reserve estimation methods
in two important aspects:
• The figures used are subjective and change when new insights are gained. However, in
view of the large number of correlated reservoir elements resulting in varying degrees
of field-level correlation, the aggregated result is relatively insensitive to individual
error. Major reversals of opinion must occur to change the overall result by a significant
amount.
• As the established risks are addressed in more detail, specific correlation coefficients
will be updated with the proper audit trail. For example, a new seismic interpretation by
a new team may result in the shared uncertainties in seismic interpretation being
removed after the new interpretation has been accepted.
Correlation matrices are a good way to lay out correlations between individual members of
an aggregation. However, there is a significant danger to create impossible correlations, or
impossible circular calculations, and thereby destroy the validity of the aggregation unless the
aggregator is cautious and attentive. For example, let’s assume three partially correlated
Reservoirs A, B, and C. If we correlate B and C to A at a rank correlation coefficient of 0.7, that
precisely defines a minimum degree of correlation between B and C.
205
This example is illustrated in Fig. 8.13. We see both B and
C have a relationship with A. The coefficient of 0.7 would be a
moderately strong correlation, so as A’s outcome is high, B’s
outcome would track similarly higher. C’s outcome would also
track generally higher. Therefore, there is also a positive
correlation between B and C, but it will not be defined by our
input; it is mathematically determined. An imprecise correlation
between B and C is roughly equal to the correlation coefficient
of A-B multiplied by the correlation coefficient of A-C. As A-
B is 0.7 and A-C is 0.7, a rank correlation coefficient is
automatically formed for B-C of 0.7 × 0.7 = 0.49.
Fig. 8.13—Three correlated elements. As we apply rank correlation between elements, there is an
important factor to think about: How best do we apply a rank
correlation coefficient across a large number of wells? Our well collections may be generally
good, middling, or generally poor. We have significant variability in the quality of our well
families, but along trend or within a sub-play, wells may be strongly correlated. Correlation
typically originates from a common natural (and unaltered) rock, containment, or diagenetic
characteristics. Drilling, completion, and producing activities may mask the “natural correlation.”
These activities include but are not limited to drilling across technology changes, the use of
different production practices, different production starts (well vintages), or infill vs. initial
drilling. Rock correlation will still be present, but it may not be fully recognized as one moves
from well to well. When we drill a productive area, we may be creating a well family. Although
we might expect a strong relationship between neighboring wells, we cannot depend on one.
If we have a well program that drills a number of wells sequentially, i.e., Wells A, B, C, D,
E, and F, with each well having a rank correlation coefficient of 0.8 with the preceding well, by
the time we get to well F, the correlation back to well A is equal to 0.8^5 or ~0.33, representing
a negligible correlation. Sequential correlation does not preserve correlation across the entire set
of wells. We need to reconfigure the correlation so that we correlate all wells to a dummy
distribution. As we are rank correlating, it does not matter what distribution shape we use.
Fig. 8.14 shows an illustration of central correlation. Each of the well distributions is
correlated to a dummy distribution prior to sampling and aggregation. In this case, they are all
correlated to the dummy at 0.8, which maintains a family correlation (the correlation between any
two members) at about 0.64. Correlation precision is not important. Unique individual
correlations can be handled by changing the correlation of an item with the dummy. All
relationships will track appropriately. When a group of nearby wells is aggregated, any correlation
must be taken into account; otherwise, the true upside potential of the group may be lost. It is
important to capture the distinct probability that the wells will behave as a group. This is where
this method can be very useful.
206
Fig. 8.14—Central correlation method.
8.4 Regional Aggregation

The aggregation framework described herein is applicable at the lowest to highest aggregation
levels. While Proved Reserves have a distinct definition, 1P-2P-3P categories typically have a
narrower range as compared to the 1C-2C-3C categories. Aggregation principles work across all
categories described in the PRMS. Regional aggregation is often the purview of governments and
governmental agencies, while public policy is often influenced by aggregated reserves and
resources distributions. Unfortunately, unless proper aggregation methodology is followed,
confusing or invalid ranges may be communicated to decision makers or cited by the public or
the press and relied upon to set, or react to, policy.
Methods described here in the Guidelines for Application of the PRMS can assist all
stakeholders in assessing and communicating valid and reliable estimates. It is an appropriate
public service to follow approved methods and communicate results in a coherent and industry-
accepted manner. However, there are choices when we aggregate at the regional level, and the
appropriate choices may produce results that seem unusual, with ranges that may be
unrealistically narrow or that show unusually large potential at the upper end.
We shall now examine the resource estimate for an area of unconventional (continuous)
resource as shown in Table 8.5. Examination of this published data will allow us to assess a
significant portfolio of real reservoir collections and discuss guidelines for communication of
aggregation results.
We will focus on the assessment units, which are defined in this example as oil
accumulations, as highlighted.
207
Table 8.5—Delaware Basin undiscovered resource (Gaswirth et al. 2018).
If we construct a quick stochastic simulation relying on the published data, we can examine
the various ways to aggregate the reservoir collections. Organizations often find themselves in a
tight spot when communicating aggregated values. A straight stochastic simulation, without
correlation, will typically provide an aggregate distribution that appears too narrow. Yet,
arithmetically summing the components across the ranges, which assumes an extreme rank
correlation of +1, as Table 8.6 shows, produces an aggregate range that is too wide—
underestimating the downside and grossly overestimating the upside.
Using the range data provided for each member of the aggregation, a continuous distribution
was developed that honored the data points, including the Mean, within a maximum 3% error.
The Mean is not replicated exactly, as the original distribution for the data is unknown, and it had
to be approximated. Running with a correlation coefficient of zero, we see the following results
for a sampling of 10,000 iterations:
F95 F50 F5 Mean

37000 45230 54900 45500
Table 8.6—Aggregated oil distribution based on data provided by Gaswirth et al. (2018), aggregated through sampling of
the component distributions.
The 90% confidence range from F95 to F5, or P95 to P5 as we call it (i.e., 37,000 to 54,900),
would be considered too narrow based on the experience and knowledge of most evaluators and
earth-science professionals, yet the simple arithmetic tallies (26,729 to 71,075) may seem overly
wide. Fig. 8.15 presents the aggregated volumes range for the data assuming different correlation
coefficients.
208
Range of Aggregated Volumes with Increasing
Correlation Coefficient
80000
70000
60000
Aggregated Volume
50000
40000
30000
20000
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Correlation Coefficient
F95 F50 F5
Fig. 8.15—Stochastic simulation results for the Delaware Basin data shown in Table 8.5.
With even minimal knowledge of the basin in question, we can rule out either correlation
end member as a valid representation of the aggregate volume range. There is no possibility of
reservoir existence and performance being fully correlated across the basin and throughout the
section. Similarly, the noncorrelated end member would ignore the common diagenetic history,
depositional environment, and sediment sourcing, which would tend to produce similarities
through the section.
What degree of correlation is appropriate? In this particular basin, and most basins for that
matter, a moderate correlation is likely to best represent the volume range across the vertical
section, given knowledge of the geologic uncertainties present and their shared characteristics
within a generally prolific section. Within a single horizon, especially in unconventional
(continuous) resource plays, a good correlation between individual well uncertainties is the norm.
Across a portfolio of geographically diverse, multibasin opportunities, a weak correlation would
be expected, if any at all. Note that correlation coefficients of less than 0.3 have little effect on
the aggregated outcome range.
As a general note, evaluators should be wary of localized diagenetic and seal failure issues,
especially in areas of structural complexity. The integrity and consistency of the area within a
part-play have a great deal to do with the level of well-to-well correlation.
When arithmetic summation is used, as in the above table, it is advisable to adjust the
referenced probability. Any total will have a place on the distribution and should be stated in
context. For example, if we assumed a 0.0 correlation coefficient in the example data, then the
arithmetic sum of the F5 in the table (P5 in PRMS nomenclature) equates to a result that is not
seen more than 1 in 10,000 times. The F95 in the table (P95 in PRMS nomenclature) was not seen
209
at all through one million iterations. To imply that the simple arithmetic sums define the F5 and
F95 aggregated outcomes is misleading.
If we assess the appropriate correlation coefficient between aggregation elements as 0.6, then
the arithmetically summed F5 volumes have an aggregate probability of approximately 1%, a
four-percentage-point difference in the distribution. If the arithmetic sum is to be used in this
case, it should be referenced as the F1 as opposed to F5. At the opposite end, the arithmetic sum
of the individual F95 values equates to the F99 of the aggregate when the correlation coefficient
is approximately 0.7. The upper and lower printed bounds of the aggregated data exaggerate the
high side potential and are too low at the lower end if an arithmetic sum is used. As seen in this
case, evaluators and their audiences should not assume the effect of correlation is symmetrical.
Due to the skewed nature of the component distributions, it will usually affect the upper end of
the aggregation more.
The guideline we can offer for the selection of appropriate correlation is that it must rely on
empirical evidence from similar aggregations, or detailed reservoir and diagenetic assessment. If
no such benchmark aggregation exists, aggregators are encouraged to reason out an appropriate
correlation coefficient based on geologic and anticipated reservoir performance indicators.
Aggregate range results should be tested for reasonableness through subject matter expert groups
having no stake in the outcome or use of the aggregation results. The degree of correlation used
across the aggregated assessment should always be stated.
8.5 Scenario Methods

A mechanical approach to aggregate reserves is the use of scenario methods. A “results tree” (or
scenario tree) is constructed to show potential outcomes and contingent (posterior) probabilities
of second and third aggregation elements. The technique is only suitable for a limited number of
elements being aggregated because the scenario trees get very complicated very quickly. The
main requirement is to have enough endpoints so as to mimic a fuller stochastic distribution; as
such, there is a sweet spot of three to four aggregate-able items. For any population less than that,
an adequate distribution will not be created, while for any more than that, the tree and population
process will be very complicated. The number of endpoints is equal to 3x, where x = aggregated
population. While it is usually easier to create an actual stochastic simulation, this method can
work for aggregations where the modeling skill set is limited.
8.5.1 Tree-Based Example of Correlation Between Reservoir Elements. This is an example

of the aggregation of volumes with a low degree of correlation using a scenario approach. Higher
degrees of correlation may be handled similarly. We will aggregate a set of three sands (M, N,
and S), for which the reservoir parameters and gross rock volumes are relatively independent.
The reason for this independence is that the reservoirs occur in different geological formations at
very different depths, so there are few factors that cause low and high cases of the sands to
coincide. We recommend giving consideration to the source of the reservoir data. Do all
measurements originate from the same source? Is there an incentive to have similar outcomes? If
so, then a higher correlation coefficient might be warranted. While it is always good to have the
entire range of possible outcomes represented and probabilistically symmetrical (P90-P50-P10),
this method allows subject matter experts to incorporate their interpretations and linkages directly
into the tool. Both of our cases are correlated at the reservoir level to simplify the example. In an
actual aggregation, it is more useful to identify the reservoir component that shares the uncertainty
between reservoirs. Table 8.7 gives low, median, and high recoverable reserves for the sands.
210
Volumes
Mean=
Low Median High
Expectation
M sands 17 23 30 23.3
N sands 29 41 54 41.3
S sands 10 15 25 16.7
Table 8.7—Uncertainty range for three oil-bearing sands (MMBO).
To construct the scenario tree for this situation, we have taken the low, median, and high
values of volume with probability values equivalent to those in Swanson’s mean approximation
in the sand with the largest volume, the N sand. As the N sand is dominant, this will ensure a
more reliable Mean. We then combine these first with the M sand and subsequently with the S
sand by eliciting contingent probabilities from subject matter experts. For example, assuming the
outcome of N is known, what are the probabilities of the low, median, and high results for M?
The presence of a correlation between the units is shown by a change in the probability of high,
median, or low outcomes given a known outcome of the previous sand. Note that the total
probability of the limbs leaving a node must equal 100%.
All distributions, if uncorrelated, will follow the Swanson’s Mean standard of 30% for the
low case, 40% for the medium case, and 30% for the high case. If N is high, M’s probability of
being the high case increases from the original 30% to 45%. If N is low, then M will tend to track
lower with a 45% probability of also being low. If N’s result is at the Median, then M will have
a higher chance of being at the Median. The probabilities that M will not track with N are reduced
from their noncorrelated state.
Let’s allow the S reservoir to have a higher correlation with the N reservoir. If the N result is
high, then S will have a 60% probability of also being its high case. There will be similar tracking
for the low case. Table 8.8 shows the correlated result probabilities that we will use in our tree
when we build it in the next step.
Table 8.8—Contingent probabilities given N reservoir results.
211
We set up the tree starting with the N reservoir outcome and then place the probability-
adjusted outcomes for the M and S reservoirs. We use the probabilities as shown in Table 8.8 to
complete the tree. This results in a scenario tree with 27 end branches for these three sands
(Fig. 8.16).
As can be seen in Fig. 8.16, there is some correlation between the occurrence of low, high,
and median cases for each of the sands (i.e., the probabilities that the M sands have a high value
are higher than if the N sands are high, etc.). At the end branches, we can read off the total volume
in each of the 27 possible combinations of N, M, and S sands, as well as the frequency
of occurrence.
Fig. 8.16—Scenario tree with altered limb probabilities due to correlation. All volumes are MMBO.
212
Note that each endpoint (limb) probability is the product of the individual branch probabilities
(e.g., 30% × 45% × 60% = 8.1%). The volume figures in the far-right column are the sum of the
three sands’ individual volumes (e.g., N low 29 + M low 17 + S low 10 = 56 MMBO).
If we now reorder this far-right column from smallest to largest and plot them against their
cumulative probability, the resulting plot (Fig. 8.17) yields the aggregated reserves at any desired
probability (confidence level).
Fig. 8.17—Aggregate correlated reserves distribution of probabilistic aggregation of scenario tree outcomes.
The probability of the branches and the dependencies between these probabilities, as
represented in the tree, should reflect the understanding of the geological processes at work. The
resulting aggregated distribution can then be used as a building block for a resources assessment
in the PRMS.
Caution is urged when assigning correlation coefficients between uncertainties. In the above
example, the probabilities of low, median, and high scenarios of the second and third sands were
specified given the knowledge of the N reservoir outcome. This is not the same as assigning a
correlation coefficient between the uncertainties from the beginning, but it provides a reasonably
reliable result for quick assessment of a low number of aggregated elements.
8.6 Summary—Some Guidelines

1. Deterministic 1P-2P-3P aggregation and probabilistic (stochastic simulation) methods
are compatible and together can provide reliable reserves estimates for valuation and
decision making in projects and portfolio management. Capability in both techniques is
required to adequately describe aggregated collections.
213
2. In summing 2P reserves values, arithmetically add the deterministic estimate of
quantities. Recognize that this will usually result in a value less than the real Median of
the aggregation, but it may be close enough for most 2P reserves distributions.
3. Arithmetic summation of individual categories of reserves for independent units leads
to a conservative estimate for the Proved total and an optimistic result for the upside
potential. Methods and tools are available to determine a more realistic value
(probabilistic or stochastic simulation, scenario trees, and customized tools) for
aggregation of distributions.
4. Adding individual categories of reserves probabilistically without fully accounting for
correlation could overstate the Proved total and understate the upside potential. The
expected (Mean) aggregate result will increase with correlation because the distribution
is skewed and the rate of change between percentiles is much larger at the upper end.
5. In estimating reserves quantities from well-performance extrapolation or DCA, always
work up from the lowest aggregation level (e.g., well or completion). Simple sums of
Proved Reserves from well-based DCA estimates may lead to overly conservative
estimates of reserves at the reservoir level of aggregation; hence, always check with an
overall reservoir performance extrapolation that includes development type, available
performance suitability, timing of prior activities, common bottlenecks, and market
conditions. Review the “history-to-forecast” interface to make sure that the methodology
has not introduced any discontinuities.
6. The PRMS allows probabilistic aggregation up to the field, property, or project level.
Typically, for reporting purposes, further aggregation uses arithmetic summation by
category, but these category aggregations should be labeled as the sum of individual
confidence levels and not associated with any particular probability or aggregated
confidence. A fully probabilistic aggregation of a company’s total reserves and risked
(i.e., mean success volumes multiplied by their associated probability of success)
Contingent and Prospective Resources may be used for portfolio analysis at the business
unit, region, or full company levels as long as the aggregation method handles the chance
and uncertainty components separately and provides clarity to the method of aggregation
presented.
7. For adding volumes with differing ranges of uncertainty and volumes that are correlated,
or in situations where discount factors are applied, it is easiest to use stochastic
simulation methods; however, for simple cases, a scenario approach may be sufficient.
8. When adding volumes, make sure they have a common standard of measurement
(pressure/temperature, calorific value).
8.7 Acknowledgments
This chapter is based on the framework and discussion provided by Wim J.A.M. Swinkels in the
previous version of the PRMS guidelines. Important feedback and editorial effort were provided
by Delores (Dee) Hinkle.
8.8 References
Carter, P. J. and Morales, E. 1998. Probabilistic Addition of Gas Reserves Within a Major Gas
Project. Paper presented at the SPE Asia Pacific Oil and Gas Conference and Exhibition,
Perth, Australia, 12–14 October. SPE-50113-MS. https://doi.org/10.2118/50113-MS.
214
Gaswirth, S. B., French, K. L., Pitman, J. K. et al. 2018. Assessment of Undiscovered
Continuous Oil and Gas Resources in the Wolfcamp Shale and Bone Spring Formation of
the Delaware Basin, Permian Basin Province, New Mexico and Texas, 2018. U.S.
Geological Survey Fact Sheet 2018–3073, 4 p. https://doi.org/10.3133/fs20183073.
Petroleum Resources Management System, Version 1.01. 2018. Richardson, Texas, USA:
Schuyler, J. R. 1998. Probabilistic Reserves Lead to More Accurate Assessments. Paper
presented at the SPE Annual Technical Conference and Exhibition, New Orleans,
Louisiana, USA, 27–30 September. SPE-49032-MS. https://doi.org/10.2118/49032-MS.
215
Chapter 9
Evaluation of Petroleum Reserves

and Resources
Charles Vanorsdale
9.1 Introduction
The valuation of reserves and resources is about converting a quantity forecast to a value forecast.
In that sense, valuation is part of “evaluation,” which is defined within the Petroleum Resources
Management System (PRMS 2018, Appendix A—Glossary) as: “The geosciences, engineering,
and associated studies, including economic analyses, conducted on a petroleum exploration,
development, or producing project resulting in estimates of the quantities that can be recovered
and sold and the associated cash flow under defined forward conditions.” Results from the
conversion of a quantity forecast into a cash-flow value assessment may then be used in internal
entity investment decision-making regarding commitment of funds for petroleum resources
development. The projects can be from either an existing developed project or a future planned
petroleum recovery project. The value results will be used to support public disclosures of reserves
and resources, subject to regulatory reporting requirements.
These guidelines are provided to promote consistency in project evaluations and the presentation
of evaluation results while adhering to the PRMS principles. Note that this chapter is not about fair
market value, the value upon which a willing buyer and willing seller agree, neither of whom are
compelled to act, and both of whom are knowledgeable of the relevant facts. In this context, a project
evaluation will result in a production schedule and an associated cash-flow schedule; the time
integration of these schedules will yield an estimate of marketable quantities (or sales) and future net
revenue or net present value (NPV) using a range of discount rates, including the entity’s internal rate
of return. The estimation of value is subject to uncertainty due to not only inherent uncertainties in the
petroleum in place and the efficiency of the recovery program, but also in the product prices, the capital
and operating costs, the execution of current and future environmental and regulatory requirements,
including abandonment, and the timing of implementation. Thus, similar to the estimation of a range
of marketable quantities (i.e., reserves or resources), the resulting value estimates will reflect a range
of outcomes. The PRMS allows the evaluator to include, and state, the variables that have significant
impact on the economic outcome, including uncertainties in volumetrics, product prices,
environmental, social, and governance (ESG) costs, and future capital costs.
This chapter will first summarize net cash-flow evaluation and then define and amplify key
terms and concepts in Section 9.3. Section 9.4 describes the input to an evaluation; Section 9.5
shows how to generate a cash-flow evaluation; and Section 9.6 outlines how to analyze that
evaluation. Section 9.7 compares the terms “economic” and “commercial.” Section 9.8 provides
an evaluation example and discusses other considerations. Section 9.9 briefly discusses
probabilistic evaluation, and Section 9.10 closes the chapter with ESG considerations.
216
9.2 Net Cash-Flow Evaluation
This chapter seeks to provide guidelines with which to satisfy compliance with the PRMS
principles for a net cash-flow (NCF) evaluation. Per PRMS § 3.1.1 Net Cash-Flow Evaluation:
“3.1.1.1 Project-based resource economic evaluations are based on estimates of future
production and the associated net cash-flow schedules for each project as of an effective date.
These NCFs should be discounted using a defined discount rate, and the sum of the future
discounted cash flows is termed the net present value (NPV) of the project. The calculation shall
be based upon an appropriately defined reference point (PRMS § 3.2.1) and should reflect the
following:
A. The forecast production quantities over identified time periods.
B. The estimated costs and schedule associated with the project to develop, recover, and
produce the quantities to the reference point, including abandonment, decommissioning,
and restoration (ADR) costs, based on the entity’s view of the expected future costs.
C. The estimated revenues from the quantities of production based on the evaluator’s view
of the prices expected to apply to the respective commodities in future periods, taking
into account any sales contracts or price hedges specific to a property, including that
portion of the costs and revenues accruing to the entity.
D. Future projected production- and revenue-related taxes and royalties expected to be paid
by the entity.
E. A project life that is limited to the period of economic interest or a reasonably certain
estimate of the life expectancy of the project, which is typically truncated by the earliest
occurrence of either technical, license, or economic limit.
F. The application of an appropriate discount rate applicable to the entity at the time of the
evaluation.”
Although the PRMS recognizes that there may be several methods by which to assign value
to a project, the PRMS guidelines apply only to cash-flow-based evaluations (PRMS § 3.1.0.1).
The cash-flow evaluation must incorporate elements A–F above through the following general
process:
1. Test that the project is economic.
2. Determine the project life.
a. Validate the economic viability for undeveloped projects.
b. Determine undeveloped project commerciality.
9.2.1 Step 1—Test That the Project Is Economic. Per the PRMS (§ 3.1.2.1), “A project with a
positive undiscounted cumulative net cash flow is considered economic. Production from the
project is economic when the revenue attributable to the entity interest from production exceeds
the cost of operation.”
The evaluator may generate a best estimate cash flow using a “forecast case” (which allows
for future reasonable variation, such as due to inflation or deflation, of prices, costs, market factors,
etc.), a “current economic conditions” (in which the prices/costs/expenditures, etc., are based on
historical conditions but may include reasonable inflation or deflation), or a “constant case”
(similar to the current economic conditions case but with prices, costs, etc., fixed as of the
evaluation date and without future inflation or deflation). In any type of evaluation, contractual
obligations must be honored.
217
In order for a project to be deemed economically viable, at a minimum, the best estimate
undiscounted cumulative cash flow must be positive, and undeveloped projects must also then be
assessed for commerciality, as below in Step 2b. According to PRMS § 2.1.2.2, “The
commerciality test for Reserves determination is applied to the best estimate (P50) forecast
quantities, which upon qualifying all commercial and technical maturity criteria and constraints
become the 2P Reserves.”
This best estimate cash flow must consider all activities necessary to support the project as
part of the field development plan required to yield the predicted production profile (e.g., the
forecast accounts for any additional water-handling facilities and their associated capital).
Production and revenue-related taxes and royalties must be considered (PRMS § 3.1.1, Step D).
The resulting forecast of revenues and expenses must indicate a positive cumulative undiscounted
NCF in order to identify the economic limit; consequently, the process is called the economic limit
test. The economic limit (EL) is defined as the production rate at the time when the maximum
cumulative NCF occurs for a project (PRMS § 3.1.3.1). After that point, the remaining expenses
are forecast to exceed the revenue attributable to the entity, and this situation results in the ensuing
cash flow being negative. Costs for abandonment, decommissioning, and restoration (ADR) as
well as income taxes (other than production taxes), any overhead not required for the specific
project, and depreciation are excluded at the EL test stage of the evaluation (PRMS § 3.1.3.2) to
verify that the project is economically producible. The impact of ADR, taxes, and depreciation
will be addressed later in this chapter.
Production forecasts for either developed (ongoing) or undeveloped projects must
demonstrate economic producibility. Undeveloped projects must also satisfy economic viability,
as below in Step 2a.
9.2.2 Step 2—Determine the Project Life. Having performed the EL test, the evaluator has
determined the date beyond which further operation of the project will no longer have a positive
NCF to the entity. However, per the PRMS (§ 3.1.1, Step D), the project life “is limited to the
period of economic interest or a reasonably certain estimate of the life expectancy of the project,
which is typically truncated by the earliest occurrence of either technical, license, or economic
limit.” Consequently, if the EL occurs after a technical constraint is reached or the
license/concession expires, then the project life (and NCF to the interest of the entity) must be
truncated by the appropriate limit.
9.2.2.1 Step 2a—Validate the Economic Viability for Undeveloped Projects. Projects that
are undeveloped have a further criterion: The project’s cumulative undiscounted NCF, truncated
as above, must exceed the projected ADR liability. (Remember, ADR costs are not included in
assessing economic producibility or the EL calculations.) If this condition is met, then economic
viability is confirmed (PRMS § 3.1.2.5), and the undeveloped project’s forecasted recovery will
be subjected to commerciality assessment to determine if it might qualify as Reserves. For ongoing
projects, the threshold of economic viability has been satisfied once the economic producibility
check of Step 1 is performed.
9.2.2.2 Step 2b—Determine Undeveloped Project Commerciality. With this step, the cash-flow
evaluation expands in scope. The PRMS spells out seven criteria required to achieve commerciality
(PRMS § 2.1.2.1), one of which is that there must be a “reasonable assessment that the development
projects will have positive economics and meet defined investment and operating criteria.” The best
estimate case NCF prepared previously, honoring the earliest truncation by contract term, operational
limitations, or the EL, is then adjusted for the time value of money by using a discounting factor.
218
This evaluation, performed for undeveloped projects, will include the ADR liability and any other
of an entity’s internal criteria (beyond PRMS requirements) such as corporate income taxes. The
resulting discounted cash flow (DCF), when summed, yields the project’s NPV. The discounting
factor is usually taken to be equivalent to the evaluating entity’s hurdle rate or weighted average cost
of capital (WACC) (PRMS § 2.1.2.1). Therefore, if the project’s NPV is less than zero, then the
project fails the commerciality test, and the projected recovery, although economic, cannot be
classified as Reserves. All other internal decision criteria must be satisfied to classify the project’s
forecasted recovery as Reserves. For example, the entity decision criteria might also include
achieving target lifting costs, return on investment, or other profitability metrics, which need to be
considered to classify the project’s projected recovery as Reserves.
A cash-flow evaluation showing that a project will return the entity’s required hurdle rate or
WACC provides supporting evidence of the entity’s commitment to the project as required in
PRMS § 2.1.2.1.
Consequently, under the PRMS, a project assessed to be commercial is necessarily economic.
In fact, because Reserves are defined as commercially recoverable quantities, assessing the
economic status of the forecast production is a prerequisite to determining commerciality and
assigning Reserves.
Petroleum reserves and resources evaluations therefore require integration of
multidisciplinary expertise in both the technical and the economic/commercial areas throughout
the life cycle of asset development. Petroleum resources, as they progress to Reserves, will change
class and/or categories throughout their life cycle due to revision of technical and commercial
uncertainties, such as through drilling activity, data acquisition and interpretation, changing costs
(among them operating, capital, and sunk costs) and product prices, changing fiscal and/or
commercial terms, and development/adoption of new technology.
9.3 Terminology of Cash-Flow-Based Evaluation

The preceding section gives the impression that there is a language specific to the cash-flow-based
evaluation of petroleum resources, and indeed certain terminology—such as with the resources
definitions themselves—must be clear in order for resulting evaluations to be consistently
performed and/or compared and conveyed using that information (whether internal or external).
This section will briefly discuss the key terms associated with a cash-flow-based evaluation as
addressed in PRMS § 3.1.2. (Section 9.7 provides a more thorough discussion of the difference
between the terms “economic” and “commercial.”)
9.3.1 Net Cash Flow (NCF). A cash flow is the calculation, at discrete intervals (typically reported
on an annual basis but perhaps calculated on a monthly basis), of the cash inflow (e.g., revenue
from product sales) and the cash outflow (e.g., operating costs, capital expenditures, taxes, etc.)
attributable to a specific project. The NCF is the difference between the inflow and the outflow to
the entity’s interest. As shown in Fig. 9.1, the annual NCF is depicted by the bars. (Factors not
depicted individually are revenue, operating costs, and capital expenditures, for the simplicity of
presentation.) The cumulative undiscounted NCF and cumulative oil recovery are represented by
the red and black lines, respectively. We can see in Fig. 9.1 that the cumulative NCF for the first
four years is negative, indicating that the sum of the capital expenditures and operating costs for
those four years exceeds the revenue. This situation is reversed in year five, when the cumulative
NCF becomes positive.
219
Fig. 9.1—Undeveloped project forecast for EL test. (Note that economic viability determination by comparison of maximum
NCF to ADR obligation is not shown.)
9.3.2 Abandonment, Decommissioning, and Restoration (ADR). ADR is defined (PRMS

Appendix A—Glossary) as “The process (and associated costs) of returning part or all of a project
to a safe and environmentally compliant condition when operations cease. Examples include, but
are not limited to, the removal of surface facilities, wellbore plugging procedures, and
environmental remediation. In some instances, there may be salvage value associated with the
equipment removed from the project. ADR costs are presumed to be without consideration of any
salvage value, unless presented as ‘ADR net of salvage.’”
ADR expenses are not used in generating the NCF when estimating the EL, but they are
included when performing project economic analysis (unless specifically excluded by contractual
terms). Close to the EL, these costs are considered to be an obligated liability that is incurred
regardless of the economic producibility (i.e., the net revenue exceeds the net expenses to the
entity’s interests) of the project. However, if the cumulative NCF (including ADR costs) of an
undeveloped project fails to yield a positive figure, then the estimated quantities are not
economically viable (PRMS § 3.1.2.5) and cannot qualify as Reserves.
In the case of production sharing contracts (PSCs), the PRMS notes (§ 3.3.2.4) that “the terms
governing cost recovery in a particular PSC may require special treatment of items such as taxes,
overhead, and ADR to determine entitlement.” The entitlement determination must be evaluated
to be able to forecast the net revenue for the project, and ADR may not be required as part of the
economic assessment for PSCs if, for instance, the host government accepts responsibility when
economic production is anticipated to continue past the contract termination date.
9.3.3 Economic Limit (EL). PRMS § 3.1.3.1 describes the EL as the production rate at the time
when the maximum cumulative NCF occurs for a project. Continuing to produce beyond the EL
220
results in negative cash flows after the EL date, and therefore the project is uneconomic from that
date onward.
For either undeveloped or developed projects, ADR expenses are not used in the NCF when
estimating the EL because economic determination of the project is not the objective at this step
(however, see the comments for undeveloped projects two paragraphs ahead). The EL of
production for the project is based on the defined economic conditions. The use of alternative
economic scenarios will likely alter the time and production rate, thereby changing the cash flow
establishing the EL. The EL at the forecast production of the low, best, and high estimates are
typically different because they have different cash flows (i.e., they have different sales quantities
and therefore different revenues). In the absence of other limits (e.g., technical/operating
conditions, concession/license term, etc.), the EL defines the production rate beyond which the
project is no longer economically producible. The cumulative production to that EL defines the
economically producible quantity subject to the final project economic determination.
It is possible that there may be periods of negative NCF during the life of a project (such as
forecasted major capital expenditures, development/infill drilling, or during operational
maintenance). Interim negative project NCF periods may be accommodated provided that the sum
of subsequent positive NCF exceeds the sum of the negative NCF, thus continuing to increase the
maximum cumulative NCF of the project. With this confirmation, the entire time period is
considered the economically productive life of the project up to the time beyond which there is no
increase in the cumulative NCF.
For an undeveloped project, once the EL is determined for the best estimate case, the
cumulative NCF at that limit must be compared against the estimated ADR cost associated with
the project (PRMS § 3.1.2.5). If the estimated ADR liability exceeds this cumulative undiscounted
NCF, then the forecast production is not economic and cannot be recognized or booked as 2P
Reserves. The same economic check must be applied for the low estimate case to claim 1P
Reserves, but using the low estimate (Proved case) EL.
In many cases, projects must be combined for determining their common EL. Projects may be
at various stages of development, and, once the undeveloped projects are confirmed as economic
on an incremental basis, those projects are then combined with the developed projects for the
combined EL determination, often at a field level (PRMS § 2.1.3.6.3). This typically may occur in
additive projects (such as developing and producing a secondary reservoir in a field where the
primary reservoir is already producing), in which an ongoing producing project may have its EL
affected by the incremental project, or for certain projects that use a common production facility.
Multiple projects that share the expense of common major capital expenditures can facilitate the
extension or truncation of Reserves booking or even enable the booking of Reserves for a project
that, by itself, would not generate sufficient revenue to overcome the cost of installing the facilities
necessary for its production. In any event, the undeveloped project or combined projects must pass
the EL test to move forward to commerciality assessment.
9.3.4 Economic Conditions. As noted in Section 9.2, the EL is dependent upon the economic
conditions assumed in the project evaluation. Generally, such economic conditions are expressed
in either a “current economic conditions case,” “forecast case,” or “constant case.”
The “current economic conditions” case (PRMS Appendix A—Glossary) refers to the use of an
arithmetic average of the product pricing and operating expenses in the calculation of the NCF, using
the average of the prior 12 months of actual data, i.e., the 12 months prior to the effective or
evaluation date. (For undeveloped projects with no economic history, carefully selected analogs may
221
be used to estimate the economic conditions.) A shorter time frame may be necessary if, in the course
of the prior 12 months, there has been a step change in prices and/or costs that is expected to continue.
One such situation may occur when operators experiment with drilling and completion techniques
to find and consistently implement the most profitable combination. Once the optimum technique is
found, this approach will be done consistently going forward. This repeating of a consistent method
may cause costs and/or timing to experience a “learning curve” effect (PRMS § 3.1.2.3) that will
enable lower future values to be recognized. A current economic conditions case may incorporate
reasonable inflation or deflation scenarios.
The forecast case, on the other hand, permits the use of price and cost escalation with time.
The forecast case must be founded on assumptions that are considered reasonable by the evaluator
to exist during the course of the forecast period or life of the project. This case may incorporate
inflation or deflation effects (PRMS § 3.1.2.2). Prices and costs should not be escalated indefinitely
and should be limited either by a cap value or time period. Because the forecast case, with its
escalation effect, is commonly used, it is often referred to as the “base case” scenario.
The constant case is similar to the current conditions case, but it holds all product pricing and
costs constant, i.e., has no inflation or deflation effects (PRMS § 3.1.2.6).
In all of these NCF case options, any contractual obligations that have defined costs or pricing
override any inflation/deflation assumptions because the contract terms apply instead.
The economic conditions must not vary among categories within a resources class. The
product price scheme used in evaluating Proved Reserves must be the same for the other Reserves
categories; otherwise, a particular category will exhibit an economic bias when the categories are
used to reflect primarily technical uncertainties. Adopting, for example, different pricing scenarios
within a resources class is known as “split conditions” (PRMS § 2.2.0.3) and is not permitted in
the application of the PRMS. Likewise, when the same economic conditions are applied to each
category, and the best estimate case is economic but the low estimate (Proved) case is not, then
the best estimate case may go on to pass the other commerciality requirements and become 2P
Reserves; however, the uneconomic Proved case cannot be Reserves, and so no 1P Reserves may
be booked (PRMS § 3.1.2.8).
A given project, however, is based on a certain development plan and commercial maturity
level, upon which investment decision-making is founded. Consequently, the project cannot consist
of both Reserves and Contingent Resources because they represent different commercial maturities.
A project containing multiple resources classes is not permitted under the PRMS (§ 2.2.0.4) because
this constitutes “split classification.” Where economic production exists beyond the limit of a
concession term, decisions must be made between the parties at or before the time of the concession
contract expiration to proceed with continued production. If there is no evidence of historical contract
extensions between the parties, or the concession terms do not call for automatic extension, then this
economic production would be a new, undeveloped project consisting of Contingent Resources;
otherwise, the quantities may be booked as Reserves without split classification.
9.3.5 Discount Rate. The determination of an EL relies on a NCF, excluding ADR and income
taxes. Once the EL has been found, and the forecasted production has been deemed economic, the
undeveloped project must be assessed for commercial viability for the defined conditions in order
to recognize Reserves. A project’s NCF will be discounted at a rate chosen by the entity to equal
or exceed its minimum attractive rate of return, which generally reflects the entity’s WACC
(PRMS §2.1.2.1). When a discount rate is applied to the NCF evaluation, the resulting cumulative
time value reflects the NPV of the future cash-flow stream. For example, if the entity’s decision
222
criteria require achieving a 10% WACC, then the resulting NPV discounted at 10%/year must
equal or exceed zero to satisfy the required criteria. If the criteria are not met, then the entity may
decide to not fund the project, and the volumes are not commercial and therefore cannot be
classified as Reserves.
9.3.6 Other Key Terms. “Economic interest” (also known as entitlement interest or “mineral
interest”) represents the share, right, or title in property (a lease, concession, or license), project,
asset, or entity. It typically also represents the percentage of the forecast sales production in the
entity resources NCF. There are several forms of economic interest, depending on the specific
regime. An economic interest in a concession regime is generally simpler to determine and may
be represented by the working interest of the entity, after deducting royalties and any interests
owned by others. In more complex regimes, such as PSCs, typically there is a variable economic
interest in the production, caused both by cost recovery mechanisms and possibly by variable profit
petroleum fraction (with values often listed in tables). This variable interest of “entitlement” for
an evaluated party cannot be calculated by simply multiplying, for example, total project volumes
or revenue (whether before or after the deduction of any royalties owed to others) by the party’s
working interest. Chapter 12—Resources Entitlement and Recognition herein describes forms of
economic interest to be considered for both the entitled quantities and for the share of production
revenue in the resources cash flow.
“Royalties” are the payments made to the mineral rights owner for the right to explore for
and produce petroleum after a discovery. Depending on the country’s regulations, mineral rights
may belong to the host government or other mineral rights owners (lessors). In agreements where
royalties are taken in kind (e.g., in terms of production) or “cash or kind,” the royalty share must
be deducted from the entitled quantity of the entity (and the entitled production in the NCF). This
treatment is proper in cases where the host or lessor retains a mineral interest in the related volume.
In agreements where payments do not relate to host or lessor ownership of a mineral interest, and
the payment termed royalty is paid in cash, the default position is to not deduct royalty from the
entitled quantity of the entity (and the entitled production in the NCF). The foregoing discussion
is focused on reporting regulations that require disclosure of net volumes after deduction of
royalties owned by others, which is the objective of the entitlement quantities–based PRMS.
However, some jurisdictions may require reporting of gross volumes before deduction of royalties,
depending on the specific purpose and use of the information reported.
Many agreements allow for the producer to lift the royalty volumes, sell them on behalf of the
royalty owner, and pay the proceeds to the owner. In such cases, the royalty quantity may also be
deducted from the entitled quantity of the entity. In some agreements, royalties are required to be
paid in cash to the mineral rights owner. There are also agreements where cash payments to the host
government are termed royalties, but they are recognized as production taxes and are not associated
with mineral rights ownership. If this is the case, then the entity’s resources NCF royalties shall not
be deducted from the entity’s economic interest, but those “production taxes” should be treated as
an expense in the entity’s NCF when determining the EL and in its DCF when recognizing project
economics and the entity’s associated Reserves. Chapter 12—Resources Entitlement and
Recognition herein contains further description of royalties.
“Royalty interest” is a mineral interest that typically is not burdened with a proportionate share
of investment and operating costs. Royalty owners are responsible for their share of production
and ad valorem taxes (i.e., taxes imposed based on production value and/or value of equipment
necessary to produce petroleum). Royalty interest may also be defined as the share of minerals
223
reserved in money, or in kind, free of expense, by the owner of mineral rights when leasing the
property or contracting another party for exploration and production.
“Overriding royalty interest” is a revenue interest, free of any cost obligation, created by the
operating entity and/or working interest owner and paid by the operating entity and/or working
interest owner out of revenue from the property. This differs from royalty interest because
ownership of the minerals is not the basis for this interest, but rather it is another form of economic
interest in the property.
“Fiscal systems or arrangements” made between the producer and the host government may
include concession agreements, joint venture agreements, PSCs, and other contracts (refer to
Chapter 12—Resources Entitlement and Recognition herein).
9.4 Required Input for a Cash-Flow-Based Evaluation

Proper evaluation calls for the preparation and validation of certain data specific to each project,
generally including technical constraints (wells, facilities, flowlines, etc.), product prices at the
sales/reference point(s), capital expenditures, operating costs, ADR costs, ownership/entitlement
and royalties, taxes, and legal/contract/fiscal terms (PRMS § 3.1.1.1).
9.4.1 Net Entitlement Sales Production Forecast. The net entitlement sales forecast is the start
of the resources NCF, representing the entity’s share of sales quantities that will generate sales
revenues in the NCF when multiplied by product prices. This sales production forecast is measured
and reported at the reference point, which is typically the point of sale of the marketed quantity.
There is a different net sales forecast for each resources category, which will generate different
sales revenues. The entity’s net sales production forecast is directly affected by the economic
interest of the entity in the project, which depends on the contract model. Chapter 12—Resources
Entitlement and Recognition provides more description on the economic interest of contracts and
agreements such as concessions.
9.4.2 Product Prices. Product price forecasts used by the evaluator to apply to the commodities in
the future must consider the impact of any sales contracts or price hedges. Fiscal arrangements made
between the producer and the host government also may be part of the actual price received for the
products. Whatever method is used for setting future prices (forward strip, internal entity estimates,
contractual specifications, fixed historical averages, etc.), they are to be applied at the reference point
(typically the sales point; see Chapter 11—Production Measurement and Operational Issues herein)
for the project and should be referenced to an appropriate benchmark (e.g., Brent, West Texas
Intermediate, Arab Light, etc.). Historical benchmark pricing is readily available on industry
websites, but these reference a particular quality of product. The project being evaluated likely will
have a different commodity quality, and so there will be a product price differential associated with
the crude quality (API gravity, sweet/sour), natural gas liquid, or gas heating value (calorific content)
that must be included when determining the applicable product price.
For gas, it is important to look at the gas composition at the reference point to ensure that the
correct differentials are being applied. If the reference point is after gas processing, dry gas and
each byproduct (e.g., propane, butane, and condensate) should be evaluated with the appropriate
price forecast.
This differential also may be based on the physical sales or reference point and the market for
the product in that area. Often, actual historical price differentials from the sales point may be
224
acquired from a lease operating statement, to ensure that the correct differentials are being applied
and to tie the evaluation to the forecasted price differentials.
The sales price may also be impacted by certain costs. “Marketing and transportation costs”
are the costs incurred to deliver a product to its sales point, and they are to be identified either
as part of the operating costs or as a reduction of the sales price if the sales point is not at the
wellhead. Both methods typically result in the same economic outcome, but they will result in
operating costs and sales price differences, and the approach must be consistent with the
financials treatment. The evaluators must clearly identify and document all assumptions used in
the evaluation price forecast because this will directly impact the projected quantities eligible
for classification as Reserves and Resources.
9.4.3 Project Capital Costs. For undeveloped and partially developed projects, any cash-flow
evaluation requires an assessment of the development plan’s future capital costs, based on an
evaluation of the development expenditures for that specific resource category. In this context, the
development costs include all the necessary project capital for wells and facilities to enable delivery
of petroleum attributed to the resource category from the accumulation(s) to a product sales point
(or to an internal transfer point between upstream operations and midstream/downstream operations).
Capital expenditures (CAPEX) are major expenses that typically encompass items such as
land acquisition, exploration, drilling and well completion, and surface facilities (gathering
infrastructure, process plants, and pipelines). Additionally, they may include environmental and
social costs required to be incurred related to the resources category in the economic evaluations.
Well drilling and completion costs are categorized in terms of tangible (subject to depreciation
allowance) and intangible (expensed portion and portion subject to amortization) well costs (for
companies that follow the U.S. Generally Accepted Accounting Principles, of GAAP; companies
reporting under the International Financial Reporting Standards, or IFRS, must consult the
reporting principles used in their respective countries). Surface facility costs are subjected to
facility-specific depreciation allowances used in calculating taxes and various incentives. This
level of detail regarding depreciation, depletion, and amortization (DD&A) usually does not enter
into any cash-flow analyses for economic and commercial assessment, but rather relates to
accounting, public disclosure, and reporting for tax purposes.
When a development plan for the project is either incomplete or not available (which may be
the case for contingent resources), CAPEX estimates may be made using one of two approaches:
• Top-down approach—This approach uses historical data from similar development
projects to estimate the costs for the current project by revising and normalizing these
data for changes in time (inflation or deflation), production quantity, facility capacity,
location, and other factors. It uses a simple “percentage-of-cost basis” established from
the review of historical or current data.
• Bottom-up approach—This is a more detailed method of cost estimation that requires a
detailed design that breaks down the processing equipment into small, discrete, and
manageable parts (or units). The smaller unit costs are summed (including other
associated costs) to obtain the overall cost estimate for the processing equipment.
Total capital investment cost required for any process equipment (or plant with several units
of equipment) is generally recognized under four categories (Clark and Lorenzoni 1978;
Humphreys and Katell 1981) when there are tax considerations:
• “Direct costs” include all material and labor costs associated with a purchased physical
plant or equipment and its installation. They include the costs of all material items that
225
are directly incorporated in the plant itself as well as those bulk materials (such as
foundation, piping, instrumentation, etc.) needed to complete the installation.
• “Indirect costs” represent the quantities and costs of items that do not become part of, but
are necessary costs involved in, the design and construction of process equipment.
Indirect costs are generally estimated as “percentage of direct costs.” Indirect costs are
further subcategorized as engineering, constructor’s fee (covering administrative
overhead and profit), field labor overhead, miscellaneous others and owners’ costs (such
as land, organization, and startup costs). Engineering indirect costs include the costs for
design and drafting, engineering and project management, procurement, process control,
estimating, and construction planning. Field labor overhead includes costs of temporary
construction consumables, construction equipment and tools, field supervision, and
payroll burden, etc.
• “Miscellaneous others” include freight costs, import duties, taxes, permit costs, royalty
costs, insurance, and sale of surplus materials.
• “Contingency” is included to allow for possible redesign and modification of equipment,
escalated increases in equipment costs, increases in field labor costs, and delays
encountered in startup.
Further, “working capital” is needed to meet the daily or weekly cost of labor, maintenance,
and purchase, storage, and inventory of field materials.
9.4.4 Operating Costs. These expenses are generally recognized under five categories
(Humphreys and Katell 1981):
• “Direct costs” are considered to be dependent on production and include variable and
semivariable components. Costs are variable when related to produced volumes (or
injected volumes for water-disposal or supplemental recovery projects). Such costs are
typically expressed in cost per unit volume (e.g., USD/bbl). Variable costs will be zero
when there is no production/injection. Examples include lifting costs and chemical
treatment costs. Semivariable costs are related to producing operations but not directly
tied to volumes. These costs have a fixed component that does not go to zero when
production ceases (e.g., plant shutdown). Examples are costs to monitor and maintain
wells and facilities. These costs go up as production operations grow (e.g., more wells
and facilities) but do not directly tie to produced volumes. These costs are often expressed
as cost per unit of operation such as USD/well/month.
• “Indirect costs” are considered fully independent of production and, thus, fixed costs.
These include plant overhead, or burden, and other fixed costs such as property taxes,
insurance, and depreciation. (Note that depreciation, the reduction of asset book value
with time, is typically part of DD&A charges used in accounting practice and is not, as
mentioned in Section 9.4.3, above, per se part of the cash-flow analyses performed by the
reserves evaluator, except indirectly, via the impact of these DD&A charges on tax
considerations.)
• “General and administration expenses (G&A),” or simply overhead expenses, are those
costs required to operate the subject property. These costs must be actual incremental
costs attributable to the project (PRMS § 3.1.3.2).
• “Distribution costs” are those operating and manufacturing costs associated with
transporting the products to market, like pipelines for crude oil, gas sales, and natural gas
226
liquids. They include the cost of containers and packages, freight, and operation of
pipelines, terminals, and warehouses or storage tanks.
• “Contingencies” constitute an allowance made in an operating cost estimate for
unexpected costs or for error or variation likely to occur in the estimate. A contingency
allowance is just as important in the operating expenses as it is in the CAPEX. However,
it must be pointed out that companies may define and categorize their operating costs
differently and may not even include some of the components in their project economic
analysis if they are not directly related to the project scope for that category.
9.4.5 Ownership and Royalties. Entitlement, as characterized by ownership (or economic)

interests and royalties, is covered in more detail in Section 9.3.6 above and in Chapter 12—
Resources Entitlement and Recognition herein. When evaluating reversionary interests based on a
payout quantity, the NCF will be used to determine when this occurs. For example, a partner in a
project may have a certain working interest until the time at which the revenue from the project
equals the capital investment, subject to the agreement payout terms, to bring the project on
production. Known as a Working Interest before Payout (WIBPO or just BPO), this interest then
reverts at payout to a Working Interest After Payout (WIAPO or just APO). The same situation
may exist for the net revenue interest attributable to the partner. Whether the reversion is from a
large interest to a smaller interest or vice versa, the result is a change in the percentage of
hydrocarbons to the interest owner, which must be reflected in not only the reserves, but also the
project NCF.
9.4.6 Taxes. Taxes that are imposed based on production value and/or value of equipment
necessary to produce petroleum must be included in the NCF. Examples are production taxes that
may include severance and ad valorem taxes in the US. Income taxes or profit-based taxes (which
may include certain types of federal and state taxes) are not included in the determination of the
entity’s EL. (Severance tax may be assessed, usually on a state level, based on either the volumes
produced, e.g., USD/bbl, or on the value as a percent of revenue. Ad valorem taxes are assessed,
usually at the county level, as a percentage of the revenue.) Similar taxes may be assessed in
different countries around the world.
9.4.7 Legal/Contract/Fiscal Terms. The revenue and cost components of any term described
above (including all other relevant economic and commercial terms) may be defined differently
from country to country due to the fiscal arrangements made between companies and host
governments, which allocate the rights to develop and operate specific oil and gas businesses.
Common forms of international fiscal arrangements are concessions (through royalties and/or
taxes), PSCs, and risk service contracts (see Chapter 12—Resources Entitlement and Recognition
herein). In general, these agreements define how project costs are recovered and how profit is
shared between the host country and the entities that take a working interest in the agreement.
Detailed knowledge of these governing regulations (regarding, for example, the permitted period
of activity; any royalties, income taxes, or other fiscal items; any other obligations, such as the
ADR funding mechanism or fiscal incentives) is critical for a credible project reserves assessment
and evaluation process.
Additionally, the same can be said of agreements between project co-owners, which can
impact an evaluated entity’s cash flows, for example, farm-ins or farm-outs and other situations
where the evaluated partner’s monetary entitlements/obligations cannot be determined simply as
227
its percentage working interest in the project multiplied by the total monetary amount in question.
Fiscal terms and other conveyances are covered in more detail in Chapter 12—Resources
Entitlement and Recognition herein.
9.5 Generating a Cash-Flow-Based Evaluation

To reiterate, the NCF evaluation should be performed as follows.
1. Production forecasts of the primary and secondary streams are generated per project
and/or per category (e.g., 1C, 2C, or 3C, Development Pending sub-class).
2. The economic conditions (current conditions, forecast case, or constant case) under which
a NCF is to be created must be decided.
3. An undiscounted NCF analysis is performed to determine the economic viability and EL
for the project. This NCF ignores income tax and (except for certain situations such as
with PSCs and as noted in step 5 below) ADR liabilities.
4. The EL must be compared against other constraints, such as operational limitations or
concession terms, and the earliest occurrence of these must be applied to constrain the
production forecast and the NCF.
5. For an undeveloped project, if the cumulative NCF exceeds the estimated ADR liability,
then the project is considered economically viable (PRMS § 3.1.2.5). For developed
projects, ADR is excluded from the economic producibility determination (PRMS §
3.1.2.1).
6. If economically viable, the commerciality of the undeveloped project must then be
assessed, starting with the application of a discount rate equal to or greater than the
entity’s WACC. The cumulative discounted NCF (the DCF), which is the sum of the
annual discounted cash flows, yields the project’s NPV, which then will be compared
with the entity’s investment decision criteria. (Note: Discounting may result in an EL
different from that determined in step 3.)
7. If the project meets or exceeds the entity’s hurdle criteria (see Section 9.6.1), and the
project satisfies all other criteria used to pass the commerciality test (as listed in PRMS §
2.1.2.1), the projected recoverable quantities attributable to the project may be classified
as Reserves.
With commercial software or the use of third-party consulting services, generating the NCF
for a project is quite easy from a mechanical standpoint. However, it is important to understand
the information required to provide the input to these calculations.
Table 9.1 below shows an example of a NCF/DCF template with the basic output. In this
example, ADR is included within the capital costs, but income taxes are excluded.
Reading the template from left to right, we see that most NCF reports employ an annual
convention, i.e., all output is in terms of an annual result (or perhaps an average in the case of the
product prices). The gross production is then tabulated by product, such as crude oil and associated
gas, or non-associated gas and condensate. As further discussed in Chapter 12—Resources
Entitlement and Recognition herein, the evaluation is conducted for the interest to which the entity
is entitled, according to the entity’s net revenue interest (NRI), expressed as Net Production =
(Gross Production × NRI), which may vary by product and the entitlement terms.
228
RUN DATE 2/1/2019
RUN TIME
RESERVES AND ECONOMICS
CLIENT MEGA OIL PREPARER CRV

FIELD/RESERVOIR BARNBURNER AS OF DATE December 31, 2018
PRICES OPERATIONS, M$ 10%

GROSS PRODUCTION NET PRODUCTION OIL GAS NET OPER SEV+ADV NET OPER CAPITAL CASH FLOW CUM DISC
YR END OIL, MBBL GAS, MMCF OIL, MBBL GAS, MMCF $/BBL $/MMBTU REVENUES TAXES EXPENSES COSTS, M$ BTAX, M$ BTAX, M$
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
SUB
REM.
TOTAL
CUM PRESENT WORTH PROFILE

DISC PW OF NET DISC PW OF NET
EUR RATE BTAX, M$ RATE BTAX, M$
0 xx 40 xx
5 xx 45 xx
WORKING INTEREST, % 100.0 NET OIL REVENUE (M$) xxx 10 xx 50 xx
REVENUE INTEREST, % 87.5 NET GAS REVENUE (M$) xxx 15 xx 60 xx
TOTAL REVENUES (M$) xxx 20 xx 70 xx
25 xx 80 xx
BTAX ROR, % xx PROJECT LIFE (YRS) xx 30 xx 90 xx
BTAX PAYOUT YEARS xx DISCOUNT RATE (%) 10 35 xx 100 xx
Table 9.1—Example NCF analysis.
Next, we have columns containing the annual product reference-point prices, which may be
held constant or escalated in accordance with a forecast scenario. As described in Section 9.4.2,
these prices should be adjusted for product quality, e.g., whether light or heavy crude, high- or
lower-BTU gas, etc., and they must honor any contractual pricing specifications.
Continuing to the right, the next columns are for the calculation of revenue and operating costs
for the project. Revenue, net to the evaluated interest, is calculated first (net sales production
multiplied by product price), and taxes are then estimated. The taxes column in Table 9.1 specifies
“severance & ad valorem” taxes, but the heading more generally could be called “production taxes.”
Net operating expenses refer to the charges to the appraised interest for maintaining the
project, whether on production or not. These costs include fixed and variable items, as has been
discussed previously in Section 9.4.4. In addition, as pointed out in the PRMS § 3.1.3.2, “Operating
costs may be reduced, and thus project life extended, by various cost-reduction and revenue-
enhancement approaches, such as sharing of production facilities, pooling maintenance contracts,
or marketing of associated nonhydrocarbons.”
The handling of lease use fuel (consumed in operations, CiO) volumes must be done properly
to avoid undercounting expenses or overcounting sales revenues. PRMS § 3.2.2.3 states that CiO
represents an operating expense reduction when used in place of fuel purchased from external
parties, although the CiO quantities themselves are not included in the project economics; in other
words, these quantities must not be treated as a revenue stream or as a cost incurred for purchase
to recognize a sales quantity. This is easily handled if the CiO quantities are not included in
229
reserves or resources. However, if an entity chooses to include CiO quantities as reserves or
resources, then the quantities must be stated and recorded separately from sales quantities to ensure
no revenue will be calculated from these CiO quantities.
Next, we have capital expenditures, which are the major expenses (as discussed previously in
Section 9.4.3) for undeveloped projects, including ADR for undeveloped projects (and even
developed projects, depending on the entity and its use of the cash flow). These costs have a greater
impact on the net present value when they occur earlier in the life of the project.
Following these columns, we have the annual undiscounted net cash flow, designated here as
“BTAX” or before tax, specifically (since this is a US project evaluation), federal income tax (more
appropriately a “BFIT” evaluation). This column is simply the summation of the annual operating
revenues less the production taxes, operating expenses, and capital expenditures.
The final column is the discounted cumulative net cash flow, in this example, using a 10%
discount rate based on the entity’s WACC. As can be found in most economic analysis textbooks,
the present worth of a single payment (for the cashflow here, the annual NCF is discounted) may
be calculated using the formula:
𝑃𝑃 = 𝐹𝐹(1 + 𝑖𝑖)−(𝑛𝑛−0.5) , ......................................................................................................... (9.1)
where
P = the present value of a future cash flow
F = the future value of cash flow
i = discount rate per interest period (decimal)
n = number of interest periods
(1 + 𝑖𝑖)−(𝑛𝑛−0.5) = discount factor assuming midyear discounting, or (1 + 𝑖𝑖)−𝑛𝑛
for an annual basis.
F = production term in the material balance equation, bbl.
Further, the discount rate may be expressed either as an annual or monthly rate. For more detail
on the monthly rate usage, reference may be made to Society of Petroleum Evaluation Engineers
Recommended Evaluation Practice 5 (Society of Petroleum Evaluation Engineers 2002).
The DCF output typically should include:
1. Identity of the project being evaluated, including
a. Entity
b. Field and reservoir(s)
c. Other identification as necessary
2. Identification of the evaluator
3. Resources class and category
4. Effective date of the evaluation
5. Discount rate applied in the evaluation
6. Annual output of columnar data described above (Table 9.1)
7. Subtotal of output of columnar data described above (Table 9.1)
8. Any remaining output for the columnar data described above beyond the tabulated time
frame
9. Total of items 6 and 7 for each column of data
10. Cumulative production, if any, of the evaluated products
11. Estimated Ultimate Recovery, being the sum of items 9 and 10 to the evaluated interest
230
12. Disclosure of application of federal/governmental income tax
13. The evaluated working and net revenue interests (reversions noted as necessary)
Optional output could include the per-product net revenue, calculated internal rate of return,
payout time, a present worth table showing the effects of varying discount rates on the project
NPV, and any concession/contract term limit.
9.6 Analyzing a Cash-Flow-Based Evaluation

At this stage, it is important to reiterate two points:
1. As mentioned previously (Section 9.2), the PRMS guidelines and definitions may be used
to estimate project value by cash-flow analyses only. Other value measures, such as
historical costs, comparative market values, key economic parameters, etc., are not
included in the PRMS guidelines.
2. The PRMS-based DCF evaluation is not a fair market value calculation; it is a net present
value estimation based on the defined economic conditions and procedures further set out
in the PRMS guidelines. The PRMS establishes no criteria for transactional value
assignment. The PRMS-based DCF evaluation may be used as a basis for the further
quantification of a fair market value, but such value is dependent upon many criteria (e.g.,
current market conditions, desirability of certain acreage position, buyer/seller
motivation, etc.) outside a principles-based system such as the PRMS.
The cash-flow-based evaluation is a necessary first step in establishing the economic viability
of a project. Once the DCF has been generated, it must then be analyzed relative to the entity’s hurdle
criteria for the project to move toward classification as Reserves (provided other contingencies have
been satisfied).
9.6.1 Investment (Hurdle) Criteria. An investment decision is dependent upon several criteria,
each specific to the entity by/for whom the evaluation is performed. The majority of the
profitability indicators are readily derived from the DCF evaluation. Such profitability indicators
are also used in portfolio ranking to plan the development sequence of projects to yield the greatest
financial return to the entity and its shareholders, if any. In addition to fulfilling the WACC hurdle
mentioned above, an entity may consider other criteria, including (but not limited to):
• The time to payout, i.e., the time required for the cumulative undiscounted net revenue to
equal the amount of the initial capital investment (some entities prefer to use discounted
payout)
• Net profit index (dollars generated per dollar initially invested, also known as “return on
investment”), calculated as the cumulative undiscounted NCF, i.e., cumulative
undiscounted net revenue, excluding CAPEX, divided by the initial undiscounted
CAPEX
• Discounted return on investment (NPV of cash flow, i.e., cumulative DCF, including
CAPEX, divided by discounted investment; or the present value ratio, PVR)
• Magnitude of NPV
• Discounted cash-flow rate of return (DCFROR, the discount rate at which the cumulative
NPV equals zero; also known as internal rate of return IRR)
Each of the criteria has advantages and drawbacks, so investment decisions usually rely on a
combination of these economic factors. For example, the PVR method is beneficial when
considering the time value of money and for choosing investment opportunities when investment
capital is constrained, but the magnitude of the calculation is meaningful only in comparison with
231
other competing projects. Similarly, although time to payout is commonly used, it fails to take into
account the time value of money or how profitable the investment is after the payout. Furthermore,
IRR is occasionally literally incalculable (for example, when all undiscounted future cash flows
are positive), and these calculations can also result in multiple IRRs.
9.6.2 Sunk Costs. Because Reserves represent quantities yet to be recovered, their NCF evaluation
is always forward-looking for the purposes of applying the PRMS principles. Sunk costs, which
are expenditures incurred prior to the effective date of the evaluation, are relevant for accounting
or tax purposes, in addition to production sharing contracts and overall project economic
determination. Some projects that, upon cash-flow-based evaluation, are sub-economic and fail to
meet the requirements to be classified as Reserves may, in subsequent years, achieve Reserves
status as capital expenditures are incurred and become sunk costs no longer impacting the future
NCF. This situation is discussed in PRMS § 3.1.2.8, where reference is made to uneconomic 1P
quantities: “As costs are incurred in future years (i.e., become sunk costs) and development
proceeds, the low estimate may eventually become economic and be reported as Proved Reserves.”
It should also be recognized that sunk costs can have relevance beyond accounting purposes.
Sunk costs can influence future cash flows via their impact on, for instance, future taxable income.
For example, a fixed asset purchased before the evaluation date can generate tax depreciation
charges and/or tax loss carry-forwards that impact income tax charges in future periods.
Some fiscal regimes such as PSCs use mechanisms based in part on sunk costs that are forecast
to be recovered as cost oil, which will influence the evaluated party’s entitlement share of project
revenue. For example, shares of revenue (“profit-oil”) might be based on an “R-Factor,” which is
often calculated as some measure of cumulative project inflows to date divided by some measure
of cumulative outflows to date, including sunk costs.
The use of sunk costs will also be important when entitlement interests are calculated on a
before- or after-payout basis.
Therefore, when sunk costs influence future cash flows used in an evaluation, this influence
must be reflected in the calculation.
9.7 “Economic” Compared to “Commercial”

According to the PRMS (Appendix A—Glossary), “A project is commercial when there is
evidence of a firm intention to proceed with development within a reasonable time-frame.
Typically, this requires that the best estimate case meet or exceed the minimum evaluation decision
criteria (e.g., rate of return, investment payout time). There must be a reasonable expectation that
all required internal and external approvals will be forthcoming. Also, there must be evidence of a
technically mature, feasible development plan and the essential social, environmental, economic,
political, legal, regulatory, decision criteria, and contractual conditions are met.”
There has been some confusion in the industry over the use of and difference between the terms
“economic” and “commercial.” As can be seen in the previous paragraph, under the PRMS, a
project is commercial when there is evidence that certain conditions are met, one of which is the
economic condition. Consequently, being “economic” is a prerequisite for a project to be
“commercial.” Commercial projects are, therefore, economic; however, an economic project is not
necessarily commercial if it fails to pass the other necessary criteria, including providing a positive
NPV at the entity’s hurdle discount rate.
232
9.8 Example
An exploratory well, ALPHA-1, results in a discovery in an area with promising seismic signature
updip from an old well that was wet in the subject reservoir. Logs from the two wells correlate,
and no evidence of faulting can be inferred, supporting the seismic interpretation and suggesting a
structural trap. With updip hydrocarbons discovered, the entity engineers and geoscientists use the
pressure data from the new well plus the old wet well to estimate an oil-water contact and generate
hydrocarbon pore volume maps for low, best, and high case scenarios. Collected fluid samples
enable the laboratory analysis of the oil formation volume factor for the proposed surface
separation process, permitting the calculation of the oil originally in place (OOIP). The reservoir
has booked reserves in nearby fields where it is being produced through pressure maintenance by
water injection.
Data are gathered to perform an economic assessment through cash-flow analysis. Based on
the well tests and analogy to the nearby fields producing from the reservoir (but in different
accumulations), a “most likely” performance projection is created. The proposed development plan
for the analysis assumes the re-entry of the discovery well and the drilling of two more producers,
three water-injection wells, and a water-supply well. There are production facilities in the vicinity
to which the new project can be connected, provided the projected water handling from Project
ALPHA can be accommodated. A forecast case is run using a midyear escalation approach.
Arrangements with the host government call for a concession term of 40 years with a bonus to be
paid to the government in the tenth year of production. The government will impose production
taxes as a percent of revenue and will retain a royalty interest. Table 9.2 summarizes the economic
analysis input assumptions, including data pertinent to both the economic and subsequent
commercial analyses. Costs are shown before inflation.
INPUT ASSUMPTIONS
Oil Price, $/bbl 50 Government Royalty 12.5% CapEx, Wells, $MM 24

Gas Price, $/MMBTU 2.50 Government Bonus, $MM 20 CapEx, Facilities, $MM 100
BTU Content, BTU/SCF 1300 Production Tax, % Revenue 10% OpEx, $M/month* 112.5
Gas Price, $/MSCF 3.25 ADR Liability, $MM 20
Oil Price Cap, $/bbl 75 Income Tax Rate, % 10% Terminal Decline Rate, %/yr 8%
Gas Price Cap, $/MSCF 5 Discount Rate, % 10% • Pressure Maintenance by Water Injection
Currency Inflation Rate, %/yr 2% Concession Term, yrs 40 • Government Bonus Non-Tax Deductible
Mid-Year Escalation ADR Incurred in Year EL + 1 * Three producers, three injectors
Table 9.2—NCF input assumptions, where OpEx is operating expenses.
Fig. 9.2 depicts the EL test graphically. The annual bars represent the undiscounted NCF, i.e.,
the difference between net revenue and net expenditures. The cumulative NCF is shown by a red
line, which dips in year 10 as a result of the bonus, but which continues to increase thereafter.
Cumulative oil recovery is depicted by the black line. The maximum cumulative NCF occurs in
year 27 (USD 103 million); beyond that date, costs exceed revenues. As a result, the economic life
of the project is 27 years, which is less than the concession term by 13 years. (Further, no technical
constraints, such as lift capacity or water handling, occur in the economic assessment.) Subtracting
the ADR, which, when adjusted for the currency inflation rate, is USD 34 million in year 28 (i.e.,
the year after the EL), from the maximum cumulative undiscounted NCF yields a positive number.
This establishes economic viability, and the potential oil reserves equal the cumulative oil recovery
at the EL, which is equal to 5.2 million STB.
233
Table 9.3 depicts the NCF as generated for the purpose of establishing the EL. A most likely
production forecast in thousands of stock tank barrels (MSTB) is shown on an annual basis in
Column B, while the associated gas production is shown in Column C in millions of standard
cubic feet (MMSCF). The values shown are to the gross (100%) interest, i.e., a working interest
of 100%. However, under the concession agreement, there is a 12.5% royalty owed to the
government, so these production figures are thereby reduced as reflected in Columns D and E to
obtain the entity’s net (entitled) oil and gas produced, respectively. Columns F and G show the
oil and gas prices in US dollars per barrel with a 2%/year inflation index applied (using a midyear
convention) to yield the inflation-adjusted prices in nominal US dollars. Initial prices are
USD 50/bbl and USD 3.25/MSCF for oil and gas, respectively. and are thus escalated until
reaching caps of USD 75/bbl and USD 5.00/MSCF, respectively. Annual revenue from net oil
and gas sales in nominal dollars is calculated in Column H as the product of Columns D and F
plus the product of Columns E and G.
Fig. 9.2—NCF graph for determination of the EL.
In terms of costs and cash outflow, Column I contains the estimated operating expenses
(OPEX), Column J represents production taxes (in this case, as a percent of net revenue, although
it could be a flat-rate levy or tax, but this does not include income tax), while Column K continues
with the capital expenditures (CAPEX) but not yet reflecting the ADR and its spend timing, again
with all columns in nominal US dollars. Column L highlights the required bonus payment in year
10 of the project.
The undiscounted NCF of Column M is the difference between Column H (Total Net
Revenue) and the summation of Columns I through L (Total Net Costs). Due to the lead time in
establishing production from the field, the first 3 years of NCFs are all negative. Once production
commences, annual NCF figures (Column M) become positive, with the exception of year 10,
when the bonus is paid. Column N is the cumulative undiscounted NCF (for EL test purposes),
reflecting the running totals from Column M. It is not until the seventh year that the cumulative
NCF becomes positive (note also Fig. 9.2).
234
A B C D E F G H I J K L M N O
Gross (100%) Net (Entitled) Net Net Prod'n Capital Net Cum. Net Economic
Prices (Nominal) Bonus
Production Production Revenue OPEX Taxes Costs Cash Flow Cash Flow Limit
Oil Gas Oil Gas Oil Gas Nominal Nominal Nominal Nominal Nominal Nominal Nominal Test
Year
MSTB MMSCF MSTB MMSCF $/BBL $/MCF $MM $MM $MM $MM $MM $MM $MM
1 50.498 3.282 - - - 10.100 (10.100) (10.100) -
2 51.507 3.348 - - - 56.658 (56.658) (66.758) -
3 52.538 3.415 - - - 52.538 (52.538) (119.295) -
4 1020.179 612.108 892.657 535.594 53.588 3.483 49.702 0.965 4.970 43.767 (75.528) -
5 808.088 484.853 707.077 424.246 54.660 3.553 40.156 0.984 4.016 9.839 25.318 (50.211) -
6 649.932 389.959 568.690 341.214 55.753 3.624 32.943 1.505 3.294 28.143 (22.067) -
7 529.771 317.863 463.550 278.130 56.868 3.696 27.389 1.535 2.739 23.115 1.048 -
8 436.965 262.179 382.344 229.406 58.006 3.770 23.043 1.566 2.304 19.173 20.220 -
9 364.233 218.540 318.704 191.222 59.166 3.846 19.592 1.597 1.959 16.035 36.256 -
10 306.487 183.892 268.176 160.905 60.349 3.923 16.815 1.629 1.682 20.000 (6.496) 29.760 -
11 260.099 156.060 227.587 136.552 61.556 4.001 14.556 1.662 1.456 11.438 41.198 -
12 222.441 133.464 194.636 116.781 62.787 4.081 12.697 1.695 1.270 9.732 50.930 -
13 191.574 114.944 167.627 100.576 64.043 4.163 11.154 1.729 1.115 8.309 59.240 -
14 166.052 99.631 145.295 87.177 65.324 4.246 9.861 1.764 0.986 7.112 66.351 -
15 144.779 86.868 126.682 76.009 75.000 5.000 9.881 1.799 0.988 7.094 73.445 -
16 126.918 76.151 111.053 66.632 75.000 5.000 8.662 1.835 0.866 5.961 79.406 -
17 111.818 67.091 97.841 58.705 75.000 5.000 7.632 1.872 0.763 4.997 84.403 -
18 98.974 59.384 86.602 51.961 75.000 5.000 6.755 1.909 0.675 4.170 88.573 -
19 87.983 52.790 76.985 46.191 75.000 5.000 6.005 1.947 0.600 3.457 92.030 -
20 78.527 47.116 68.711 41.227 75.000 5.000 5.359 1.986 0.536 2.837 94.868 -
21 70.351 42.211 61.557 36.934 75.000 5.000 4.801 2.026 0.480 2.295 97.163 -
22 63.248 37.949 55.342 33.205 75.000 5.000 4.317 2.067 0.432 1.819 98.982 -
23 57.050 34.230 49.919 29.951 75.000 5.000 3.894 2.108 0.389 1.396 100.378 -
24 51.619 30.972 45.167 27.100 75.000 5.000 3.523 2.150 0.352 1.021 101.399 -
25 46.842 28.105 40.986 24.592 75.000 5.000 3.197 2.193 0.320 0.684 102.083 -
26 42.623 25.574 37.295 22.377 75.000 5.000 2.909 2.237 0.291 0.381 102.464 -
27 38.886 23.331 34.025 20.415 75.000 5.000 2.654 2.282 0.265 0.107 102.571 27.000
28 35.563 21.338 31.118 18.671 75.000 5.000 2.427 2.327 0.243 (0.143) 102.429 -
29 32.600 19.560 28.525 17.115 75.000 5.000 2.225 2.374 0.222 (0.371) 102.057 -
30 29.951 17.970 26.207 15.724 75.000 5.000 2.044 2.421 0.204 (0.582) 101.476 -
SUM 6,073.55 3,644.13 5,314.36 3,188.62 334.19 50.16 33.42 129.13 20.00 101.48
Table 9.3—NCF for determination of the EL.
Although the evaluator may simply scan the economic output and/or Fig. 9.2 to determine the
time at which the maximum undiscounted NCF is achieved, for the purpose of this example, we
have also included Column O in the table specifically to illustrate the determination of the
maximum NCF and economic life. Column O is a simple IF statement comparing the
corresponding value of Column N to the MAX of the range of values in Column N, and, when
they are equal, the IF statement returns the value of the year (from Column A); otherwise, the cell
is left blank. This column identifies year 27 as the year in which we reach the maximum cumulative
undiscounted NCF.
Finally, a comparison of the cumulative NCF (USD 102.6 million) to the estimated ADR
liability (USD 20 million adjusted for inflation to the EL in year 28 is USD 34 million) satisfies
the economic viability criterion required to consider the undeveloped project’s quantities as
Reserves (PRMS § 3.1.2.5). The next step is to assess the project’s commerciality. The following
text describes the NPV calculation used by this entity for investment decision.
We will consider the impact of ADR and income taxes for this example, assuming that the
ADR costs are incurred in the year following cessation of production at the EL.
Columns P and Q in Table 9.4 reproduce Columns D and E in Table 9.3 (net oil and net gas
produced, respectively) but consider only the economic quantities using the economic flag
(Column O). Columns R and S are the total net revenue from Column H and the total net costs
from the sum of Columns I through L, again considering only the economic quantities. The
undiscounted NCF for the economic life is calculated in Column T, and the ADR liability is
incurred in Column U by comparing the maximum NCF year (Column O) to the production year
(Column A), and, when equal, the ADR costs are imposed in the following year (as per the initial
assumptions).
Columns V through AA are concerned with tax considerations.
235
A P Q R S T U V W X Y Z AA AB AC
Income Tax Considerations
Economic Economic Total Total Total Losses Losses to
ADR Newly Taxable Income Cum. Undisc. Cum. Disc. Net
Net Oil Net Gas Net Net Net Carried Total Carry
Liability Incurred Income Tax Net Cash Flow Cash Flow
Produced Produced Revenue Costs Cash Flow Forward Forward
Year
MSTB MMSCF Nom. $MM Nom. $MM Nom. $MM Nom. $MM Nom. $MM Nom. $MM Nom. $MM Nom. $MM Nom. $MM Nom. $MM Nom. $MM Nom. $MM
1 - - - 10.100 (10.100) - 10.100 - 10.100 (10.100) - (10.100) (10.100) (9.630)
2 - - - 56.658 (56.658) - 56.658 10.100 66.758 (66.758) - (66.758) (66.758) (58.740)
3 - - - 52.538 (52.538) - 52.538 66.758 119.295 (119.295) - (119.295) (119.295) (100.139)
4 892.657 535.594 49.702 5.935 43.767 - 5.935 119.295 125.230 (75.528) - (75.528) (75.528) (68.786)
5 707.077 424.246 40.156 14.838 25.318 - 14.838 75.528 90.367 (50.211) - (50.211) (50.211) (52.299)
6 568.690 341.214 32.943 4.800 28.143 - 4.800 50.211 55.010 (22.067) - (22.067) (22.067) (35.637)
7 463.550 278.130 27.389 4.274 23.115 - 4.274 22.067 26.342 1.048 0.105 - 0.943 (23.253)
8 382.344 229.406 23.043 3.870 19.173 - 3.870 - 3.870 19.173 1.917 - 18.198 (14.810)
9 318.704 191.222 19.592 3.557 16.035 - 3.557 - 3.557 16.035 1.604 - 32.630 (8.391)
10 268.176 160.905 16.815 23.311 (6.496) - 3.311 - 3.311 13.504 1.350 - 24.784 (11.564)
11 227.587 136.552 14.556 3.118 11.438 - 3.118 - 3.118 11.438 1.144 - 35.078 (7.780)
12 194.636 116.781 12.697 2.965 9.732 - 2.965 - 2.965 9.732 0.973 - 43.837 (4.852)
13 167.627 100.576 11.154 2.845 8.309 - 2.845 - 2.845 8.309 0.831 - 51.316 (2.580)
14 145.295 87.177 9.861 2.750 7.112 - 2.750 - 2.750 7.112 0.711 - 57.716 (0.813)
15 126.682 76.009 9.881 2.787 7.094 - 2.787 - 2.787 7.094 0.709 - 64.101 0.790
16 111.053 66.632 8.662 2.701 5.961 - 2.701 - 2.701 5.961 0.596 - 69.466 2.015
17 97.841 58.705 7.632 2.635 4.997 - 2.635 - 2.635 4.997 0.500 - 73.963 2.948
18 86.602 51.961 6.755 2.585 4.170 - 2.585 - 2.585 4.170 0.417 - 77.716 3.656
19 76.985 46.191 6.005 2.548 3.457 - 2.548 - 2.548 3.457 0.346 - 80.827 4.190
20 68.711 41.227 5.359 2.522 2.837 - 2.522 - 2.522 2.837 0.284 - 83.381 4.588
21 61.557 36.934 4.801 2.506 2.295 - 2.506 - 2.506 2.295 0.230 - 85.447 4.880
22 55.342 33.205 4.317 2.498 1.819 - 2.498 - 2.498 1.819 0.182 - 87.083 5.091
23 49.919 29.951 3.894 2.497 1.396 - 2.497 - 2.497 1.396 0.140 - 88.340 5.239
24 45.167 27.100 3.523 2.502 1.021 - 2.502 - 2.502 1.021 0.102 - 89.259 5.336
25 40.986 24.592 3.197 2.513 0.684 - 2.513 - 2.513 0.684 0.068 - 89.875 5.396
26 37.295 22.377 2.909 2.528 0.381 - 2.528 - 2.528 0.381 0.038 - 90.218 5.426
27 34.025 20.415 2.654 2.547 0.107 - 2.547 - 2.547 0.107 0.011 - 90.314 5.434
28 - - - - - 34.477 34.477 - 34.477 (34.477) - (34.477) 55.837 2.926
29 - - - - - - - - - - - - 55.837 2.926
30 - - - - - - - - - - - - 55.837 2.926
SUM 5,228.51 3,137.11 327.50 224.93 102.57 34.48
Table 9.4—DCF for assessment of commerciality.
Column AB presents the cumulative undiscounted NCF including the effects of income taxes
and ADR liability. Using the present value formula presented earlier in this chapter, assuming the
midyear discounting option, the entity’s annual 10% discounting factor is applied to the yearly
NCFs, resulting in annual discounted NCFs (DCFs). These annual DCFs are summed for a running
cumulative value, as shown in Column AC. The total cumulative DCF of USD 2.9 million is
equivalent to the post-tax net present value (NPV), which may be used for investment decisions
or portfolio ranking purposes. Because the NPV is greater than zero, the entity’s required rate of
return has been exceeded.
Finally, the bottom row shows the summation of the potential reserves net to the entity
(excluding the royalty interest) and the cumulative undiscounted NCF (before income taxes and
ADR) in addition to the cumulative DCF after taxes and ADR. A summary of the output analysis
is shown in Table 9.5.
OUTPUT RESULTS
Economic Life, Years 27 Net Oil Reserves, MSTB 5,229

Is Cumulative Undiscounted NCF > ADR? Y Net Gas Reserves, MMSCF 3,137
Is Cumulative Discounted NCF > 0? Y Cumulative Discounted NCF, $MM (w/ADR) 2.926
Cumulative Discounted NCF, $MM (pre-ADR) 5.434
Table 9.5—Cash flow evaluation results.
236
It is necessary to point out that this cash-flow evaluation is only part of the commerciality
assessment. Unless all of the requirements spelled out in PRMS § 2.1.2.1 are satisfied, including
“legal, contractual, environmental, regulatory, and government approvals are in place or will be
forthcoming, together with resolving any social and economic concerns,” the economically
producible quantities must stay in the Contingent Resources class.
9.9 Probabilistic Evaluation

Evaluation may, of course, be conducted using either deterministic or probabilistic methods.
Although the generation of cash-flow evaluations is straightforward, the accuracy of the estimates
is dependent on the property-specific input data as well as the expertise of, and effective
collaboration among, the multidisciplinary evaluation team members.
Each component of the project NCF terms (such as production rate, product price, CAPEX,
OPEX inflation rate, taxes, and interest rate) has some element of uncertainty. There may also be
some value of information (VoI) component or cost-benefit assessment, such as whether to drill
an appraisal well, acquire additional seismic data, or implement pressure-maintenance operations
at the outset of production. There are several ways to perform cash-flow analysis under uncertainty
conditions, and the reader is directed to Chapter 7—Probabilistic Resources Estimation herein,
such as for the expected monetary value (EMV) example shown in Section 7.4.1.
9.10 Environmental, Social, and Governance Issues

“Environmental, social, and governance (ESG)” is a general term of reference that encompasses a
range of factors, some of which must be integrated into a project’s investment and decision-making
process. The term has seen increased usage among stakeholders and the investing community, but
many of the elements associated with ESG are not new to PRMS.
The term ESG is not explicitly mentioned within the PRMS, but PRMS §1.2.0.10 clearly
requires consideration of certain components of ESG pertaining to a project to be able
recognize Reserves:
“The commercial viability of a development project within a field’s development plan is
dependent on a forecast of the conditions that will exist during the time period encompassed by
the project (see Section 3.1, Assessment of Commerciality). Conditions include technical,
economic (e.g., hurdle rates, commodity prices), operating and capital costs, marketing, sales
route(s), and legal, environmental, social, and governmental factors forecast to exist and impact
the project during the time period being evaluated” (emphasis added).
Other key PRMS references are in §2.1.2.1, where the commerciality criteria are provided,
and in the Glossary definition of Defined Conditions used for an evaluation.
“Governmental factors” and “governance” are not interchangeable expressions. Governance
typically refers to the manner in which entities conduct their corporate business, which may in turn
affect a project, e.g., commitment to fund and execute within a reasonable time frame.
The PRMS advocates conveying the project uncertainties and risks examined as part of the
commerciality assessment process, and this includes ESG factors as well.
9.10.1 ESG Within the PRMS Context. The PRMS is a project-based system, and the concept
of a project includes all of the factors that may potentially impact the recovery of hydrocarbon
resources from that project. The PRMS requires that Reserves for a project be commercial, and
commerciality includes more than technical maturity and economic performance of the project.
Historically, ESG emphasis has been placed on compliance and demonstrating “license to operate”
237
in a community. It is beyond the scope of this version of the Guidelines for Application of the
PRMS to provide detailed guidance on how to evaluate the wide spectrum of ESG factors that
relate to resources assessments in various global geographic areas. ESG requirements vary locally
and regionally, and it is incumbent upon the project owners to be informed of those requirements
and their potential impacts, if any, on their specific project(s).
The effects of ESG considerations may influence resources classification, categorization, and
maturity sub-classes, such as in Chapter 2—Petroleum Resources Definitions, Classification, and
Categorization Guidelines herein, and potentially other topics (probabilistic evaluation, aggregation,
and possibly entitlement) as projects are evaluated under different ESG environments and varying
contractual regimes.
9.10.2 ESG Impact on Economic Evaluation. As they pertain to this chapter, ESG factors may
result in additional costs, revenues, and/or taxes. For example, there are potential active carbon
emission credit trading structures that have been regulated by some jurisdictions. While entities
generally must invest capital to meet carbon emission reduction goals, the carbon emission credits
create the opportunity for another source of revenue for some projects. To the extent that such
revenue is attributable to the oil and gas operations of a specific project (as opposed to a corporate-
level transaction), the PRMS allows that revenue source (and any associated capital investments
and/or additional operating expenses) to be a part of the project’s commerciality assessment.
In some cases, ESG factors may result in additional taxes. These taxes may take the form of
production taxes, where the tax is directly related to production or emissions of the project, or they
may impact the entity’s corporate income tax. Most project evaluations are done on a “before
corporate income tax” basis. Income taxes are generally applied at the corporate level, and it can
be difficult to quantify the true economics of a project on an “after corporate income tax” basis.
However, there are some situations, such as PSCs, where petroleum evaluators must consider the
“after corporate income tax” aspects of a project in order to calculate the entitlement share of
production. In these cases, the corporate income tax associated with a project is considered to be
an event that occurs inside the “ring-fence” (or contract area) of the project. By the same token,
any specific ESG cost or revenue that can be determined to be inside the “ring-fence” of a project
should be a part of the commercial evaluation of that project.
In summary, ESG factors cover a wide spectrum of topics, and, as such, they are incorporated
in an entity’s investment and decision-making process. The PRMS has always required ESG
factors to be reviewed as key input in evaluating a project’s commerciality.
9.11 Conclusion
The estimation of resources and reserves is subject to uncertainty, not only due to inherent
uncertainties in the petroleum initially in place and the efficiency of the recovery program, but also
due to the associated cash-flow assumptions that affect the future net revenue (and NPV).
Documentation of the product prices, the capital and operating costs, and the timing of
implementation of projects is a fundamental step in the quantification of marketable quantities that
result in resources and reserves estimates and their associated classification as Resources or Reserves.
These factors are forecast for the project over time, and evaluators must clearly identify and
document the assumptions used in their evaluation because these assumptions directly affect the
classification as Reserves or Resources. This chapter has sought to spell out the necessary
requirements for generating cash-flow-based evaluations that are compliant with PRMS guidelines.
238
The author wishes to acknowledge the contribution of Yasin Senturk, author of this chapter in the
previous version of the Guidelines for Application of the PRMS, and the insights of Ken Kasriel,
Steve Gardner, Rawdon Seager, Rod Sidle, Dan DiLuzio, Dan Olds, John Lee, Steve McCants,
Tim Smith, Dave Kemshell, and Monica Clapauch Motta.
9.13 References
Clark, F. D. and Lorenzoni, A. B. 1978. Applied Cost Engineering. New York City, New York,
USA: Marcel Dekker.
Humphreys, K. K. and Katell, S. 1981. Basic Cost Engineering. New York City, New York,
USA: Marcel Dekker.
Society of Petroleum Evaluation Engineers. 2002. SPEE Recommended Evaluation Practice
#5—Discounting Cash Flows. Houston, Texas, USA: Society of Petroleum Evaluation
Engineers. https://spee.org/recommended-evaluation-practices/.
9.14 Bibliography
Arnold, K. and Stewart, M. 1991. Surface Production Operations, Design of Oil-Handling
Systems and Facilities, Vol. 1. Houston, Texas, USA: Gulf Publishing Company.
Arnold, K. and Stewart, M. 1989. Surface Production Operations, Design of Gas-Handling
Systems and Facilities, Vol. 2. Houston, Texas, USA: Gulf Publishing Company.
Campbell, J. M. and Campbell, R. A. 2001. Analyzing and Managing Risky Investments.
Norman, Oklahoma, USA: John M. Campbell.
Higgins, R. C. 2001. Analysis for Financial Management. New York City, New York, USA:
Irwin McGraw-Hill.
Newendorp, P. D. and Schuyler, J. R. 2000. Decision Analysis for Petroleum Exploration,
second edition. Aurora, Colorado, USA: Planning Press.
Schuyler, J. R. 2004. Decision Analysis Collection. Aurora, Colorado, USA: Planning Press.
Seba, R. D. 1998. Economics of Worldwide Petroleum Production. Tulsa, Oklahoma, USA:
OGCI Publications.
Society of Petroleum Evaluation Engineers. 2002. SPEE Recommended Evaluation Practice
#7—Escalation of Prices and Costs. Houston, Texas, USA: Society of Petroleum Evaluation
Engineers. https://spee.org/recommended-evaluation-practices/.
Society of Petroleum Evaluation Engineers. 2007. Canadian Oil and Gas Evaluation Handbook
(COGEH), Vol. 1. Calgary, Alberta, Canada: Society of Petroleum Evaluation Engineers.
Also available (with subscription) as: Canadian Oil and Gas Evaluation Handbook
(COGEH), consolidated third edition (online). Calgary, Alberta, Canada: Society of
Petroleum Evaluation Engineers. https://speecanada.org/coge-handbook-view/.
239
Chapter 10
Unconventional Resources
Estimation
Dilhan Ilk
10.1 Introduction
Unconventional resources, primarily tight gas/tight oil and shale gas/oil, have dominated oil and
gas production in North America since the early 2000s, transforming the industry and triggering
major changes to worldwide oil prices ever since. Although production from unconventional
reservoirs is common in North America, there are many underlying questions yet to be addressed.
Industry understanding of fluid flow concepts, storage mechanisms, and the extent of hydrocarbon
drainage areas remains quite uncertain, complicating the estimation of reserves and resources.
Table 10.1 presents a summary of the main differences between conventional and
unconventional reservoirs (modified from Berg 2013). It is important to mention that some of these
points may not be applicable to all unconventional (or low-permeability) reservoirs, but in a
broader sense, the list is a good indicator of characteristic differences that would ultimately lead
to modification of evaluation and assessment methodologies.
Table 10.1—Main differences between conventional and unconventional reservoirs (modified from Berg 2013).
As mentioned earlier, production from unconventional reservoirs has become common in

North America and is emerging worldwide (e.g., Argentina, Australia, China, and Middle East).
In addition, very large oil and gas volumes do exist in resources such as oil shale, extra heavy oil,
240
and bitumen, although their commercial recovery would often require new or improved technology
and higher product prices. For some of these resources, such as oil shale and gas hydrates, no
commercial recovery methods have yet been developed to produce these resources, although some
pilot projects have demonstrated methods to extract in-place volumes.
10.1.1 Assessment and Classification Issues. The Petroleum Resources Management System
(PRMS 2018) resources definitions and classification framework are intended to be appropriate
for all types of petroleum accumulations regardless of reservoir type, completion and production
technology applied, or processing requirements. However, differences in concepts, methodologies,
and techniques are more pronounced in the evaluation and assessment of unconventional reservoirs
due to their unique characteristics.
Once an accumulation has been discovered, the appraisal phase assessment begins by
evaluating core, log, and seismic data to evaluate in-place volumes and the extent of the potentially
productive area. However, undiscovered hydrocarbon volumes, classified as Prospective
Resources, must rely on the application of analogs and/or modeling techniques. It is important to
note that there is significant uncertainty at this phase of evaluation, and, as such, estimations of
recoverable resource quantities must include an estimate of the associated uncertainty, which
would be allocated to the PRMS categories by using probabilistic, deterministic, or both methods.
A significant difference between conventional and unconventional reservoirs is the evaluation
of recovery efficiency. In addition to the importance of reservoir quality, for unconventional
reservoirs, completion effectiveness plays a more important role in determining the recovery factor
compared to conventional reservoirs. In general, recovery must be considered as a function of the
development plan combined with a specific completion design/technology. However, the nature
of the unconventional reservoir and the limited drainage area of the wells within it necessitate
planning for more vertical wells, or longer horizontal laterals, relative to conventional reservoirs.
Fig. 10.1 illustrates conceptual recovery efficiencies and net present value profiles based on
specific development scenarios.
Once a discovery is made, and its development scenario is specified, recoverable quantities
can be estimated. The most convincing evidence to establish discovery is to have a production test
with flow of hydrocarbons to the surface from the reservoir of interest. (Establishing a discovery
is discussed in PRMS § 2.1.1.1.) In the absence of a production test, the PRMS allows for discovery
based on other sufficient evidence (e.g., core, log, sampling, nearby analogs). If analogy is used,
analog data and rationale must be clearly documented and compared with the data from the subject
reservoir. In the aggregate, the properties of the subject reservoir should be similar or better than
the analogous reservoir.
As indicated in the PRMS, the extent of the discovery within a pervasive accumulation is
based on the evaluator’s reasonable confidence based on distances from existing experience;
otherwise, quantities remain as undiscovered (PRMS § 2.4.0.4). Extrapolation of reservoir
presence or productivity beyond the immediate vicinity of a control point (or discovery well)
should be limited unless there are clear engineering and geoscience data to validate areas around
discovery well(s) or control point(s) (PRMS § 2.4.0.3). Contingent Resources are then categorized
in relation to the range of uncertainty associated with estimates of recoverable quantities. Estimates
of recoverable quantities must consider the effects of the proposed development plan, reservoir
properties, fluid behavior, and completion design on production profiles. Often, this impact may
not be easily characterized due to early stage of development and lack of data and well control.
241
Therefore, it is critical to integrate uncertainty analysis and appropriate analog data into the
evaluation to address the range of uncertainty and associated categories.
Fig. 10.1—Recovery factor (R.F.) and net present value (NPV) profiles based on various development scenarios (colored
dots) and product prices (colored lines).
The PRMS recommends focusing on gathering data and performing analyses to clarify and
mitigate key contingencies that prevent commercial development. Assigning project maturity sub-
classes is recommended. Reserves must satisfy four criteria: discovered, recoverable, commercial,
and remaining (as of the evaluation’s effective date) based on the development project(s) applied
(PRMS Table 1). The same criteria apply for both conventional and unconventional reservoirs.
Resources/reserves categorization should reflect the range of uncertainty in the estimates of
recoverable quantities associated with the development project with a single set of defined conditions
applied. Use of different commercial assumptions across categories within a resources class is
referred to as “split conditions” and is not allowed under the PRMS guidelines. Development
projects in unconventional reservoirs may include large numbers of drilling and completion of wells
to fully appraise a field or play. However, due to various commercial factors, certain portions of a
development may fall outside a reasonable development time frame, and so those specific portions
can be associated with other development projects (PRMS § 2.1.2.3). In this situation, because the
recovery is beyond the reasonable time frame, it is considered sub-commercial and reclassified as
Contingent Resources (PRMS Appendix A—Glossary, “Sub-Commercial”).
Typical evaluation methodologies for unconventional reservoirs involve integration of data
from various disciplines. It must be remembered that recovery from unconventional reservoirs is
a combined function of reservoir, completions, and development plan (e.g., well spacing).
Advanced methods are commonly utilized to evaluate well performance and make decisions on
future development options. Fig. 10.2 presents a conceptual view of the performance-based life
242
cycle for an unconventional reservoir. This figure demonstrates that reservoir, geoscience,
completions, and production data are all integrated and evaluated continuously to update and
execute specific actions related to completion design and well spacing in an unconventional play.
Fig. 10.2—Conceptual view of performance-based life cycle for an unconventional reservoir. This schematic illustrates an
evaluation of an unconventional reservoir using “performance-based” methodologies relying on production data while
integrating geoscience, reservoir, fluid, and completions data.
The following sections (Sections 10.2 through 10.6) have been prepared by several different
authors to provide an overview of each resource type and information on evaluation approaches. In
addition to updating the sections on tight/shale gas and oil and coalbed methane from the prior (2011)
version of the Guidelines for Application of the PRMS, a new section on Evaluation Methodologies
has been included. Sections 10.2 through 10.6 address the following subjects:
• 10.2 Tight Gas and Oil
• 10.3 Shale Gas and Oil
• 10.4 Evaluation Methodologies for Tight/Shale Gas and Oil
• 10.5 Coalbed Methane
• 10.6 Other Unconventional Oil
Roberto Aguilera
10.2 Tight Gas and Oil Formations

10.2.1 Introduction. This section in the prior (2011) version of the Guidelines for Application of
the PRMS discussed only tight gas formations (TGFs), but now, tight oil formations (TOFs) are
also included. The word “petroleum” as used in this section includes oil, gas, and natural gas
liquids. The word “tight” as used herein includes sandstones, carbonates, and shales. Tight and
shale reservoirs are distinguished as follows: In tight petroleum formations (TPFs), petroleum has
243
migrated from a source rock to the TPF. In a shale reservoir, petroleum generated in the shale
remains in the shale (Aguilera 2014). This sub-section does not describe shales because they are
covered in more detail below in Section 10.3.
The US Natural Gas Policy Act of 1978 (Tight Formation Gas, 1978) defined a TGF as having
in-situ reservoir permeability equal to or less than 0.1 md (Kazemi 1982; Aguilera and Harding
2007). For purposes of this section, the definition is expanded such that a TGF includes “a reservoir
that cannot be produced at economic flow rates nor recover economic volumes of natural gas
unless the well is stimulated by a large hydraulic fracture treatment or produced by use of a
horizontal wellbore or multilateral wellbores” (Holditch 2006, p. 86). The same expanded
definition can be used for the case of TOFs, but there is not a specific permeability cutoff for TOFs.
Industry experience, however, suggests the permeability cutoff for TOFs is generally larger than
that for TGFs (for example, up to 1 md) with the exception of shales, which have permeabilities
that are much smaller and can reach the level of nanodarcies.
The industry generally divides TGFs into (1) basin-centered gas accumulations (BCGAs), also
known as continuous-type gas accumulations (Law 2002; Schmoker 2005), and (2) gas reservoirs
that occur in low-permeability, poor-quality reservoir rocks in conventional structural and
stratigraphic traps (Shanley et al. 2004). The PRMS (Appendix A—Glossary) defines a
continuous-type deposit as:
“A petroleum accumulation that is pervasive throughout a large area and that
generally lacks a well-defined OWC or GWC [oil- or gas-water contact]. Such
accumulations are included in unconventional resources. Examples of such deposits
include ‘basin-centered’ gas, tight gas, tight oil, gas hydrates, natural bitumen, and
oil shale (kerogen) accumulations.”
The same division can be used in the case of TOFs, in which situation, they may be more
appropriately termed basin-centered petroleum accumulations (BCPA). This is illustrated in
Fig. 10.3, which shows both TGFs and TOFs below conventional oil and gas formations in a BCPA.
Fig. 10.3—TGFs (shown as tight sand gas) and TOFs (shown as tight sand oil) below conventional oil and gas
reservoirs in a BCPA (National Energy Board of Canada 2011).
244
10.2.2 Reservoir and Hydrocarbon Characteristics. The primary definition used in this chapter
assumes that TGFs, including sandstones and carbonates, are characterized by permeabilities of
less than 0.1 mD. The hydrocarbons in these rocks are primarily methane with some impurities,
but there are also occurrences of associated gas condensate. TOFs, including sandstones and
carbonates, are characterized by low permeabilities of up to 1 mD. The hydrocarbons in these
rocks are primarily light oil.
Fig. 10.4 shows an example of flow units in an oil formation that includes a TOF. The graph
presents data from the Cardium Formation of the Pembina Field (Canada) published by MacKenzie
(1975). The crossplot shows porosity vs. permeability, where rp35 is the pore throat radii in microns
at 35% cumulative pore volume (Kolodzie 1980). Type I and Type II rocks in Fig. 10.4 correspond
to a conventional reservoir. Type III rocks represent a tight oil halo around the conventional
reservoir in the Pembina Field. Historically, oil production from this low-permeability halo was
not possible. In fact, MacKenzie (1975) indicated that Type III was a “poor rock” that should not
be included in the estimate of reserves. With the advent of horizontal drilling and multistage
hydraulic fracturing, however, economic oil production from this halo became feasible. This could
be visualized as oil being released from a “permeability jail” from that part of the TOF that is
affected by multistage hydraulic fracturing processes. Note that the TOF (Type III rock) reaches a
permeability close to 1 mD.
rp35
microns
1.0E+02 4.5
Type I
1.0E+01 1.8
1
PERMEABILITY (MD)
Type II
0.69
1.0E+00
Type III
1.0E-01 0.20
F
HR, B
1.0E-02
0.04
1.0E-03
0 3 6 9 12 15 18 21 24 27 30
POROSITY
Fig. 10.4—Flow units crossplot showing different rp35 flow units of Cardium Formation sandstones, Pembina oil field, Canada.
Data points for rock of Types I (asterisks) and II (red circles) were tabulated originally by MacKenzie (1975) and correspond to
a conventional oil reservoir. Porosities and permeabilities for Type III rocks (purple circles) were extracted from permeability-
porosity (k/φ ) data published by MacKenzie (1975) and correspond to a tight oil halo (TOF). Ranges of porosities and
permeabilities for Cardium tight oil (opena squares) were published by Hamm and Struyk (2011). Source: Aguilera (2014).
Permeability is not the only factor that plays a role in petroleum production from TPFs. A
cursory examination of the pseudosteady-state radial flow equation illustrates that petroleum rate
is a function of many other physical factors, including pressure, fluid properties, reservoir and
surface temperatures, net pay, drainage and wellbore radii, skin, and the non-Darcy constant
245
(discussed for TGFs by Holditch 2006). Furthermore, a TPF can be deep or shallow, high or low
pressure, high or low temperature, naturally fractured, contained within a single layer or in multiple
layers (discussed for TGFs by Holditch 2006), a continuous BCPA without a water leg (discussed
for TGFs by Law 2002; Schmoker 2005), or with characteristics of a conventional trap under
hydrodynamic influences (discussed for TGFs by Shanley et al. 2004; Aguilera et al. 2008). To
succeed and improve recoveries from TPFs, it is necessary to identify the location and preferential
orientation of natural fractures, to distinguish clearly between water- and petroleum-bearing
formations, to efficiently drill into and stimulate one or multiple zones, and to enhance the
connectivity between wells and their associated drainage volumes (Kuuskraa and Ammer 2004).
10.2.2.1 Continuous Tight Petroleum Formations. Continuous BCPAs are defined by the US
Geological Survey (Schmoker 2005) as “those oil or gas accumulations that have large spatial
dimensions and indistinctly defined boundaries, and which exist more or less independently of the
water column.” In addition, they commonly have low matrix permeabilities, are in close proximity
to “conventional” reservoir rocks, have low recovery factors (Schenk and Pollastro 2002), and are
visualized as a collection of petroleum charged cells. All of these cells are capable of producing
petroleum, but their production capabilities change from cell to cell, with the highest production
being obtained from cells with connected natural fractures and/or higher matrix permeabilities.
There are four key elements that define a BCGA as published by Law (2002), and these are
adapted for the BCPA as follows:
• Abnormal pressure
• Low permeability (generally ≤0.1 md for the case of tight gas and 1 md for tight oil)
• Continuous petroleum saturation
• No downdip water leg
If any one of these elements is missing, the reservoir cannot be treated as a continuous petroleum
accumulation. Note that lithology is not part of the four requirements listed above; the same four
elements have been reported for both clastic and carbonate reservoirs.
10.2.2.2 Conventional Tight Petroleum Traps. An opposite view to the concept of continuous
BCPA discussed above has been presented by Shanley et al. (2004, p.1083). These authors state
explicitly that “low-permeability reservoirs from the Greater Green River basin of southwest
Wyoming are not part of a continuous-type gas accumulation or a basin-center gas system in which
productivity is dependent on the development of enigmatic sweet spots. Instead, gas fields in this
basin occur in low-permeability, poor-quality reservoir rocks in conventional traps.”
The model used by Shanley et al. (2004) to explain their theory is called “permeability jail.”
The concept was developed originally by A. Byrnes of the Kansas Geological Survey based on
laboratory work conducted at room-temperature conditions and at 4,000 psi overburden stress
(Shanley et al. 2004). The “permeability jail” concept indicates that a range of saturations exists,
within which the relative permeabilities to gas and water are equal to zero; that is, the relative
permeabilities do not cross each other as in the case of conventional reservoirs (or, theoretically,
they might cross at extremely low permeabilities).
The controversy over whether TGFs are basin-centered features or occur in low-permeability
conventional traps is important because the estimates of original gas in place (OGIP) volumes and
mobile gas are much larger in a BCGA compared with those in discrete conventional traps.
Could the TGF “permeability jail” concept have application in the case of TOFs? Probably,
yes; if it works for gas, then it should work for oil; i.e., it should work for petroleum. How did
oil get into the pores? Probably through normal early migration from a source rock and charge
of the pores. Following storage of oil in pores, the pore throats were diminished or destroyed
246
through geologic time, for example, due to burial, creating the jail from which liquids (oil and
water) cannot escape.
An example of a TOF is provided by the Cardium Formation in the Pembina Field of Canada.
Oil has been produced from conventional Cardium oil reservoirs for several decades, but there is
a halo of very low permeability TOF surrounding the good-quality rocks.
10.2.3 Assessment Methods. The integration of geoscience and engineering aspects is of

paramount importance in exploring for and assessing TPFs. Folding, faulting, natural fracturing,
in-situ stresses, multilayer systems, mineralogy and petrology, connectivity and continuity,
permeability barriers, and interbedded coals and shales are just some of the aspects that must be
taken into account when evaluating TPFs (Aguilera et al. 2008). These are affected by the
dominating tectonics, which in the case of the Rocky Mountain basins are wrench/extensional,
while in the Western Canadian Sedimentary Basin, they are compressional (Zaitlin and Moslow
2006). Uncertainty evaluation methods should be used whenever possible while assessing TPFs
(see Chan et al. 2012).
10.2.3.1 Exploration Methods. Geoscience focuses on how to identify storage and source
rock potential, evaluate thermal maturation history, and locate swarms of natural fractures, positive
closures, and “sweet spots” of higher matrix permeability. Once these are located, and natural
fracture orientations are determined, wells are drilled in a way that intersects the natural fractures.
Inducing formation damage must be avoided as much as possible, which typically involves the use
of underbalanced drilling. However, even if the reservoir is not damaged, stimulation(s) of the TPF
will likely still be required to establish economic production. To be commercial under PRMS
guidelines (§ 2.1.2.1), in addition to technical development feasibility, the project must include
economic, legal, environmental, social, and governmental viability.
10.2.3.2 Geophysics. Seismic velocity reductions can indicate zones of high porosity, while
variations in seismic velocity with direction (azimuthal anisotropy) can be related to fractures in
the rocks. Wide-azimuth seismic acquisition and processing techniques may allow the detection
of natural fractures, which appear as wavy or sinusoidal reflectors on the seismic data. Although
ideally good seismic data should be used, Wang et al. (2015) have shown a feasible application of
using mediocre-quality poststack three-dimensional (3D) seismic data and attributes to improve
the characterization of a TGF in the Lower Permian Xiashihezi Formation, Ordos Basin, China.
The reservoir evaluation indicates that the Pareto principle (80/20 rule) helps to enhance the insight
needed to identify the trend of higher-productivity regions in the Ordos Basin.
Ant tracking (Pedersen et al. 2002; Liang et al. 2015) is another approach that offers hope for
locating fracture swarms. The technique has been found to be useful for automatic determination
of fault surfaces from conditioned fault-enhancing attributes. In those instances where the fractures
are fault related, the method can provide indirect indications of areas where the fractures
are located.
10.2.3.3 Geomechanics. An integrated approach using geomechanical methods, shear wave
splitting, P-wave azimuthal velocity anomalies, cores, and image logs (Billingsley and Kuuskraa
2006) has proven to be useful for locating natural fractures in three distinct geologic settings and
tight gas basins in the US: the Piceance and Wind River basins in the Rocky Mountains, and the
Anadarko Basin in western Oklahoma. Under favorable conditions, this technology allows fracture
density and apertures to be estimated. This technology was reported to improve ultimate recoveries
significantly in lenticular gas plays of the Rulison Field in the Piceance Basin from 0.9 Bcf/well
247
in 1956–1972 to 2.0 Bcf/well subsequently. The number of dry holes also dropped from 45% to a
low percentage (Billingsley and Kuuskraa 2006).
The recognition of fractures, slots (Soeder and Randolph 1987; Byrnes et al. 2006a, 2006b;
Aguilera 2008), and the best porosities allows optimum positioning of drilling targets and,
consequently, a reduction in capital and operating costs (Aguilera and Harding 2007).
10.2.3.4 Hydrodynamic Studies. These studies, often summarized as crossplots of pressure
from wireline formation tests vs. depth (Masters 1984; Aguilera and Harding 2007), must be
conducted to determine if the TPF is over- or underpressured, whether it has downdip (or updip)
water, if it is continuously petroleum saturated, and what the approximate size is of the TPF. This
work is useful in determining whether the TPF is a continuous BCPA (updip water) or a
conventional structural or stratigraphic low-permeability trap (downdip water leg). This work also
is very important in planning the development strategy of the reservoir. If the TPF is a continuous
petroleum accumulation, then large problems with water production probably will not be an issue
(Solano et al. 2011). However, if the hydrodynamic study shows the presence of a downdip water
leg, then it is reasonable to anticipate that eventually there will be water-production problems
(Shanley et al. 2004).
TGFs can act in some cases as a natural gas storage facility that feeds gas to a higher-
permeability medium that in turn feeds gas to the wellbore. This happens in the Western Canada
Sedimentary Basin, with the Cadomin Formation conglomerate feeding gas to the wellbore. As
the Cadomin Formation pressures drop, the surrounding TGF starts feeding gas into the higher-
permeability conglomerate (Zaitlin and Moslow 2006), which in turn moves gas to the wellbore.
10.2.3.5 Petrophysics. Although porosities are lower in TPFs, this does not necessarily
translate into lower calculated petroleum saturations. The reason for this is that there are lower
values of the Archie cementation exponent, m, in TPFs, resulting from the presence of fractures
and slot pores (Byrnes et al. 2006a, 2006b; Aguilera 2008, 2018). The recovery efficiency,
however, would be generally lower than that in a conventional petroleum reservoir due to the low
matrix permeabilities.
An excellent and valuable compilation of rock properties for the Mesaverde Group (gas) has
been published by Byrnes et al. (2006a, 2006b) for the Green River, Piceance, Powder River, Sand
Wash, Uinta, Washakie, and Wind River basins in the Rocky Mountains region of the US. Their
work involves routine in-situ porosity, permeability, and grain-density measurements, along with
special core analyses, including cementation and saturation exponents, cation exchange capacities,
mercury injection capillary pressures, drainage critical gas saturations, thin sections, and core
descriptions. Ideally, the same type of information should be collected for all TPFs, along with the
most recent generation of well logs, including image logs and nuclear magnetic resonance logs.
The work of Byrnes et al. (2006a, p. 9) also shows that the value of the cementation exponent,
m, becomes smaller as porosity decreases. They related the low values of m to the presence of slot
pores in the rocks, and they stated that “this pore architecture is similar to a simple fracture that
exhibits cementation exponents near m = 1.” The slot porosity can be visualized as grain-bounding
fractures that result from uplifting and cooling (Billingsley and Kuuskraa 2006).
10.2.3.6 Well Testing. The planning and analysis of well tests require specialized methods
due to the very low permeabilities of TPFs. Methods for single- (Lee 1987) and dual-porosity gas
reservoirs (Shahamat and Aguilera 2008) using type curves are available for this purpose. These
methods assume that radial flow is reached during the test, which generally implies the analysis of
vertical wells such as those drilled in some thick TPFs. However, horizontal wells are now
prevalent, and linear flow is typical in these instances (for basic principles, see, for example,
248
Palacio and Blasingame 1993; Arevalo-Villagran et al. 2006). Under favorable circumstances,
estimates of permeability and OGIP can be determined with a flowing-gas material balance
(Rahman et al. 2006). Given that well spacing is smaller in TPFs than in conventional reservoirs,
single-well simulators can provide reasonable results in some instances.
10.2.3.7 Production Decline. Analysis of decline curves using pressure-drop normalized
petroleum rates (see Section 10.4.3) can provide good results for estimating performance in
vertical, slanted, and horizontal wells, especially if wells have been producing for several years. If
normalization is not possible because of the lack of pressure data, hyperbolic decline can be used
to obtain generally reasonable results, provided it is modified with an exponential terminal decline
rate. It is important to monitor the forecast production period, so that estimates of ultimate recovery
are not skewed by very long production periods.
In addition to the theoretical models mentioned above, there also have been many useful and
practical empirical models based on actual recordings of production rates in TPFs. Production
examples from the Nikanassin TGF have been published by Solano et al. (2011) and Zambrano et
al. (2016). Production examples from the Cardium TOF have been published by Hamm and Struyk
(2011) and Aguilera (2014). Additional approaches are presented below in more detail in Section
10.4—Evaluation Methodologies for Tight/Shale Oil and Gas.
10.2.3.8 Machine Learning. Given the large volumes of available data in petroleum
reservoirs, machine learning algorithms are increasingly being used in efforts to understand all
types of petroleum formations, including TPFs. An important segment of machine learning deals
with decline rates of petroleum production. Although machine learning is extremely valuable, it is
important to emphasize that it must not be used without consideration of the reservoir physics.
Otherwise, results might be greatly misleading.
10.2.4 Drilling, Completion, and Stimulation Issues. Knowledge of in-situ stresses and natural
fracture(s) strike and dip is important. Vertical wells are not as efficient for intersecting vertical
fractures, which tend to have high inclinations at the depth of interest. This has led to a clear
preference for slanted and particularly horizontal wells in TPFs. An important additional benefit
is that a larger surface area of the reservoir is contacted by horizontal wells. The accepted concept
in TPFs is that the horizontal well must be drilled perpendicular to the open fractures (and
perpendicular to the maximum compressional stress). If more than one set of open fractures is
present, a properly designed slanted, horizontal, or multilateral wellbore can maximize petroleum
production and recovery by intersecting as many fracture sets as economically possible.
In conventional drilling, the mud weight is chosen to exceed the reservoir pressure to avoid
potential blowouts. In TPFs, however, mud invasion can result in large values of skin factor because
these formations are highly susceptible to damage. The problem is exacerbated because of the
complex geology of TPFs, which includes natural fracturing (causing fluid leakoff and potential sand
screenouts), folding and faulting (resulting in high stresses that could make initiation of the hydraulic
fractures difficult or impossible), and channel sands and interbedded coals and shales (resulting in
leakoff into cleats or unexpected fracture-propagation paths) (Bennion et al. 1996).
As a result, underbalanced drilling appears as a reasonable approach for drilling TPFs. In
underbalanced drilling, the usual mud is replaced by fluids such as inert gases and foams to make
the hydrostatic pressure exerted on the reservoir lower than the reservoir pressure. This eliminates
fluid invasion through the fractures and, consequently, minimizes damage to the TPF. Downhole
sensors near the drill bit gather and send information to the surface, which permits the bit to be
249
steered through the best portions of the reservoir, improving the probability of success (Bennion
et al. 1996).
Unfortunately, underbalanced drilling is not a panacea in TPFs because it can sometimes
induce severe unforeseen damage. Some of the potential problems include (Craig et al. 2002): fluid
retention, adverse rock/fluid and fluid/fluid interactions, countercurrent imbibition effects, glazing
and mashing, condensate dropout, and entrainment from rich gases, fines mobilization, and solids
precipitation.
Hydraulic fracturing jobs (single or multistage) are necessary in most cases in TPFs, even
when drilling slanted or horizontal wells. However, water retention is a significant problem in
some TPFs. As a result, many potential hydraulic fracturing technologies have been attempted in
the past, including fluids such as pure oil, CO2-energized oil, and crosslinked, water-based
polyemulsion and water-based foam (Craig et al. 2002; Rahman et al. 2006). Many operators must
experiment with the fluid types, proppant concentrations, etc., to find the method that works best
for their TPF.
10.2.5 Processing and Marketing. A general observation based on experience is that where there
is “conventional gas,” there is also “tight gas” (Aguilera et al. 2008). As stated by Salvador (2005),
“tight-sand accumulations should occur in all or nearly all petroleum provinces of the world.” As
a result, the processing and marketing of tight gas could proceed hand-in-hand with that of
conventional gas. Stranded gas, both from conventional and unconventional reservoirs (including
TGFs), requires special handling and economic considerations due to the very large investments
required. TOFs are also present in many places around the world, and processing and marketing
of tight oil could also proceed hand-in-hand with that of conventional oil. In all cases, the PRMS
guidelines would still apply.
10.2.6 Commerciality Issues. Economic considerations in TPFs have to take into account special
drilling, stimulation, and completion practices. In the case of gas wells, transient-flow periods can
last for several years before encountering any reservoir boundary or the production effect of an
offset well. Larger numbers of wells, whether vertical, slanted, or horizontal, per unit area are
always required in TPFs compared to conventional reservoirs (PRMS § 2.4.0.2). In order to move
some of the huge TPF resources into Reserves, efforts need to focus on many technological
improvements that have the potential to reduce costs and increase production rates. Generally,
companies operating in TPFs have been very successful at reducing costs and increasing petroleum
production rates. “Learning curves” along these lines as well as uncertainty have been discussed
in the literature (Chan et al. 2012) and are now defined in the PRMS as “demonstrated
improvements over time in performance of a repetitive task that results in efficiencies in tasks to
be realized and/or in reduced time to perform and ultimately in cost reductions” (PRMS Appendix
A—Glossary). In the case of gas, the handling of liquids, even in continuous accumulations
without downdip water, is an important consideration that must be taken into account when
producing TGFs in order to optimize production.
10.2.7 Classification and Reporting Issues. The PRMS classification, categorization, and
definitions are applicable to TPFs. Depending upon the assessment of commerciality, including
the quality and completeness of geoscience, engineering, and economic data, the quantities
resulting from the project’s evaluation could be classified either as Reserves (categories 1P/2P/3P)
or Contingent Resources (categories 1C/2C/3C). The undiscovered petroleum can be classified as
250
Prospective Resources (categories 1U/2U/3U). Once a project satisfies all the required
commerciality criteria, the associated Contingent Resources can be classified as Reserves. (One of
the PRMS criteria is that development proceeds within a recommended period of 5 years from the
initial classification date, unless a longer time period is justified. Contractual obligations, such as
rig availability and mineral extraction agreements, as well as the complexity of the development,
such as an offshore environment or governmental negotiation, may justify a longer period.)
However, given the characteristics of TPFs discussed previously, there are some differences
with respect to conventional reservoirs that should be highlighted, including the following:
• In spite of low porosities, the expansive continuity of a BCGA suggests that the volume
of gas initially in place is generally much larger in TGFs located within BCGAs compared
with that in conventional reservoirs. To avoid being overly optimistic (Schmoker 2005,
p. 2), the “assessment scope needs to be constrained from that of crustal abundance to
resources that might be recoverable in the foreseeable future.” Similar principles can be
applied in the case of TPFs. The petroleum volume of a BCPA would be classified as
total petroleum initially in place in the PRMS guidelines. At a smaller scale, it could be
divided between Discovered Petroleum Initially in Place and Undiscovered Petroleum
Initially in Place. Although there would be little doubt about the existence of the TPFs,
the uncertainties associated with economic productivity (such as the presence of natural
fractures, higher matrix permeability, low values of water saturation, resource maturity,
and the size of individual well drainage areas) will all affect whether the accumulation
can progress from Prospective Resources to Contingent Resources to Reserves.
• The recovery efficiency, as a percentage of the total petroleum initially in place in the
entire BCPA without a water leg, is likely much lower (on the order of 10%) than that in
a conventional reservoir. However, the recovery efficiency from a given property (lease
or license area or study area) located in a sweet spot with high porosity and high
permeability (maybe due to natural fractures) within the continuous accumulation can
reach 50% or more in the case of gas (see, for example, Jenkins 2009). The bulk of the
resources may be categorized initially as Contingent Resources but can move very rapidly
to Reserves if the project’s commerciality threshold is met. However, care must be
exercised because it takes hard data to step out far from discovery or appraisal wells to
the bulk of the resources. For a given unconventional property, it is also important to
remember that altogether a small percentage of the wells will contribute to the bulk of the
petroleum production.
Creties Jenkins
10.3 Shale Gas and Oil

10.3.1 Introduction. Shale gas and oil are produced from organic-rich mudrocks (siliceous,
calcareous, or a combination), which serve as the source, seal, and reservoir for hydrocarbons.
Shales have very low matrix permeabilities (tens to hundreds of nanodarcies) and typically require
multistage hydraulic fracture stimulations in horizontal wells to produce at economic rates. Any
sedimentary basin with a working petroleum system and shale source rocks has the potential for
shale gas and/or oil production. The challenge is to locate sufficiently large areas where the rock
and fluid characteristics of the shale are favorable and where wells can be drilled, completed, and
produced commercially.
251
10.3.2 Reservoir Characteristics. Organic-rich shales are complex rocks that exhibit
submillimeter-scale changes in mineralogy, grain size, pore structure, fracturing, and other
properties that control hydrocarbon storage and productivity (Jenkins 2016a). For many decades,
shales were considered to be homogeneous, isotropic, and laterally extensive because the degree
of heterogeneity was not observable in outcrops, cuttings, or even thin sections. Only in the past
decade have we begun to unlock the mysteries of nanoscale variability in these rocks with scanning
electron microscopes and other high-resolution tools. This work shows that we should not only be
linking variability in well performance to variations in drilling and completion practices, but also
to variations in rock and fluid properties within the shale. Shales are not, as one engineer described
them, “pristine until disturbed by drilling and completion activity.”
Shales, and other unconventional accumulations, are sometimes referred to as “statistical
plays,” implying that well performance variability is primarily the result of random variations in
reservoir properties from well to well. This is misleading because, while there is an element of
randomness, there are reservoir property trends controlling the average performance of well groups
that can be mapped and quantified. Understanding these trends and the information they tell us
about the better-producing areas, or intervals/landing zones, will allow us to concentrate drilling
in locations with the greatest production potential.
The complexity of these reservoirs means that a comprehensive suite of information is needed
to fully characterize them for the purposes of wellbore design, spacing, and construction
(Table 10.2). Collecting, analyzing, and integrating this information, and using it to help make
good decisions whether to proceed or exit from a project, can be very cost-effective relative to
relying on expensive trial-and-error drilling.
Category Information
Geology Outcrop descriptions, bedrock geology, mudlogs, formation tops, mineralogy, structural geology, faults,
fractures, stratigraphy, biostratigraphy, potential geohazards, topographic maps, geocellular models
Geochemistry Kerogen types, total organic carbon, RockEval pyrolysis, kinetics, borehole temperatures, thermal
maturity, biomarkers, isotopes, analysis of saturates-aromatics-resins-asphaltenes (SARA), gas
chromatography
Geophysics Two-dimensional (2D) seismic lines, 3D seismic volumes, vertical seismic profiles, check shots,
microseismic data, gravity, magnetics, remote sensing (satellite images, aerial photos), topography, in-
situ stress magnitudes and orientations
Cores Core descriptions, electrical properties, routine core analyses, sorption isotherms, gas contents
(desorption), capillary pressure, formation compressibility, digital rock physics, X-ray diffraction (XRD),
scanning electron microscopy (SEM), nuclear magnetic resonance imaging (NMR)
Wireline logs Spontaneous Potential (SP), spectral gamma ray (GR), array resistivity, microlog, caliper, density,
neutron, dipole sonic, image, geochemical, nuclear magnetic resonance, pulsed neutron
Geomechanical Principal stress orientations and magnitudes, elastic properties (Young’s modulus, Poisson’s ratio),
properties natural fracture orientation and density, unconfined compressive strength, geomechanical models
Reservoir Fluid types, contaminants (CO2, H2S), pressure-volume-temperature (PVT) properties, static reservoir
pressures, permeabilities, flowing bottomhole pressures, reservoir boundaries (faults), well testing (e.g.,
diagnostic fracture injection tests or DFITs), analogous reservoir data, analytical/numerical models
Drilling Mudgas volumes and compositions, rates of penetration, drilling histories, wellbore configurations, fluid
kicks and losses, drilling problems, formation pressure, drilling practices, cementing, casing integrity
tests, directional surveys
Completions Fracture staging, perforation strategies, proppant and fluid types/volumes, pumping schedules, treating
pressures, fracture diagnostics, microseismic data, tiltmeters, completion histories, refracturing
operations
Production Bottomhole (preferably) flowing pressures, well rates (oil, water, gas), production logs,
radioactive/chemical tracers, compression, artificial lift, choke sizes, water source and disposal intervals,
wellbore schematics, distributed temperature and acoustic surveys, facility diagrams and constraints,
nodal analysis
Health, Safety and Wellbore integrity, waste disposal, pollution/depletion of groundwater, air and water quality, spills, noise
Environment (HS&E) and light pollution, seismicity, well and site abandonment
General Well locations, lease boundaries, lease terms, regulatory environment
Table 10.2—Summary of useful information for characterizing and quantifying shale gas and oil reservoirs.
252
10.3.3 Drilling and Completions. Although initial shale wells may be vertical in order to evaluate
reservoir properties with cores and logs, appraisal and development wells are almost exclusively
horizontal. These wells, which can be 15,000 ft long or more, are landed in intervals with favorable
properties, including geomechanical features, allowing hydraulic fractures to be generated and
propagated into adjacent rocks in order to drain them (Jenkins 2016b). The dominant fluid used
for stimulation is water containing a friction reducer (slickwater), although gel (linear or cross-
linked) may be used to increase the proppant carrying capacity, build greater fracture width, and/or
reduce the amount of water used.
In addition to selecting the appropriate fracturing fluid, decisions need to be made regarding
the length of the well, length of the stages, perforation density, pounds of proppant per stage,
barrels of fracturing fluid per stage, and treating pressures, among others. These factors are
sometimes referred to as the “drilling and completion intensity,” whereby more perforations,
proppant, and fluid pumped in shorter stages at higher pressures result in greater intensity.
Increasing the intensity in recent years has been rewarded with higher productivity and value, but
there is a point of diminishing returns where stress shadows, interference between stages, and
interference between wells can reduce project value.
This situation is most obvious in parent-child well relationships. In the US, horizontal wells
are typically drilled at a well density of one well per square mile to hold acreage by production,
followed by development infill drilling, if initial production characteristics are favorable. Many
companies have overestimated the forecasted production of these bounded infill wells, believing
that their rates would be similar to the initial unbounded wells. In addition, multiple companies
have implemented a cube strategy, whereby tens of wells are spaced horizontally and stacked
vertically to drain multiple horizons in a given area. While some of these projects have met
expectations, many have not, because the wells were too close together, the drilling and completion
intensity was too great, and/or reservoir heterogeneities such as natural fractures contributed to
connecting the wells. Recognition of the uncertainties associated with parent-child well
relationships enters into the reserves categorization or resources classification process.
Other production issues that exacerbate commercial recovery include condensate banking in
rich gas and rock compaction as the reservoir pressure depletes with production. Effects such as
these may contribute to a loss in productivity in excess of 50% (for example, see Ayyalasomayajula
et al. 2005), unless properly managed.
These disappointments are best mitigated by demonstration (pilot) projects. For example,
establishing a drilling and completion protocol and holding it constant for tens of wells will facilitate
understanding the range of well performance caused by changes in reservoir properties and
determining whether the average well in this group is a commercial success. Failure to do this may
result in a false positive (leading to noncommercial development) or a false negative (walking away
from a commercial development project). The same applies to implementing the cube strategy.
Demonstration projects using different well stacking/spacing combinations in different areas of the
field are recommended, but wells will need to produce long enough in each cube to ensure that the
initial rate and subsequent decline of the average well meet expectations and are commercial.
Given the tens of thousands of existing shale wells, there is abundant information available
from multiple shale plays regarding well performance and associated drilling and completion
practices. This information also includes the time and cost for drilling and completing wells, and
the ways in which these factors decrease as companies become more efficient in a new shale play.
Analyzing these data with statistical modeling and machine learning can provide insights as to
253
those factors that are the key drivers in maximizing value. These factors, in turn, can be used as a
starting point for developing best practices for a target shale reservoir.
10.3.4 Commerciality Issues. A key difference between conventional accumulations and shales
is that shales have a much greater chance of discovery, given that a trap is not necessary. This does
not mean, however, that the chance of a commercially successful project for shales is greater than
for conventional accumulations. In fact, they tend to be similar because while it is relatively easy
to get a flow of hydrocarbons from an organic-rich, thermally mature shale, it is much more
challenging to obtain economic producibility over critically sized areas.
Over the past decade, many shale projects have failed to deliver the value promised to
stakeholders. A common problem is failure to drill enough wells prior to development to
understand whether the development project will be a commercial success. This problem is
exacerbated by a desire to minimize the time and money spent in predevelopment. Similarly, a
large-scale project based on results that are more representative of “sweet-spot” wells is almost
certain to underperform.
With this in mind, shale projects should be evaluated in a series of stages (Fig. 10.5). The
decision to move from one stage to the next should be tied to economic thresholds such as
recovering the cost of individual horizontal wells in the deliverability stage and attaining
commerciality with pad drilling or cube projects in the demonstration stage. A development project
can then be implemented assuming the project is competitive with other opportunities in the
company’s portfolio. This process incrementally exposes larger amounts of investor capital in a
responsible manner.
Fig. 10.5—Decision tree illustrating the staged approach, where HC is hydrocarbon. (Note that Stage 3—Demonstration
Success considers not only the Present value (PV) > 0 requirement, but other PRMS commerciality criteria as well.)
254
In each stage, there is a need to identify the key uncertainties and risks, collect the data needed
to quantify these uncertainties and risks, and generate a probabilistic assessment of potential
outcomes and their associated value in order to compute an expected value for the project (Jenkins
and McLane 2019). Estimates of resources and reserves are linked with this staged process.
Prospective Resources are estimated prior to drilling the discovery well, Contingent Resources are
estimated thereafter, and these are converted to Reserves when the demonstration and development
projects are determined to be commercially successful.
Two other elements important to commercial success are assurance and performance reviews.
An assurance process within companies includes clear guidelines and workflows to ensure the
fundamental quality of technical work and peer reviews and assistance by subject matter experts.
Performance reviews help in understanding whether the expectations of a given project stage were
met, and if not, what should be done differently to achieve greater agreement. Applying these
techniques to assess, approve, carry out, and monitor projects is critical for establishing
consistency in performance and generating confidence in the investment community.
Dilhan Ilk
10.4 Evaluation Methodologies for Tight/Shale Oil and Gas

10.4.1 Introduction. Analysis and forecasting of production data in unconventional reservoirs
continue to be problematic because there is considerable uncertainty related to our current lack of
understanding of the relevant fluid flow phenomena. There are many unknowns, such as the link
between flow in nano- and macroscale features, the effect of natural fractures and stress
fields/geomechanics, fracture propagation, permeability, pressure-dependent reservoir properties
(e.g., permeability, porosity, and fracture conductivity), and phase behavior at nanoscale, which
are all primary sources of uncertainty in long-term production performance. Recently, a great
amount of research focusing on these unknowns has been performed to understand and relate these
issues to well performance analysis and forecasting, but significant uncertainty still remains,
particularly for short-term production data.
Generally speaking, conventional decline curve equations are utilized with some adjustments
(e.g., the modified hyperbolic equation, Ilk et al. 2008) for forecasting production in
unconventional reservoirs. These equations are empirical in nature, and attempts to link the results
of batch processing of large numbers of wells and their decline curve relations to detailed reservoir
engineering work are not very common. While it is tempting to use these relations due to the notion
that they are “easily understood and repeatable,” a thorough understanding of uncertainty in
forecasts and the effects of well/reservoir properties on performance are generally missed. On the
other hand, methods attempting to account for well/reservoir properties, fluid properties, and
drainage area complexities are often considered to be elaborate, laborious, and unreliable within
the context of production forecasting for large numbers of wells. However, these additional steps,
when taken on key wells, often provide time-saving insight in the evaluation of a larger population.
10.4.2 Overview. From a historical perspective of decline curve analysis (DCA), Arps’ hyperbolic
and exponential relations have been the industry standard for production forecasting and
estimating ultimate recoveries for over 70 years. However, in unconventional reservoirs, the
application of these relations may be problematic due to invalid assumptions. The main
assumptions forming the basis of traditional DCA include the following:
255
• Any changes in operating conditions or field development during the producing life of
the well are minor.
• No major changes occur in well productivity (such as changing skin throughout
production).
• Drainage area remains constant.
• Production is achieved against a constant bottomhole flowing pressure.
• A boundary-dominated flow regime (reservoir depletion) has been achieved.
It is common to observe misapplication of traditional DCA in very low permeability reservoir
systems by ignoring these assumptions, frequently resulting in overestimation of reserves,
especially if the hyperbolic relation is extrapolated with decline exponents (widely known as b
exponents or b-parameter) greater than 1. To prevent overestimation, the hyperbolic equation may
be spliced with an exponential decline at late times, resulting in a modified hyperbolic equation.
This approach remains nonunique and may yield widely varying estimates of reserves, mainly due
to selection of b exponents and ambiguity when selecting minimum terminal decline values
(particularly in the absence of mature analogs).
Numerous authors [Ilk et al. 2008 (power-law exponential); Valkó 2009 (stretched
exponential); Clark et al. 2011 (logistic growth model); Duong 2011, 2014; Mishra et al. 2014;
Fulford and Blasingame 2013 (transient hyperbolic equation); Artus and Houzé 2018; Artus et al.
2019] have proposed various empirical rate decline relations, which are principally based on a
characteristic feature of the data and appear to match the early time transient and transitional
flow regimes.
Each model may best be described as empirical (i.e., no direct link with theory) and generally
centered on a particular flow regime and/or characteristic behavior. As such, all of these relations
may produce good matches across the entire range of production, but each relation may diverge in
the forecast and hence result in a broad range of estimates of ultimate recovery. In addition, each
equation may not apply to every play due to the unique production characteristics of the particular
unconventional reservoir.
These decline curve relations may be used in combination to obtain a range of outcomes rather
than a single estimated ultimate recovery value. The range of outcomes may be associated with
the uncertainty related to production forecasts and can be evaluated as a function of time.
Fig. 10.6a presents an example of the application of various decline curve relations to a shale gas
well on a classic semilog rate-time plot, while Fig. 10.6b presents results on a semilog rate-
cumulative plot. Note that the forecasts from each of the decline curve relations yield a range of
estimated ultimate recovery values, which is mainly a result of the different characteristic
behaviors of these decline relations. Aside from the range of estimated ultimate recovery values,
no direct link to reservoir and completion parameters can be obtained.
256
(a)
(b)
Fig. 10.6—(a) Production forecasting of a shale gas well using various decline curve relations (semilog rate and time plot).
Abbreviated decline curve relations: PLE refers to “power-law exponential” decline, MH refers to “modified-hyperbolic,”
“DNG” refers to Duong, “LGM” refers to logistic growth method, and THP refers to “transient hyperbolic.” (b) Production
forecasting of a shale gas well using various decline curve relations (semilog rate and cumulative production plot).
Fig 10.7 illustrates the various flow regimes that may be encountered in a multistage fractured
horizontal well. It should be noted that this sample illustration is only conceptual and was derived
using an analytical or numerical solution for a horizontal well with multiple transverse fractures
(for first derivation, see van Kruysdijk and Dullaert 1989). There are many variations of this
solution yielding similar behavior (i.e., flow regimes). The following model parameters were
assumed to generate the response on Fig. 10.7:
k = 0.0005 md (500 nd) (permeability)

xf = 300 ft (fracture half-length)
Fc = infinite conductivity (fracture conductivity)
s = 0 (skin factor)
nf = 40 (number of fractures)
A = 640 acres (well spacing)
257
Fig. 10.7—Typical (conceptual) flow regimes for a horizontal well with multiple transverse fractures, where SRV is
stimulated reservoir volume.
Fig. 10.8 illustrates actual generic production responses (over 3 years and plotted in terms of
pressure drop–normalized rate vs. production time) from several unconventional plays in North
America. It should be noted that generic production responses are based on the average behavior
of certain well groupings (typically consisting of 100 or more wells) and should not be considered
as definitive for each play. In other words, the duration of the linear flow regime and late time
decline behavior may be different than what is plotted here due to many factors such as well
spacing, completion design, and reservoir and fluid properties. These plots help compare actual
well behavior to theoretical responses.
For each play illustrated in Fig. 10.8, an uncertainty region beyond historical production is
imposed to illustrate uncertainty with respect to future production behavior. This uncertainty
region can be used to derive decline curve parameters associated with uncertainty in estimates of
ultimate recovery. Early time behavior appears to be mostly impacted by well cleanup effects. It
is interesting to note that, after cleanup, actual production behavior observed in various
unconventional plays is typically consistent with the first flow regime (linear flow), but it may
differ at a later time. Another point to consider is that linear flow associated with the “post-SRV”
is not commonly observed.
258
Fig. 10.8—Typical production responses (in terms of pressure drop–normalized rate) in various shale/tight oil and gas
plays in North America.
A generic workflow can be considered for analysis and forecasting of well performance,
preferably using high-frequency production rate data, such as at least daily pressure and rate data,
ideally with individually metered wells. Although it cannot be universally applied to all scenarios,
it is consistent and repeatable in many cases. First, data quality control and well performance
diagnostics are performed. In this stage, outliers and inconsistent data are identified and may be
removed to achieve a better identification of performance trends. Production diagnostics are
259
important when interpreting flow regimes, quantifiable measures of productivity (productivity
metrics), well groupings, and characteristic group behavior. Similarly performing wells are
identified and well groups are created before representative wells are picked for detailed model-
based analysis.
Next, model-based analysis is performed to analyze and forecast well performance. During
this step, it is important to consider the most appropriate model to use and incorporate relevant,
play-related factors that affect flow behavior, such as stress dependencies, drainage area patterns,
and pressure-volume-temperature and multiphase behavior. In this analysis, ranges are estimated
for the unknown parameters, which will result in a range of solutions at the end of process rather
than yielding a single answer. Finally, model-based analysis results (i.e., forecasts) can be
extended to the other wells in the well group.
Sections 10.4.3 through 10.4.5 present detail on the steps of the generic workflow.
10.4.3 Production Diagnostics. The first step in the workflow is to fully evaluate the production
data using diagnostic tools. The data utilized for this process are not only production data (time-
rate, time-rate-pressure), but also well completion data and reservoir data. A complete diagnostics
study should include the following activities:
Single-well basis:
• Review of data quality and consistency
• Identification of data features and characteristics (for example, effects of tubing
installation, effects of offset well fracturing, etc.)
• Identification of flow regimes
Multiwell basis:
• Comparison well performance through multiwell plots
• Identification of well performance indicators (productivity metrics)
• Identification of groups of wells with similar behavior
• Selection of representative wells for detailed model-based analysis
Throughout the data quality and consistency checking process, off-trend or spurious data can
be removed (or isolated) on diagnostic plots for flow-regime diagnosis. In addition, erratic rate
data can be removed from semilog time-rate plots to improve numerical differentiation quality and
obtain D and b parameters as functions of time (Ilk et al. 2008). For reference definitions of D-
and b-parameters are provided below (where q is rate):
1 𝑑𝑑𝑑𝑑
𝐷𝐷 ≡ −
𝑞𝑞 𝑑𝑑𝑑𝑑
𝑑𝑑 1 𝑑𝑑 𝑞𝑞
𝑏𝑏 ≡ � �≡− � �.
𝑑𝑑𝑑𝑑 𝐷𝐷 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑/𝑑𝑑𝑑𝑑
A visual inspection of production history plots of rate and time and flowing (surface or
bottomhole) pressure and time is very helpful to reveal inconsistencies (such as rates and pressures
increasing at the same time) prior to analysis. If inconsistencies exist and are not identified, one
can end up analyzing artifacts that have nothing to do with reservoir behavior. In all cases, care
should be taken to ensure no undue bias is applied that may exclude data points that are, in
fact, applicable.
260
Identification of flow regimes is the key objective of production diagnostics. It is incumbent
upon the evaluator to know when to use, or not to use, certain rate-decline relationships as a
function of the flow behavior stage of the well. As indicated earlier, the conditions under which
the Arps model may be applied are limited. In other words, there is no “one size fits all” empirical
rate-decline model. Therefore, it is critical to utilize diagnostic plots to identify characteristic
decline behavior, which could indicate an appropriate decline curve model to use or a decline
parameter (e.g., Arps’ hyperbolic exponent) specific to an area of interest.
For this purpose, log-log plots of rate and pressure drop–normalized rate are utilized. Pressure
drop is calculated by subtracting flowing bottomhole pressure from initial pressure (or average
reservoir pressure, to be theoretically valid). Practically speaking, average reservoir pressure
measurements are not common in unconventional reservoirs, and flowing surface pressure data
can be utilized in the absence of bottomhole pressures. On the time scale, standard production time
or material balance time (MBT) is used. MBT is the ratio of cumulative production to
instantaneous rate. The goal is to identify “power-law” signatures, which are observed as
straightline trends on a log-log plot. The observed slope of power-law signatures could translate
to possible flow regimes such as:
• Quarter slope → Bilinear flow (finite conductivity fracture)
• Half slope → Linear flow (infinite conductivity fracture)
• Unit slope → Depletion-type flow (this is only valid for the MBT plot)
Fig. 10.9 presents a schematic of “power-law” signatures using a diagnostic MBT plot
(sometimes referred to as a “Blasingame plot”). On this plot, red points indicate linear flow
behavior, while yellow points suggest bilinear flow behavior, and gray points indicate depletion
flow behavior (see Acuna 2016, 2017; Acuna et al. 2018; Chu et al. 2017; Raghavan and Chen
2019 for other interpretations of “power-law” signatures).
Fig. 10.9—Schematic Blasingame plot for flow regime identification in rate or pressure drop–normalized rate and material
balance time (MBT) format.
261
In addition to production history plots (i.e., rate and production time, flowing pressures and
time), the following diagnostic plots are recommended to assist interpretation:
• [Log-log] Rate (or pressure drop–normalized rate) and time for flow-regime identification
• [Log-log] Rate (or pressure drop–normalized rate) and MBT for flow-regime identification
• [Cartesian] Reciprocal rate (or rate-normalized pressure drop) and square root time for
well productivity assessment (for productivity metrics)
• [Semilog] Rate (or pressure drop–normalized rate) and cumulative production for
production extrapolation
Although these plots are useful in many cases, the use of additional plots, which could be
relevant for specific circumstances (e.g., gas-oil ratio as a function of time, water rates as a function
of time, water-oil ratio versus cumulative oil production, derivative plots such as D parameter and
b parameter, etc.), is encouraged. Secondary phase relationships are necessary for the forecasting
of condensate (in a gas-condensate reservoir) and solution gas (in an oil reservoir), while water-
oil or water-gas ratios assist in identifying well cleanup. The evaluator should also monitor changes
in choke settings, which may influence flow rates and secondary ratios.
Multiwell diagnostic plots are utilized to identify characteristic group behavior, which may
imply a unique flow regime applicable for each well in the grouping. An identified unique flow
behavior, such as late time decline behavior, can be translated to a decline curve parameter (e.g.,
b factor) and used for each producing well and further may be utilized as analog behavior. Once
the characteristic features of well performance are identified, model-based analysis can be
performed to estimate well/reservoir properties, completion efficiency, and, most importantly,
future production.
The key step in this process is to group wells exhibiting certain features together. Along these
lines, important things to consider for well groupings are wells producing in a similar geological
area, similar fluid properties, and similar completions. (For more about assessing analogs, see
Chapter 4—Assessment of Petroleum Resources using Deterministic Procedures herein.) Next,
flow regimes and the other related features of each well are identified using single-well diagnostic
plots. After this exercise, wells can be plotted on the same plot to visualize if they exhibit similar
trends together or not. At this point, normalization by certain parameters can be used to eliminate
differences and to obtain a unique trend. The normalization parameters could be proxy parameters
(e.g., the cumulative production value of each well at a specified time) or well
completion/subsurface parameters (e.g., lateral length, amount of proppant pumped, propped
thickness, and initial reservoir pressure, among others).
10.4.4 Model-Based Analysis. For this discussion, model-based analysis refers to the use of
analytical and numerical models to analyze and model time-rate-pressure data in order to estimate
dynamic well/reservoir properties such as permeability, effective fracture half-length, fracture
conductivity, etc., through history matching (Stalgorova and Mattar 2012, Ilk et al. 2012, Atadeger
et al. 2020). Once a history match (such as shown in Section 10.4.5) is achieved, model-based
forecasting yields rate (or pressure) forecasts as a function of time.
Model-based analysis will require high-frequency time-rate-pressure data (in particular,
bottomhole pressure data). In the absence of bottomhole pressures, surface pressure data will need
to be converted to bottomhole pressure by using correlations or nodal analysis. Conversion to
bottomhole conditions may add complications to this analysis, especially in the case of multiphase
flow and artificial lift.
262
Pressure-volume-temperature (PVT) and static reservoir properties are also required to
construct the model. Actual pressure-volume-temperature laboratory reports are preferrable
(particularly for condensate and volatile oil cases), but correlations may be acceptable in the case
of dry gas. With ongoing research in progress for phase behavior in nanopores, it might be difficult
at this time to conclude how this behavior affects modeling. Petrophysical interpretation is required
from well logs and cores to estimate porosity, net reservoir thickness, and initial water saturation.
Net reservoir thickness is considered to be a parameter with high uncertainty, and different
interpretations are typically possible, such as propped fracture height. In order to quantify this
value, microseismic interpretation and geomechanical analysis may be utilized in support of
petrophysics.
Finally, well completion data are required to define well/reservoir geometry. Well completion
data should mainly include horizontal well length, number of fracture stages, and the number of
perforation clusters per stage. When multiple wells are modeled, the amount of proppant and fluid
pumped could be utilized to relate model-based analysis results to model parameters such as
effective fracture half-length.
There are general issues related to models in unconventional reservoirs. Some of the important
ones are listed below:
• Nonlinearities: Stress (pressure)-dependent rock properties, multiphase flow
• Complex fracture configurations, fracture geometries
• Interference between wells
• Uncertainty in drainage area (stimulated reservoir volume concept, discrete fracture
networks)
• Very low permeability
Most analytical models require special transformations to deal with some of the challenges
listed above. Numerical models, on the other hand, should be able to account for these challenges
with proper gridding and due to their ability to incorporate information from all sources (such as
geology, geomechanics, geophysics, and completion diagnostics, etc.). Generally speaking,
numerical models can be used to mitigate the limitations of analytical models in terms of
well/fracture geometry, nonlinearities, and complex diffusion processes. The main problem with
numerical models is to properly address high pressure gradients caused by the very low
permeability of the system. Correspondingly, this issue needs to be addressed by implementing
very refined gridding around the high-pressure-gradient locations such as the interface between
fractures and the formation rock. References in this section provide a comprehensive list of various
analytical, semi-analytical, empirical, and complex models for modeling flow behavior of
multifracture horizontal wells producing in unconventional reservoirs.
10.4.5 Well Performance Analysis and Forecasting Example. In this subsection, examples
are shown to demonstrate key aspects of previous subsections with data from producing wells
from an unconventional gas play in North America. In the first example, we focus on issues
involved with projecting ultimate recovery with limited data. Fig 10.10 illustrates curve fits to
production data at specific times throughout production history with specific b factors used for
production forecast.
263
Fig. 10.10—Production forecasts at specific times throughout production history.
It can be observed that the b factor continuously decreases compared to the value used in the
initial forecast. To summarize this observation, each forecast is plotted on one plot with actual
(most recent) production. As observed in Fig. 10.11, earlier forecasts with higher b-factor values
significantly overestimate the production.
Fig. 10.11—Production forecasts at specific times with most recent production history.
264
This observation points out the importance of understanding flow regimes encountered
throughout a well’s production history. In this example, decreases in the b factor used for
production forecasting are related to changes in flow regimes. Evaluators should be aware of this
situation when dealing with wells with limited production history.
The second example focuses on analysis of and forecasting for a group of wells with similar
behavior. Fig. 10.12 presents a log-log flow-regime analysis plot (rate and MBT) for wells
producing in a shale gas play. Each colored dots on Fig 10.12 represent a specific well behavior.
It can be observed that most of the wells show similar behavior, but the magnitude of production
responses is different.
Fig. 10.12—Flow-regime analysis plot for a group of wells.
To approximate the characteristic behavior of this group, rates are normalized by a specific
parameter (or a combination of parameters). In this example, cumulative production at 6 months
for each well is used as the normalizing parameter. Cumulative production at 6 months can be
considered as a proxy for including the effects of completions and localized reservoir properties.
Production behavior of this group is shown after normalization in Fig. 10.13 (note each colored
dot represents a specific well behavior after normalization).
Note in Fig. 10.13 that almost all wells exhibit linear flow behavior, and none of the wells
falls on the unit slope, which would indicate depletion. This observation may indicate challenges
with production forecasts based on existing data (with decline curves) as transition from linear
flow is expected at some point in future. Accordingly, model-based analysis is then utilized to
account for changing flow regimes.
Model-based analysis can be performed for each well; however, this may not be practical, as
it typically takes more time to perform a comprehensive analysis for a single well compared to
DCA. Therefore, it may be sufficient to perform model-based analysis for one “representative”
well from this group of wells as long as this group of wells exhibits a unique behavior.
Accordingly, in this example, we selected a representative well for model-based analysis.
265
Fig. 10.13—Flow-regime analysis plot for a group of wells after normalization.
After selecting the well, we start with data quality control and identification of specific events
or issues throughout the well’s history. Fig. 10.14 shows the rate and calculated bottomhole
pressure history from this well. There appear to be no major issues with production and rate, and
the calculated bottomhole pressures are in good agreement. Consequently, analysis of this well
should not be problematic.
10000
5000
5000
4000
3000
2000
1000
0
0 100 200 300 400 500 600 700 800 900
Production Time (days)

Fig. 10.14—Rate and calculated bottomhole pressure data for selected well for analysis.
266
After removing outliers, flow-regime analysis is performed next using diagnostic plots. For
this well, we observe linear flow behavior for almost its entire production history. Because one
can expect deviations from linear flow at a later time, production forecasting based on this
observation may potentially overestimate volumes. Considering this uncertainty, we impose an
uncertainty region on the diagnostic plot to illustrate potential flow regimes that may be
encountered at later times. Possible interpretations for potential decline behavior in this example
include potential “longer-term” linear flow, which could be used to model a “high case” forecast,
or an immediate transition to a steeper decline trend as depicted by the “unit slope” on Fig. 10.15
to help model a “low case” forecast. Characterization of the uncertainty using diagnostics helps
the evaluator to estimate the ultimate recovery values that can guide assessment of 1P, 2P, and 3P
(or 1C, 2C, and 3C) quantities once these empirical model evaluations are tied to actual well
recovery outcomes.
Fig. 10.15—Flow-regime analysis plot for the selected well.
Model-based analysis is then performed for the selected well. Although there are many models
described in literature used for analysis and forecasting, this example will be used to illustrate a
history match obtained with a relevant model suitable for a multifracture horizontal well. It is
recommended to obtain history matches using different combinations of modeling parameters to
account for uncertainty and nonuniqueness and to compare these results to the physical
characterization of the subject reservoir and fluids to identify those matches that compare
favorably to these properties.
Typically, there is significant uncertainty associated with model parameters such as
permeability, effective fracture half-length, and drainage area. Therefore, uncertainty analysis is
strongly recommended—this can be achieved by experimental design (see Collins et al. 2015) or
assuming probability distributions on model parameters (see Anderson et al. 2012). It is always
beneficial to incorporate additional sources of information such as reservoir characterization
267
studies and/or completion diagnostics (microseismic, tracer tests, geomechanical models) to
constrain key model parameters such as effective fracture half-length and number of contributing
fractures.
Fig. 10.16 illustrates the history match of rate and calculated bottomhole pressure data (with
only one realization of model parameters) accompanied by history matches on diagnostic plots.
Fig. 10.16—History match with one realization of model parameters. Bottom left plot shows normalized pressure drop data
(blue points) and its derivative (orange points). Bottom right plot shows normalized rate data (orange points), normalized
rate-integral data (blue points), and derivative of rate-integral data (green points). Solid lines indicate model response.
Once history matches are achieved, production is forecast with various realizations of the
model parameters. These forecasts can be achieved deterministically or probabilistically, which
should ultimately reflect the uncertainty associated with reserves and resources categories. Low,
best, and high case forecasts are achieved for this well using different values for reservoir
permeability, as presented in Fig. 10.17.
268
Fig. 10.17—Low, best, and high case production forecasts for selected well.
Finally, model-based forecasts can be extended to the other wells in this well grouping,
provided that wells in this group exhibit common characteristics, which are mainly functions of
similar completions and, more importantly, similar reservoir and fluid properties.
(Correspondingly, this group of wells can be considered as an analog well set.) Extension of
forecasts can simply be performed by a process similar to “type curve matching,” as illustrated in
Fig. 10.18. Model responses can be “shifted” on time and rate axes to match the production data
of other wells and consequently achieve production forecasts.
269
Fig. 10.18—Extension of low, mid, and high case production forecasts for other wells in the well grouping, where RTA is
rate-transient analysis.
10.4.6 Estimating Recoverable Quantities in Unconventional Resources. Unconventional

resources exist in petroleum accumulations that are pervasive throughout a large area and are not
significantly impacted by hydrodynamic influences. Although this definition implies a potentially
larger extent of resources in place, performance and productivity can be significantly different
across the play, which can greatly affect recoverable quantities. The PRMS (§ 2.4.0.2) indicates
that there is a need for increased sampling density to define uncertainty of in-place quantities and
variations in reservoir and hydrocarbon quality. Along these lines, extrapolating reservoir presence
or productivity beyond control points is not recommended in the absence of technical evidence to
support the conclusion. Increased well control/density may demonstrate consistency of well
production behavior across an area of interest, thereby bolstering the technical evidence.
As unconventional resource plays are scoped, discovered, and developed, well control and
supporting geoscience and engineering data will be acquired and interpreted. As understanding of
the play evolves, its maturity of development can be subdivided by phases. Estimation of
recoverable quantities in unconventional reservoirs should be conducted within the context of
development maturity. For simplicity, the recommended steps in estimating recoverable quantities
in unconventional reservoirs can be described as follows:
1. Early phase: This phase consists of a few wells (usually exploratory wells, either vertical
or horizontal) with limited production history. This period also may be referred to as the
exploration phase. In this phase, more emphasis is placed on data collection and
evaluation of uncertainty in reservoir and fluid properties across the area of interest. The
following steps are recommended for estimating recoverable quantities:
a. Evaluate pilot hole well log data.
b. Evaluate reservoir properties based on core and log data interpretation.
c. Assess geochemistry of the resource for oil/gas richness and maturation.
d. Review/evaluate/integrate geology, geophysics, petrophysics, and geomechanics
data across area of interest.
e. Estimate/review in-place quantities.
f. Establish uncertainty ranges on reservoir/completion parameters for dynamic
modeling.
g. Generate potential production profiles for undeveloped locations using
well/reservoir model and/or analog data (based on specific development plan[s]).
270
Note that approaches based on volumetrics (i.e., recovery factor multiplied by in-place
quantities) may not be applicable due to complexities in estimating the influence of well
spacing, depletion, completion design, and production operations on production profiles
and estimated recoveries from undeveloped areas.
2. Intermediate phase: This phase consists of several wells (on the order of 10 or more),
generally with longer-term production histories, but the area of interest is far from fully
developed, and completion design and well spacing are yet to be optimized. For this
phase, more emphasis is placed on evaluating spatial distribution of reservoir properties
and evaluation of potential well completion and field development scenarios. The
following steps are recommended for estimating recoverable quantities:
a. Evaluate reservoir properties based on core and log data interpretation of several
key wells.
b. Assess thermal maturity.
c. Review/evaluate geology, geophysics, and geomechanics data.
d. Estimate/review in-place quantities.
e. Analyze and forecast recoverable quantities for producing wells using model-based
analysis (i.e., rate/pressure-transient analysis).
f. Evaluate potential completion design and sensitivities using model-based
approaches and analogs.
g. Generate potential production profiles for undeveloped locations using
well/reservoir model and compare those to analogs.
3. Mature phase: This phase consists of many wells (typically more than 100), most with
long-term production histories with established trends, completion design, and well
spacing practices. For this phase, generally speaking, the distribution of reservoir
productivity is better understood, and most of the efforts are concentrated on performance
analysis of existing wells using DCA. Model-based analysis can be performed for select
key wells. Statistical methods are generally applicable to assess consistency of well
performance and extrapolate reservoir presence and productivity beyond immediate
offsets to producing well locations. Statistical methodology commonly used in the
industry is presented, for example, in Society of Petroleum Evaluation Engineers (SPEE)
Monograph 3 (SPEE 2010). Utilization of the “learning curve” (PRMS § 2.4.0.5) is much
more common during this phase. Evaluation steps followed during this period generally
consist of the following:
a. Review reservoir properties based on core and log data interpretation.
b. Review geology, geophysics, and geomechanics data.
c. Review in-place volumes.
d. Generate/review type well areas (i.e., areas with similar well performance and
analogous well behavior).
e. Forecast well performance of producing wells using DCA.
f. Evaluate potential field development scenarios, completions, and well spacing
options,
i. generally, by using observed (empirical) data, and
ii. occasionally, by utilizing models.
g. Generate potential production profiles for undeveloped locations by performance-
based analysis results and statistical methods using SPEE Monograph 4
methodology (SPEE 2016).
271
Table 10.3 summarizes the application of performance-based methodologies at each phase of
development maturity.
Performance-Based Methodology
Phase of Diagnostics Model-Based Analysis Production Profiles for

Development Undeveloped Locations (Areas)
Maturity
Early phase Diagnostics are of limited Use models or analogs to Rely on model results; use
use. address reservoir and distribution of sensitivity
completion uncertainty. analysis results.
Intermediate phase Full diagnostic Rate-transient analysis Production profiles are

interpretation can be (model-based analysis) is generated using rate-transient
performed on some wells. primary method of analysis.
analysis.
Mature phase Full diagnostic DCA is primary method Statistical methods (SPEE
interpretation can be (rate-transient analysis Monograph 3; SPEE 2010) are
performed on all (or can be used for select used in consideration of
majority) of wells. wells). depositional environment.
Table 10.3—Performance-based methodologies by development phase of maturity.
Assignment of resources and classification of undeveloped locations require interpretation

of productivity beyond existing location(s) and assessment of the corresponding uncertainty. Various
techniques and combinations of methodologies, which have been field tested and demonstrated to
provide results with high confidence, can be utilized to interpret reservoir productivity and justify
assigning resources categories to undeveloped locations beyond direct offsets.
Fig. 10.19 illustrates examples for the assignment of resources and classification of
undeveloped locations in unconventional resource plays assuming horizontal well development. It
is noted that this example considers the “deterministic incremental method”; however, results from
other methods such as the “deterministic scenario,” “geostatistical,” or “probabilistic” methods
should yield similar remaining recoverable quantities for the area of interest. In the example below,
classification and assignment are implied as a function of development phase.
Multiple scenarios are illustrated, and it is worth mentioning that there is no clear rule or
procedure for the number of locations and distance from existing producers to be considered.
Put simply, this task is based mainly on documented evidence of the consistency of well
performance, interpretation of reservoir productivity, and availability of critical data, and the
evaluator’s confidence in associating these with the appropriate uncertainty. It is also worth
mentioning that increasing consistency in well performance, corroborated by supporting multiple
data sources, and better understanding of the distribution of reservoir properties throughout the
development area could allow for resources assignment to incremental locations at greater
distances beyond producing locations, such as the Proved Undeveloped location in the upper-right
corner of the Mature Phase stage.
272
Fig. 10.19—An example illustration of assigning resources to undeveloped locations in the deterministic incremental approach.
Christopher R. Clarkson
10.5 Coalbed Methane

10.5.1 Introduction. The term “coalbed methane” (CBM) refers to natural gas hosted in coal.
Coal, in turn, is defined as a “readily combustible rock containing more than 50% by weight and
more than 70% by volume of carbonaceous material formed from compaction and induration of
variously altered plant remains similar to those in peaty deposits” (Schopf 1956). While gas
produced from coal tends to be “dry,” often consisting primarily of low-molecular-weight
hydrocarbons, with methane being the dominant hydrocarbon component, other non-hydrocarbon
components such as carbon dioxide may occur in significant amounts.
Referring to the definition of unconventional reservoirs provided in Table 10.1, coal qualifies
as unconventional because it commonly serves as both the source rock and the reservoir for its gas,
with much of the gas generated from biogenic or thermogenic processes retained in the coal in the
adsorbed state, CBM generally occurs as a continuous, laterally extensive accumulation, and
special extraction methods may be required to achieve commercial levels of production. A further
component often attributed to unconventional reservoirs (although not applicable to all
unconventional reservoirs) is low matrix permeability (often <<0.1 md, although system
permeability, which includes natural fractures, may be much higher than this). The high organic
matter content of coal is responsible for many of the unique reservoir properties of coal, which are
elaborated upon below.
10.5.2 CBM Reservoir Characteristics. Some reservoir characteristics (fluid storage and flow)
unique to two-phase (gas + water) CBM reservoirs are illustrated in Fig. 10.20. These
characteristics include gas storage dominated by adsorption (top left), existence of saturated or
undersaturated conditions (top right), gas flow via diffusion in the coal matrix, providing a source
to natural fractures through which both gas and water flow to the well (bottom left), and the
possibility of gas effective permeability increases due to relative permeability and absolute
permeability growth during depletion (bottom right).
273
Pure Component Isotherms and Mixture Isotherm Determinination of Saturation Level
800 600
Desorption canister test

600 - Saturated coal
Gas Content (scf/ton)
Gas Content (scf/ton)

400
400
Desorption canister test

200 - Undersaturated coal
200
Desorption pressure
0 0
0 500 1,000 1,500 0 500 1,000 1,500 2,000
Pressure (psia) Reservoir Pressure (psia)
CH4 CO2 Mixture Mixture adsorption isotherm Gas content - saturated Gas content - undersaturated
Effective Permeability to Gas Change with Pressure

1,000
Effective Permeability to Gas (md)

Diffusion of gas in matrix
100
10
Darcy flow of
gas and water
in fracture
1
1,600 1,400 1,200 1,000 800 600 400 200 0
Reservoir Pressure (psia)
Field data Exponential fit to field data
Fig. 10.20—Some reservoir characteristics unique to two-phase coal reservoirs. (Top left) Gas storage due to adsorption,
modified from Clarkson (2021). (Top right) Undersaturated vs. saturated coals, modified from Clarkson (2021). (Bottom left)
Single-phase flow (diffusion) of gas through matrix and two-phase flow of gas and water through the fracture system,
modified from Clarkson (2018), after Clarkson and Bustin (2011). (Bottom right) Effective permeability to gas growth due to
changes in relative permeability and absolute permeability (matrix shrinkage), modified from Salmachi et al. (2018), after
Clarkson et al. (2007b).
10.5.2.1 Gas Storage Mechanisms. A variety of gas storage mechanisms may occur in coal
(Clarkson 2018, 2021), including:
• Adsorption onto the internal surface area of organic matter (within micro-/mesoporosity)
• Free-gas storage in matrix organic and inorganic matter porosity (within meso-
/macroporosity)
• Free-gas storage in fractures (microfractures, larger-scale natural fractures)
• Solution gas storage (absorption) in entrained fluids (i.e., water, residual oil)
• Solution gas storage (absorption) in organic matter
A commonly used pore size classification system for coal is the International Union of Pure
and Applied Chemistry (IUPAC) system (Thommes et al. 2015), in which micropores are
274
classified as pores <2 nm in diameter, mesopores are pores between 2 and 50 nm in diameter, and
macropores are pores >50 nm in diameter.
In high-thermal-maturity (rank) coals, the pore structure in the organic matter tends to be
dominated by microporosity (pores <2 nm in size), which results in a large internal surface area,
causing adsorption (mechanism 1) to be the dominant means of storage. However, as noted by
Clarkson (2018), the relative importance of the five storage mechanisms listed above is a function
of coal properties (organic matter content of the coal, its composition and thermal maturity) and
pressure-temperature and hydrodynamic conditions, amongst other factors. Normally, free gas is
negligible compared to adsorbed gas storage and is usually ignored in CBM reservoirs because of
low fracture pore volumes and high water saturations. The exception is for some dry CBM
reservoirs, in which free-gas storage may be more significant (Bustin and Clarkson 1999; Bustin
and Bustin 2009, 2011). Gas in solution in pore fluids is also usually ignored. Absorption of gas
in organic matter could be significant, particularly for CO2, but this is often not distinguished from
the adsorbed gas contribution in adsorption isotherm experiments.
Adsorption isotherms, used to quantify adsorbed gas storage as a function of pressure at a
fixed temperature, are measured in the laboratory using well-known methods (McLennan et al.
1995; Mavor and Nelson 1997; Seidle 2011; Clarkson 2018) in order to quantify adsorption of the
dominant components of the coal gas (CH4 and CO2 are depicted in Fig. 10.20, top left). However,
because coal gas is usually composed of gas mixtures (not just methane), the adsorption of this
mixture must either be measured in the laboratory or predicted using multicomponent adsorption
models (Clarkson 2018). The most popular adsorption model for correlating laboratory adsorption
data in the petroleum engineering literature is the Langmuir model (for single- and
multicomponent adsorption).
While adsorption isotherm measurements and adsorption modeling are sufficient for
quantifying gas content of “saturated” coals, the in-situ gas content of the coal under the conditions
of “undersaturation” is less than that estimated with an adsorption isotherm at the initial reservoir
pressure of the coal (Fig. 10.20, top right). Practically, this means that a coal seam will need to be
dewatered until the pressure reaches the desorption pressure, after which gas will flow along with
water to the well. In order to evaluate the state (saturated vs. undersaturated) in which the coal is
in prior to production, conventionally, a combination of “direct” (adsorption isotherm
measurement) and “indirect” (gas desorption canister testing) methods is used, and the results are
compared to ascertain the saturation level (Seidle 2011; Clarkson 2018).
As noted above, the gas storage mechanisms in coal are a function of coal properties, pressure-
temperature, and hydrodynamic conditions of the coal reservoir. Adsorption in coal is similarly
affected by coal properties and in-situ conditions; for example, adsorption is known to be a
function of the amount/thermal maturity/composition of the organic matter. A goal of a coal
reservoir characterization program, as it pertains to gas-in-place estimation, should be to establish
the primary controls on gas content in the laboratory and then use the resulting correlations to
predict gas content variability in the field. There are, for example, well-established procedures
(e.g., Mavor and Nelson 1997) for estimating coal gas content using bulk density logs (combined
with laboratory data), allowing in-situ gas content to be estimated for noncored intervals in a well.
The unique gas storage properties of coal (e.g., adsorption, saturation level, etc.) need to be
accounted for not only in gas-in-place calculations, but also in material balance calculations, and
in models used for rate-transient analysis (RTA) and reservoir simulation, as discussed in
Section 10.5.4.
275
10.5.2.2 Gas- and Water-Flow Mechanisms. The simple conceptual model shown in Fig.
10.20 (bottom left) is often used to illustrate gas and water flow through coal. It is commonly
assumed that coal is a dual-porosity medium (fractures and rock matrix) with a regular, orthogonal
set of macroscale natural fractures (cleats).
When coals are initially saturated, and the pressure is dropped in the natural fracture system
during initial production, gas and water will flow through the cleats to the well. Transport of gas
and water is commonly modeled using Darcy’s law altered to account for relative permeability
changes caused by dewatering (reduction of water saturation in the cleats). If the coals are initially
undersaturated, dewatering of the fractures (with water flow through the fractures also modeled
using Darcy’s law) must occur in order to reach desorption pressure, after which gas flows from
the matrix to the fractures as in the saturated coal case. It is commonly assumed that only single-
phase flow of gas occurs through the coal matrix, and that the dominant gas transport mechanism
in the matrix is diffusion (Mavor 1996). Various methods have been proposed for handling matrix
transport in simulation, including: equilibrium formulations, which ignore matrix transport and
assume desorption is instantaneous; pseudosteady-state formulations, which assume that there is
no concentration gradient in the matrix, but that the average concentration changes at each
timestep; and unsteady-state formulations, which account for concentration gradients. In several
commercial CBM plays, such as the prolific Fruitland Coal Fairway in the San Juan Basin of
Colorado/New Mexico, the fracture/cleat spacing is typically quite small, and desorption can be
assumed to be instantaneous because of the small matrix block size. However, in lower-
permeability coal reservoirs, where fracture spacing is wider, the latter two formulations may be
more appropriate.
A truly remarkable aspect of coal that can potentially affect long-term production
characteristics of CBM is the dynamic nature of the fracture system (cleats) during production.
While it is well known that fractured rock typically exhibits stress sensitivity (i.e., fracture absolute
permeability decreases as a function of the increase in effective stress during depletion), in some
commercial CBM plays (e.g., the aforementioned Fruitland Coal Fairway, and the Bandana
Formation coals in Australia), fracture absolute permeability may actually increase with depletion.
This latter phenomenon, referred to as “matrix shrinkage” (Fig. 10.21), is caused by desorption of
gas from the matrix blocks, which in turn causes the matrix blocks to shrink volumetrically. This
in turn causes the fracture apertures between matrix blocks to dilate. In some two-phase field cases,
it may therefore be possible for effective permeability to gas to grow with depletion due to a
combination of relative permeability to gas increasing and/or absolute permeability growth due to
matrix shrinkage. An example is shown in Fig. 10.20 (bottom right) from a Fruitland Coal Fairway
well. In this case, the early time (pressures >600 psi) effective permeability growth is likely
dominated by relative permeability increases, whereas late time increases (after the coal has been
dewatered) are attributable to matrix shrinkage (Clarkson 2018). There are several analytical
models now available that can account for the combined effects of compaction, acting to close
fractures, and desorption-dependent permeability (matrix shrinkage), acting to dilate fractures. For
example, the Palmer-Higgs model, which extended the original Palmer-Mansoori model for the
case of vertically cleated coals exhibiting transversely isotropic elastic behavior, is one such model
that has been applied to field data (Moore et al. 2015; Clarkson 2021).
Fractures may also affect long-term production characteristics by causing multilayer behavior
(differential depletion between coal layers) and permeability anisotropy. The effect from
multilayer behavior occurs because fracture density may vary from coal seam to coal seam (as
depicted in Fig. 10.20, bottom left) (Clarkson 2018, 2021). In addition, permeability anisotropy is
276
caused by permeability being different in the directions of the two major fracture systems (the
more continuous “face cleat” and less continuous “butt cleat”).
Matrix
block
Matrix shrinkage due to desorption
Fig. 10.21—Illustration of the concept of matrix shrinkage caused by fracture absolute permeability increases (due to
fracture aperture dilation), modified from Clarkson et al. (2010).
A more thorough discussion of gas storage and gas/water transport mechanisms can be found
in various CBM summary papers, book chapters, and books published in the past decade (e.g.,
Clarkson and Bustin 2011; Seidle 2011; Clarkson 2018, 2021). In these works, laboratory and field
techniques used to evaluate coal properties are also reviewed. As with shale gas and oil reservoirs
discussed in Section 10.3.2, the complexity of CBM reservoirs necessitates that a comprehensive
characterization program (analogous to Table 10.2) be developed to assist with field
development optimization.
Because of the unique storage and transport mechanisms associated with CBM reservoirs,
CBM wells can exhibit unusual production profiles. The production characteristics of a CBM well
exhibiting two-phase flow are illustrated using an example from the Fruitland Coal Fairway
(Fig. 10.22).
Production Rates: Two-Phase Fruitland Coal Well 1

10,000
2 4 1000
Water Rate (STB/D)
Gas Rate (Mscf/D)
3
1,000
5 100
100 10
0 1,000 2,000 3,000
Time (days)
Field gas rate Field water rate
Fig. 10.22—Production profile of a two-phase CBM well (Fruitland Coal Well 1) in the San Juan Basin, Colorado/New Mexico.
The numbered “periods” of well production are discussed in the text. Note that monthly production rates for gas are provided
until ~2,300 days, after which daily production rates are shown. Figure is modified from Clarkson et al. (2007b).
277
During period 1 in Fig. 10.22, the well is dewatering, and gas production increases primarily
due to increasing effective permeability to gas and decreasing flowing pressures (see Fig. 10.26
for flowing pressure profile for this well). Period 1 is referred to as the “negative decline” period,
during which conventional (positive) decline curve methods cannot be applied. After a first peak
in production (period 2), the well enters a normal (positive) decline period (period 3). For this
example, conventional DCA cannot be performed until several months after peak production is
reached (>1,000 days after first production in this example). Due to a combination of well
restimulation, wellhead compression installation/upgrades, and surface infrastructure
optimization, the well reaches a second peak in production (period 4), after which a second normal
decline period is entered (period 5). Clearly, conventional approaches for forecasting wells, such
as DCA, are challenged by such complex production characteristics. For this well, a combination
of material balance analysis and reservoir simulation was required to provide confident reserves
estimates during the first several years of production. Analogous to the discussion in Section 10.4.6
regarding development maturity (early phase–intermediate phase–mature phase), the reserves
assessment methods applied to this single well evolved over its production history from material
balance and model-based approaches earlier in the well life to DCA methods after a clear decline
trend was established (see Section 10.5.4).
In areas outside of the Fruitland Coal Fairway, wells may exhibit very different production
characteristics. For example, the dry coal well in Fig. 10.23 (which exhibits negligible water
production) is located only a few miles from the Fruitland Coal Fairway, where wells have
production characteristics more similar to that shown in Fig. 10.22. This kind of variability in
production characteristics, even in the same basin, makes selection of analogs to support
exploration activities quite challenging. Dry coal wells also exist in the Horseshoe Canyon play in
Alberta, Canada, and they exhibit a more conventional decline profile, analogous to shallow
gas wells.
Gas Production Rates: Single-Phase Fruitland Coal Well
1,000 1,000
Flowing Bottomhole Pressure (psia)
Gas Rate (Mscf/D)
100 500
10 0
0 1,000 2,000 3,000
Time (days)
Field gas rate Field FBHP
Fig. 10.23—Production profile for single-phase (gas) CBM well in the San Juan Basin. This well is located outside the
Fruitland Coal Fairway, where production profiles are similar to that shown in Fig. 10.22. Figure is modified from Clarkson
et al. (2007a), where FBHP is flowing bottomhole pressure.
278
10.5.3 Drilling and Completions. Unlike shale reservoirs, where the primary drilling/completion
method used for exploitation is a horizontal well hydraulically fractured in multiple stages
(multifractured horizontal wells) (Section 10.3.3), CBM reservoirs have been developed with a
wide variety of well architectures, completion methods, and stimulation treatments. The primary
criteria used in selection of the drilling/completion method for coal is system permeability, which
is highly variable from basin-to-basin and field-to-field, and which is a function of drilling depth
and degree of natural fracturing, among other factors. In a comprehensive study by Palmer (2010),
the importance of permeability for CBM completion selection was outlined. He detailed the typical
completion methods used for various ranges of system permeability, including <3 md (“tight coal
completions”), 3–20 md (“low-permeability completions”), 20–100 md (“high-permeability
completions”), and >100 md (“ultrahigh-permeability completions”). Although these permeability
limits for the different completion styles are somewhat subjective, it is useful to use permeability
as a framework for completions decisions. The reader is referred to Palmer’s work for the typical
completion methods in each of these ranges.
10.5.4 Evaluation Methods for Reserves/Resource Estimation. Consistent with the discussion in
Section 10.4.6 for tight/shale reservoirs, methods applied for reserves and resources estimation for
CBM wells will vary according to development phase. In addition, as noted in Section 10.5.2 for
two-phase CBM wells, the evaluation methodology depends on the period observed in the well’s life
cycle (Fig. 10.22). Therefore, in the following, reserves/resources evaluation methods for CBM are
discussed in the context of the phase of exploration/development (early/intermediate/mature) that is
occurring, as well as the period of the well life that is being observed.
10.5.4.1 Early Phase (Development and Well Life). As noted in Section 10.4.6, during this
phase of development, an emphasis is placed on data collection and evaluating uncertainties. As
such, any forecasting performed should be stochastically based to properly represent the
uncertainty. The evaluation methods discussed in this section for this phase apply to exploration
activities in a CBM prospect prior to exploration well completion and production/evaluation, but
after core and log data collection in the exploration wells, and other wells penetrating the targeted
coals, has enabled resource mapping to be performed (coal thickness, density, gas content, gas-in-
place) and the prospect(s) to be defined.
The variability of key CBM reservoir properties from basin-to-basin and even field-to-field
necessitates a stochastic approach to CBM exploration. Failure to reach economic CBM
production is often related to lack of permeability, resulting in subeconomic rates.
Because CBM plays are often characterized as continuous accumulations, it is recommended
that CBM prospects be broken up into cells, the sum of which equals the total resource evaluated,
as recommended by Haskett and Brown (2005), Schmoker and Klett (1999), and Schmoker (2002).
The cell size nominally would be assigned to be a typical well drainage area. Following Haskett
and Brown (2005, p. 3), it is also recommended that CBM resource assessment be “run from
exploration through the pilot phase, appraisal, development and production” and that a “full value
chain” assessment be performed, including multiple decision points.
For CBM exploration and appraisal, a key step is the design of the pilot program (Roadifer
and Moore 2009). Uncertainties associated with production forecasting include relative
permeability, absolute permeability, and the effect of stress and desorption on permeability during
depletion, permeability anisotropy, and multilayer effects (Section 10.5.2). It is for these reasons
that pilot projects are needed particularly for undersaturated coal reservoirs, where interior wells
are bounded by exterior wells to accelerate dewatering. The interior wells need to achieve
279
significant (economic) gas rates, and effective permeability to gas must be established before
reserves bookings can be contemplated. The pilot projects need to be designed to reduce the
uncertainty in key reservoir parameters and to test various completion/drilling technologies to
determine those that are most cost-effective.
An example approach for CBM prospect evaluation in a low-permeability, slightly
undersaturated, two-phase CBM reservoir was provided by Clarkson and McGovern (2005) and is
illustrated in Fig. 10.24. The various components of the modeling approach were summarized in
that work and by Clarkson (2018). After preliminary resource (multiple layers) mapping (Fig.
10.24, top left), “pods” (areas of relatively uniform gas-in-place and apparent reservoir quality)
are selected for further evaluation. A key factor in the assessment of each pod in the prospect is
the stochastic modeling of a “type” well assigned to each pod using a single-well, multilayer
(including coal and noncoal layers, if necessary), analytical simulator coupled to Monte Carlo
simulation. At each Monte Carlo iteration, key uncertain parameters (permeability, degree of
undersaturation, etc.) are selected from input distributions, and type-well gas and water forecasts
are generated (Fig. 10.24, top right). A decision tree is used at each Monte Carlo iteration to
advance the prospect analysis (Fig. 10.24, bottom): At decision node 1, the success or failure of
the exploration well (E well) representing the pod is decided, based on well deliverability (for
undersaturated coal wells, water cumulative production is used as a proxy); at decision node 2, the
success or failure of the pilot project (here, the default pilot project is a five-spot well pattern
designed to accelerate dewatering of a central, bounded well) is dependent on type-well
economics; at decision node 3, the ideal spacing of the development wells is selected based on
desired well interference time. For the latter, at each iteration, the time to reach boundary-
dominated flow for various spacings is calculated; this is compared to the input desired well
interference time, and the optimal spacing to accelerate dewatering is decided upon. An economics
module is used to provide both single development well (D well) economics and full-cycle
(exploration + development) economics. See Clarkson and McGovern (2005) for further
description of the method and an example application.
This method incorporates the components of unconventional reservoir exploration
recommended by Haskett and Brown (2005), including cell-based resource evaluation, use of pilot
projects, generation of a distribution of well forecasts using a model that honors model input
uncertainty (and captures the unique production characteristics of undersaturated/saturated two-
phase CBM wells), and a full value chain assessment. In principle, the same approach can be used
if horizontal wells are required for resource development using analytical modeling approaches
suggested by Clarkson (2018, 2021).
While the above example demonstrates the utility of stochastic CBM well forecasting for
CBM prospect evaluation, stochastic modeling may also be required for predicting well
performance in undeveloped areas, or for early D-well forecasting after some initial production is
available. For example, during the early dewatering phase (Fig. 10.22, period 1), when gas
production is still increasing, the peak production rate and timing and the postpeak decline
characteristics are still unknown; stochastic modeling can be used at this stage to predict a range
of outcomes for the well.
280
TOTAL 160-acre OGIP (MMcf)
14
12
Mean 1 standard deviation
10
Gas Rate (MMscf/D)

2800-3200 8
2400-2800
2000-2400
1600-2000 6
1200-1600
800-1200
4
400-800
0-400
2
E-wells
0
Tim e (years)
32
40
80
12
16
20
24
28
0
00
0
POD1, at each MC iteration:

Success – development
3
Spacing?
Success - pilot 2
D-well economics
Exploration
1
well Cum water production
Failure – no development
Failure – Don’t pilot
Fig. 10.24—Components of the CBM prospect analysis tool developed by Clarkson and McGovern (2005), modified from
Clarkson (2018). See text for description of each component (where OGIP is original gas in place, MC is Monte Carlo
method, and Cum is cumulative).
10.5.4.2 Intermediate Phase (Development and Well Life). As noted in Section 10.4.6, during
this phase of development for tight/shale reservoirs, model-based approaches are most commonly
applied. This is also true for CBM reservoirs, although material balance methods are also
commonly used, depending on the availability of shut-in pressure data.
The example two-phase CBM well in Fig. 10.22 was drilled during the first wave of
development in the Fruitland Coal Fairway before infill drilling, but it was one of several hundred
wells drilled in the field over a short period of time; therefore, it could be classified as an
intermediate or mature development phase well according to the criteria provided in Section
10.4.6. However, it serves to illustrate application of evaluation methodologies that can be applied
during intermediate stages of well life. Again referring to Fig. 10.22, after the early dewatering
period (period 1), material balance and model-based (RTA and numerical simulation) methods can
be used to assess reserves for this and analogous wells. Application of these methods is described
briefly below. Although deterministic methods are illustrated, stochastic approaches can also be
used for RTA and model history matching and forecasting. Importantly, consistency between the
parameters (such as OGIP) obtained from the various methods is sought in order to gain confidence
in the analysis. In order to achieve this consistency, it is recommended to follow the RTA workflow
281
outlined by Clarkson (2013); Clarkson (2021) also includes discussion of development of RTA
methods specifically for CBM wells.
Of the unconventional reservoir types discussed in this chapter, (static) material balance
methods are more commonly used in CBM reservoirs because of the relatively high system
permeabilities of several of the commercial CBM plays in the US, Canada, and Australia. High
permeability is required to achieve estimates of reservoir pressure during well shut-ins in a short
period of time. There have been a number of material balance equations (MBEs) developed
specifically for CBM reservoirs, including those by King (1993), Jensen and Smith (1997), Seidle
(1999), Clarkson and McGovern (2001), Ahmed et al. (2006), Firanda (2011), and Thararoop et
al. (2015), amongst others. All CBM MBEs at a minimum account for gas desorption, but they
vary in their ability to account for free-gas storage, and additional drive mechanisms. Fig. 10.25
provides a demonstration of the application of the MBE presented by Jensen and Smith (1997) to
the two-phase CBM well in Fig. 10.22. A pressure observation well (POW) located near the
producing well (at an equivalent 160-acre spacing) was used to estimate reservoir pressures for
MBE calculations in Fig. 10.25. The producing well was also shut-in periodically throughout its
history, resulting in reservoir pressure estimates very consistent with the POW pressures measured
during the same time frame. A straightline fit to the Jensen and Smith MBE plot, which is a plot
𝑝𝑝
of the material balance pressure function ( , where 𝑝𝑝𝐿𝐿 is Langmuir pressure from the adsorption
𝑝𝑝+𝑝𝑝𝐿𝐿
isotherm, and 𝑝𝑝 is the estimated reservoir pressure) vs. cumulative production, results in an OGIP
estimate of 17.9 bscf. Given an abandonment pressure assumption/calculation, the estimated
ultimate recovery for the well can be calculated.
Static Material Balance Plot: Two-Phase Fruitland Coal Well 1

1.0
Coal Material Balance Pressure Function
0.8
(Dimensionless)
0.6
0.4
0.2
0.0
0.0E+00 3.0E+06 6.0E+06 9.0E+06 1.2E+07 1.5E+07 1.8E+07
Cumulative Gas Production (Mscf)
Fig. 10.25—Use of the Jensen and Smith CBM MBE for estimating OGIP for the two-phase Fruitland Coal Well 1 in
Fig. 10.22, modified from Clarkson (2021). Reservoir pressures corresponding to the data points were estimated from an
offset pressure observation well (POW).
Model-based methods that have been applied to two-phase CBM reservoirs include RTA
(straightline methods, type-well profiles) and analytical/semi-analytical/numerical modeling (for
history matching and forecasting). These models need to account for the unique CBM reservoir
properties described in Section 10.5.2, including desorption, nonstatic fracture permeability, etc.
Application of analytical model history matching/forecasting to Fruitland Coal Well 1 is shown in
282
Fig. 10.26 (top). In order to achieve the history match, gas rates were input, and the well flowing
bottomhole pressures and offset POW pressures were matched; Clarkson (2021) also showed a
match to this well using flowing pressures as input (gas and water rates were matched in that case).
The pressure-dependent effective permeability trend shown in Fig. 10.22 (bottom right) was
required to achieve the match; details of the model matching of this well were provided by
Clarkson (Clarkson et al. 2007b; Clarkson and Bustin 2011; Salmachi et al. 2018; Clarkson 2021).
A forecast was then generated for the well assuming a constant flowing bottomhole pressure. The
modeled gas-in-place estimate is 18.1 Bscf, which is in good agreement with the static MBE
analysis (Fig. 10.25). Fig. 10.26 (bottom) also shows a flowing material balance (FMB) analysis
(a straightline RTA method) of the same well. The version of the FMB equation used in this plot
was modified from Agarwal et al. (1999) by Clarkson (Clarkson et al. 2008; Clarkson 2021) to
account for changes in effective permeability to gas and desorption. The pressure-dependent
effective permeability trend shown in Fig. 10.20 (bottom right) was used in the plot. The OGIP
obtained from the intercept of the plot (~18.1 MMscf) is in good agreement with model history
matching and static material balance analysis.
A second example of the application of model-based methods is provided for the single-phase
Fruitland (non-Fairway) coal well (Fig. 10.23) in Fig. 10.27. No flowing pressure data were
available for this well for the first 21 months of its life; therefore, calculated sandface pressures
from a simulation history match discussed by Clarkson and McGovern (2005) were used in the
analysis. No shut-in pressure data were available for this well, so static material balance analysis
could not be performed. Unlike Fruitland Coal Fairway wells (such as that analyzed in Fig. 10.26),
which typically do not exhibit evidence of transient flow, non-Fairway wells tend to have lower
permeability and may exhibit periods of transient flow. Therefore, prior to undertaking the
analysis, flow-regime identification was performed to determine the flow regimes that were
occurring for this well (Fig. 10.27, top left). For this purpose, a log-log plot of rate-normalized
′
pseudopressure difference (𝑅𝑅𝑅𝑅𝑅𝑅𝑝𝑝𝑝𝑝 ) and its semilog (radial) derivative (𝑅𝑅𝑅𝑅𝑅𝑅𝑝𝑝𝑝𝑝 𝑅𝑅
) vs. desorption-
corrected material balance pseudotime (for a discussion of this flow-regime identification method
for coal, see Clarkson 2021) was used; the appearance of a +1 (unit) slope on the plot suggests that
the well is primarily in the boundary-dominated flow regime. With this interpretation, a flowing
material balance plot (used to analyze boundary-dominated flow) was applied to estimate OGIP
(Fig. 10.27, top right).
Unlike the FMB analysis in Fig. 10.26, a static effective permeability to gas was assumed, but
gas desorption was accounted for in the calculations. For this well, OGIP is estimated to be
approximately 2,960 MMscf from the x-axis intercept of the plot. Type-curve analysis (Fig. 10.27,
bottom left) was performed using the Fetkovich type curves (Fetkovich 1980) and the methods of
Clarkson (2013, 2021) for CBM. Note that material balance pseudotime (corrected for desorption)
is used in the analysis; this causes the depletion data to follow the b = 1 stem because material
balance pseudotime converts the data to an equivalent constant-rate case for liquids (which was
demonstrated by Palacio and Blasingame 1993). Use of real time results in b < 1. Because transient
flow is not evident, the Fetkovich type-curve match cannot be used to obtain a unique estimate of
permeability and skin; however, as pointed out by Clarkson et al. (2007a), the well is not
hydraulically fractured, and therefore large values of dimensionless outer boundary radius (reD)
(5,000 and 10,000) can be selected for the transient stem matching, resulting in permeability and
skin estimates of 9.8 md/−0.2 (reD = 5,000 stems) and 10.7 md/0.5 (reD = 10,000 stems).
283
Model Match and Forecast: Two-Phase Fruitland Coal Well 1
10,000 2,000
Gas Rate (Mscf/D)
Pressure (psia)
1,000 1,000
100 0
0 1,000 2,000 3,000
Time (days)
Field gas rate Model gas rate Field FBHP
Model FBHP POW 1 POW 2
Model reservoir pressure
Flowing Material Balance Plot: Two-Phase Fruitland Coal Well 1

5.0E-02
Normalized Gas Rate [(scf/D)/(psi2/cp)]
4.0E-02
3.0E-02
2.0E-02
1.0E-02
0.0E+00
0 5,000 10,000 15,000 20,000
Normalized Cumulative Production (MMscf)
Fig. 10.26—Use of model-based methods to analyze the two-phase Fruitland Coal Well 1 in Fig. 10.22. (Top) Analytical
model history match of flowing pressures and offset pressure observation well (POW) pressures, and forecast at constant
flowing bottomhole pressure (FBHP). (Bottom) Flowing material balance analysis; the FMB equation was altered to include
changes in effective permeability to gas and desorption, modified from Clarkson (2021).
Finally, an analytical model history match and forecast is performed (Fig. 10.27, bottom right)
using flowing pressure data as input. The OGIP used in the modeling (~3 Bscf) is consistent with
FMB analysis. The permeability and skin obtained from the matching are 10.9 md and −0.3,
respectively, and these values are in reasonable agreement with the range from type-curve analysis.
Using the skin estimate from modeling, permeability can be estimated from the y-axis intercept of
the FMB plot using the procedures of Clarkson (2021); the resulting value (9.5 md) is in reasonable
agreement with model history matching. This example highlights the need to obtain consistent
estimates of properties from all RTA methods to gain confidence in the analysis.
284
Flowing Material Balance Plot: Single-Phase Fruitland Coal
Flow-Regime Identification Plot: Single-Phase Well
Fruitland Coal Well 1.0E-01
1.0E+06
Normalized Gas Rate [(scf/D)/(psi2/cp)]

Gas RNPpg, RNP'pg_R [(psi2/cp)/(Mscf/D)]
8.0E-02
1.0E+05 6.0E-02
4.0E-02
1.0E+04
+unit slope: 2.0E-02
boundary-
dominated flow
0.0E+00
1.0E+03 0 500 1,000 1,500 2,000 2,500 3,000
10 100 1,000 10,000
Material Balance Pseudotime (days) Normalized Cumulative Production (MMscf)
RNP RNP'
Model Match and Forecast: Single-Phase Fruitland Coal Well

Fetkovich Type Curve Match: Single-Phase Fruitland Coal Well
1,000 1,000
10
Flowing Bottomhole Pressure (psia)

Gas Rate (Mscf/D)
1
b=1 100 500
qDd
reD = 5,000
0.1
10 0
b=0 0 1,000 2,000 3,000 4,000 5,000
0.01 Time (days)
0.0001 0.001 0.01 0.1 1 10
tDd Field gas rate Model gas rate Field FBHP
Fig. 10.27—Use of model-based methods to analyze the single-phase CBM well (non-Fairway Fruitland coal) in Fig. 10.23.
(Top left) Flow-regime identification plot. The dominant flow regime is identified to be boundary-dominated flow. RNP is
rate-normalized pseudopressure difference, and RNP’ is its semilog derivative. (Top right) Flowing material balance
analysis. (Bottom left) Fetkovich type-curve analysis, where qDd is dimensionless decline rate, tDd is dimensionless decline
time, reD is dimensionless outer boundary radius, and b is hyperbolic decline exponent. (Bottom right) Analytical model
history match. FBHP is flowing bottomhole pressure. Figure is modified from Clarkson et al. (2007a).
10.5.4.3 Mature Phase (Development and Well Life). As discussed for tight/shale reservoirs
in Section 10.4.6, during this phase of development for tight/shale reservoirs, DCA is the most
commonly applied method, although model-based (RTA) methods may also be applied to select
wells to “calibrate” DCA. In the workflow for the latter concept, after data quality control/filtering
steps and flow-regime identification, it is suggested that the analysis progress from model-based
RTA methods (e.g., straightline, type-curve methods), which are used to derive initial estimates of
fracture/reservoir properties and hydrocarbons in place, to model-based history matching and
forecasting, which use the RTA-derived property estimates as initial input. As a last step, if DCA
methods are used, the empirical DCA model can be fit to the model-based forecast. For example,
if the Arps hyperbolic decline model is used for empirical forecasting, then the Di and b values can
be determined through this calibration, which is particularly important for CBM because several
CBM properties such as desorption, multilayer behavior, etc., can affect the decline model
parameters. Rushing et al. (2008) performed comprehensive analyses of the influences of CBM
properties on decline characteristics.
An example application of this DCA calibration approach is shown in Fig. 10.28 for a single-
phase (gas) Horseshoe Canyon coal well (western Canada). Prior to this analysis, the workflow
described above (flow-regime identification followed by straightline and type-curve analysis to
derive initial permeability, skin, and OGIP estimates) was applied; these steps were illustrated by
285
Clarkson (2018, 2021) for this well and are not repeated here. Using the RTA-derived input as a
starting point, an analytical model history match was performed. Next, the Arps hyperbolic decline
model was matched to the same data as well as the analytical model forecast. Using this approach,
the b value was estimated to be ~0.4, which is within range of that expected for a low-pressure gas
well (Fetkovich et al. 1996). An independent estimate could also be derived from Fetkovich type-
curve analysis by matching the data to the dimensionless (Arps model-derived) depletion stems.
This approach results in a b value estimate of 0.4–0.6. The uncertainty with the latter analysis is
due to the fact that the well is still in early stages of depletion. The b values >0.5 (a general limit
for low-pressure, conventional gas wells) are commonly observed for CBM wells because of the
aforementioned combination of reservoir properties (desorption, etc.).
Model Match and Forecast: Horseshoe Canyon Coal Well

100
Gas Rate (Mscf/D)
10
1
0 2,000 4,000 6,000 8,000 10,000
Time (days)
Field gas rate Model gas rate Arps gas rate
Model Match and Forecast: Horseshoe Canyon Coal Well

100
Gas Rate (Mscf/D)
10
1
100 1,000 10,000
Time (days)
Field gas rate Model gas rate Arps gas rate
Fig. 10.28—Comparison of analytical and empirical (Arps hyperbolic) model forecasts to aid in the calibration of the latter.
The comparison was performed using the data from a single-phase (gas) Horseshoe Canyon coal well. (Top) Semilog plot.
(Bottom) Log-log plot.
286
A similar procedure has been applied to two-phase CBM wells in the Fruitland Coal Fairway.
Two-phase Fruitland Coal Well 2 (Fig. 10.29), like Well 1 (Fig. 10.22), is located in the Fruitland
Coal Fairway, but it was drilled later during more mature stages of development of the field. This
is evident from the fact that some dewatering of the well had occurred prior to initial production
(the well came on production at peak production rate, unlike Well 1, which had to dewater first),
and the initial pressure was several hundred pounds per square inch below initial reservoir
pressures in the field. As with Well 1, an analytical model was used to history match flowing
pressures and offset POW pressures; unlike Well 1, daily flow rates and flowing pressures were
used for this purpose. A pressure-dependent effective permeability trend, analogous to that shown
in Fig. 10.20 (bottom right), was again required to achieve the match. Further, as with Well 1, a
two-phase FMB plot was used to verify the model OGIP (not shown; for additional modeling
details, see Clarkson et al. 2007b, 2010).
Model Match: Two-Phase Fruitland Coal Well 2

10,000 1,000
Gas Rate (Mscf/D)
Pressure (psia)
1,000 500
100 0
0 1,000 2,000 3,000 4,000
Time (days)
Field gas rate Field FBHP Model FBHP
POW Model reservoir pressure
Fig. 10.29—Use of model-based methods to analyze the two-phase Fruitland Coal Well 2. As with Well 1, an analytical
history match of flowing bottomhole pressures (FBHPs) and pressure observation well (POW) pressures was performed
for model calibration.
In order to generate an Arps hyperbolic decline forecast for this well, the Arps model was first
fit to analytical model and numerical model forecasts for wells of the same vintage as Well 1
during the terminal decline period to constrain the range for b values (similar to what was done for
the Horseshoe Canyon well in Fig. 10.28). An Arps hyperbolic decline model was then fit to the
terminal decline period of Well 2 (Fig. 10.30) at the point where flowing pressures became
approximately constant (at ~2,000 days; see Fig. 10.29). The b value selected for the well forecast
was consistent with those obtained from model calibration for earlier-vintage wells.
The above-described empirical model calibration method could be applied to a select number
of wells in the field, representing different reservoir regions, to ensure consistency in empirical
model forecasting for those regions of the field.
287
Arps Forecast: Two-Phase Fruitland Coal Well 2
10,000
Gas Rate (Mscf/D)
1,000
100
1,000 3,000 5,000 7,000 9,000
Time (days)
Field gas rate Arps gas rate
Arps Forecast: Two-Phase Fruitland Coal Well 2

10,000
Gas Rate (Mscf/D)
1,000
100
1,000 10,000
Time (days)
Field gas rate Arps gas rate
Fig. 10.30—Arps hyperbolic decline model forecast for two-phase Fruitland Coal Well 2. (Top) Semilog plot.
(Bottom) Log-log plot.
10.5.5 Additional Exploration and Development Considerations. The unique CBM properties
affect all stages of development planning. The two-phase flow nature of most CBM plays means
that well spacing, well geometry, and well orientation should be designed to accelerate dewatering,
which will, in turn, increase effective permeability to gas, initiate gas production, and reduce the
time to peak gas production. Care must be taken, however, not to overdrill or overdevelop, leading
to pure acceleration with infill drilling. Critical data gathered during the exploration (early) phase,
such as gas contents, isotherm data, pressures (flowing and shut-in), and effective/absolute
permeability data, must continue to be collected during intermediate and sometimes mature stages
288
of development because of the heterogeneity (vertical and lateral) of CBM plays. Collection of
these key data is necessary to inform development and business decisions.
Because CBM is often composed of a mix of gas components, and these components adsorb
to the coal to different degrees (Fig. 10.20, top left), produced gas compositions may evolve during
production. For example, in the Fruitland Coal Fairway of the San Juan Basin, initial gas
compositions consisted of up to 10% CO2 (mole fraction), with the balance being mostly CH4.
During production, the CO2 concentrations in the produced gas increased because CO2, the more
strongly adsorbed component, was released at lower pressures. Because of the potential for
evolving gas compositions during depletion, facilities may be needed to scrub non-hydrocarbon
gases (such as carbon dioxide) to meet market specifications, as is the case in the Fruitland Coal
Fairway play.
Surface operations must also be planned carefully to account for production behavior.
Facilities must be designed to dewater coal wells (artificial lift) and to gather, transport (trucking
or water-gathering system), and treat (subsurface or surface disposal) large amounts of water,
particularly in the early life of a field. Compression must be considered to assist with early
dewatering and to optimize well performance.
Finally, because coal can serve as a source rock for other reservoirs, including noncoal
intervals adjacent to gas-producing coal seams, it is possible that adjacent or interbedded noncoal
intervals could contribute to CBM production through vertical crossflow, even if the noncoal
intervals are not targeted for stimulation. This could be enhanced in particular if the noncoal facies
consist of organic-matter-rich mudstones capable of significant adsorbed gas storage (Bustin and
Bustin 2016). As noted by Bustin and Bustin (2016) in their study of noncoal facies contribution
to total gas in place in the Mannville coal measures of western Canada, this contribution may not
be recognized and is often not considered in resource assessments for CBM plays.
10.5.6 Commerciality Issues. A primary consideration for commerciality is the resource size,
related to the thickness and gas content of the coals. The depth of the coal is an important factor
affecting both gas content (through pressure and temperature) and absolute permeability, which
generally decreases with depth due to the stress sensitivity of the coal fracture apertures. Economic
production of CBM is, generally, limited to depths <4,000 ft for this reason. Factors affecting the
timing of first significant gas production (above the economic limit rate in order to pay out
operating costs)—such as degree of undersaturation—will impact commerciality. In extreme
cases, it could take well over a year for early development wells to achieve significant gas rates.
Commerciality will also be affected by factors controlling time to peak production and peak gas
rate, such as effective permeability to gas, which changes with saturation and reservoir pressure.
In CBM projects, the following factors are important: Infrastructure must be sufficient to
gather and dispose of high initial water volumes; sufficient compression must be installed to
improve CBM recovery and assist with well dewatering; artificial lift should be planned for and
included in operating costs; facilities should be designed to scrub non-hydrocarbon gas from
produced gas to meet market specification (where applicable); and environmental and regulatory
concerns must be addressed.
10.5.7 Classification and Reporting Issues. CBM classification and reporting are aligned with
the principles and practices used for tight/shale oil and gas detailed in Section 10.4.6. Please also
see Barker (2008) for examples specific to CBM.
289
John Etherington, Charles Vanorsdale
10.6 Other Unconventional Oil

10.6.1 Introduction. This chapter has already discussed tight oil and shale oil within the scope of
“unconventional oil.” Arguably, these types of unconventional oil reservoirs garner the most
attention due to their worldwide prevalence and the resource potential they hold. However, these
oils are often low viscosity, medium-to-light gravity crude that facilitates transmissivity (kh/µ)
when stored in low-permeability/tight rock. Crude oil may be divided broadly into categories based
on density and viscosity. For the purposes of this section of Chapter 10, unconventional oil
categories based on density and viscosity criteria may be defined as follows.
• Heavy crude oil is generally defined as having a density in the range of 10 to 23 °API
with a viscosity that is typically greater than 100 but less than 10,000 cp. Although heavy
oil is often recovered in thermal enhanced oil recovery (EOR) projects, it is typically not
a continuous accumulation and often does not require upgrading. Therefore, heavy crude
is defined herein as conventional resources with regard to assessment methods and
classification under the PRMS guidelines. The definition is included here merely to
differentiate “heavy” from “extra-heavy” oil.
• Extra-heavy oil density is less than 10 °API with a viscosity ranging from 1,000 to
10,000 cp. While mobility is limited, accumulations typically have defined oil/water
contacts and exhibit normal buoyancy effects. Extra-heavy oil (EHO) is not necessarily a
“continuous-type” deposit, but it is herein classified as unconventional resources because
it typically requires significant extraction effort (upgrading).
• Natural bitumen typically has a density less than 10 °API and a viscosity greater than
10,000 cp measured at original temperature in the deposit and at atmospheric pressure on
a gas-free basis. In its natural viscous state, it is normally not recoverable at commercial
rates through a well and requires the implementation of improved recovery methods such
as steam injection. Near-surface deposits may be recovered using open-pit mining
methods. Bitumen accumulations are classified as unconventional resources because they
are pervasive throughout a large area and are not currently affected by hydrodynamic
influences such as the buoyancy of petroleum in water. Natural bitumen requires
upgrading to synthetic crude oil or dilution with light hydrocarbons prior to marketing.
• Oil shales are fine-grained sedimentary rocks (shale, siltstone, and marl) containing
relatively large amounts of solid organic matter (known as “kerogen”) from which
significant amounts of shale oil and combustible gas can be extracted by destructive
distillation. All current commercial extraction projects use surface mining techniques.
Similar to extra-heavy oil, oil shale is herein classified as unconventional resources
because it requires upgrading.
A general diagram delineating heavy oil, EHO, and bitumen (tar sands) according to API
gravity and viscosity, along with some comparative examples, is given in Fig. 10.31.
290
Fig. 10.31—General differentiation of oils based on viscosity and API gravity.
10.6.2 Reservoir Characteristics, Risk, and Uncertainty. 10.6.2.1 Extra-Heavy Oil. About 90%
of the world’s known accumulations of EHO are in the Orinoco Oil Belt of the Eastern Venezuelan
Basin, with over 1.3 trillion STB initially in place (Dusseault et al. 2008). A subsequent study
(Attanasi and Meyer 2010) placed the Venezuelan EHO in-place figure at 2.1 trillion STB,
comprising 98% of the total world endowment. It is clear that the extent of the resource is
considerable, but its assessment contains much uncertainty.
Individual sand bodies in the Orinoco accumulations range in thickness up to 150 ft. The
majority of oil-bearing beds are 25 to 40 ft thick, with high porosity (27 to 32%), good permeability
(up to 5 darcies), and good lateral continuity (Dusseault 2001). The major uncertainties are fault
compartmentalization and water encroachment. With the exception of isolated areas within the
belt, the oil is mobile at reservoir conditions.
In the blt, cold production of EHO is normally achieved through multilateral (horizontal) wells
that are positioned in thin but relatively continuous sands, in combination with electric submersible
pumps and progressing cavity pumps. Horizontal multilateral wells maximize the borehole contact
with the reservoir. EHO mobility in the Orinoco Oil Belt reservoirs is typically greater than that
of bitumen in the Alberta sands because of higher reservoir temperatures, greater reservoir
permeability, higher gas/oil ratios, and the lower viscosity of EHO relative to bitumen. The
recovery factor for an EHO cold-production project in the Orinoco Oil Belt is estimated to be
approximately 7 to 15% of the in-place oil (Fiorillo 1987).
While upside secondary recovery with thermal projects is forecast, these incremental volumes
would be classed under the PRMS as Contingent Resources until pilot projects are complete and
thermal projects are sanctioned. Publicly available reserves reported as Proved were substantially
increased in 2010 and 2011 in anticipation of widespread application of thermal recovery, but this
has not materialized.
The majority of Orinoco production is diluted using viscosity reducers, typically naphtha,
either at the wellhead or injected downhole, and transported to the Caribbean coast for upgrading
to remove the naphtha and high sulfur content prior to sale. The resulting “syncrude” then has a
gravity of approximately 32 °API. Consequently, economics must incorporate upgrading costs
either as integrated projects or through reduced pricing at the field-level custody-transfer point.
291
10.6.2.2 Bitumen. Natural bitumen is the portion of petroleum that exists in the semisolid or
solid phase in natural deposits. It usually contains significant sulfur, metals, and other non-
hydrocarbons.
Natural bitumen occurs in at least 598 deposits in 23 countries (Attanasi and Meyer 2010).
The largest known bitumen resource, accounting for approximately 70% of the world total, is in
western Canada, where Cretaceous sands and underlying Devonian carbonates covering a 30,000-
square mile area contain over 1,800 billion bbl of bitumen initially in place (Alberta Energy
Regulator 2020). Current commercial developments are confined to the oil sands. Individual sand
beds in the western Canada oil sands can form thick and continuous reservoirs of up to 250 ft with
a net/gross ratio of over 80%. More often, there are stacked series of 50- to 150-ft-thick sands with
intervening silts and clays. It is common for the sands to have high porosity (30 to 34%) and
permeability (1–5 darcies). The sand grains are often floating in bitumen with minor clay content.
Western Canadian oil sands may contain a mixture of bitumen, EHO, and heavy oil, the properties
of which differ between and within reservoirs.
Two processes are principally used to extract the western Canada bitumen: open-pit surface
mining and various subsurface in-situ recovery methods.
In surface mining, the overburden is removed, and the oil sands are excavated with very large
“truck and shovel” operations. The oil sands are transported to a processing plant, where the ore
is subjected to a series of hot water froth flotation and/or solvent processes to separate the sand
and bitumen. While the process can recover more than 95% of the bitumen in the sand, the
intermixing of clays and the mine-layout requirements reduce the actual recovery factor to about
75%. In fact, the Alberta Energy Regulator prescribes a minimum recovery factor of 90% for the
“as-mined” ore when the bitumen content is greater than 11% by weight (as per the Canadian Oil
and Gas Evaluation Handbook [SPEE 2018], a set of resources evaluation guidelines closely
aligned with the PRMS). Surface mining is typically considered where the depth to the top of the
oil sands is less than 215 ft.
Bitumen that is too deep for surface mining is typically produced using in-situ thermal
recovery processes similar to those used in heavy oil projects. Such projects require a reservoir
depth in excess of 500 ft to provide an impermeable cap to contain the required steam pressure
that provides adequate reservoir energy and temperature. In cyclic steam stimulation operations, a
volume of steam is injected into a well, some period of time (soak time) is allowed to pass, and
then the bitumen, the viscosity of which has been significantly reduced by the high-temperature
steam, is produced from the same well. This process can be repeated ten or more times in the same
well, and the recovery efficiency in these projects is estimated to be 5 to 25% of the original
bitumen in place, depending on the number of pore volumes of steam injected.
Instead of cyclic steam stimulation, most new in-situ projects employ steam-assisted gravity
drainage using a pair of vertically offset horizontal wells. The upper wellbore is used for steam
injection, creating an expanding steam chamber. The thermally mobilized bitumen drains into the
lower wellbore, from which it is produced. A typical project uses well pairs with horizontal lengths
of 2,500 to 3,500 ft, and the injector is placed about 15 ft above the producer. The wells are drilled
in patterns from pads consisting of 5 to 10 well pairs spaced 300 to 500 ft apart. Expected
production rates are 800 to 2,000 BOPD per well. Recovery efficiencies range from 40 to 75% of
oil initially in place (Etherington and McDonald 2004) within the steam chamber; however, at the
reservoir level, the recovery efficiency may only be less than 40% of that range (SPEE 2018).
292
Bitumen, due to its density and immobile character, may require different methods to delineate
deposits and estimate in-place volumes than those used for other conventional oil assessments.
Conventional production decline and material balance calculations do not apply.
For surface mine planning, a closely spaced grid of core holes is required to support a detailed
volumetric assessment. The total cores are analyzed in laboratories to determine the weight percent
of bitumen, which is typically 10 to 14 wt%. The versatility of PRMS to adapt to surface mining
operations is demonstrated in this example. The PRMS Reserve categorization (i.e., P1, P2, P3) is
defined based on confidence in the recovery of product quantities. For surface mining, a key
uncertainty is the presence and bitumen content in the deposit. Thus, confidence is tied to the level
of core hole sampling (e.g., number of holes and distance between them) that describes ore content
in the intended development area. For example, Proved Reserves may require a 1,600-ft grid (61-
acre spacing), while Probable Reserves would be assigned to areas with a 3,200-ft grid (247-acre
spacing). Further, Proved Reserves in the McMurray Formation of Alberta, Canada, are based on
40-acre spacing if 3D seismic data are not available, but 80-acre spacing if they are available;
Probable Reserves would be assigned for 160-acre spacing, with or without 3D seismic data, and
Possible Reserves would be assigned for 320-acre spacing, with or without seismic data (SPEE
2018). Thickness and condition of overburden, and volume allowances on the lease for mine layout
and tailing ponds are examples of key factors affecting mine economics that would likely be
unfamiliar to engineers focused on conventional reservoirs.
The assessment methods for in-situ bitumen-production operations require close well spacing
and core analysis but are supplemented by high-resolution 3D seismic data and complete wireline
log suites. Thermal processes, such as steam-assisted gravity drainage, are sensitive to reservoirs
with associated gas and/or top- or bottomwater zones that may act as potential thief zones. Water
zones rob the steam chamber of energy that would be otherwise available to heat the bitumen and
result in higher operating costs and poorer oil recoveries.
Similar to improved-recovery projects in conventional reservoirs, reserves attribution requires
“a favorable production response from the subject reservoir from either (a) a representative pilot
or (b) an installed portion of the project, where the response provides support for the analysis on
which the project is based” (PRMS § 2.3.4.4). The difference in bitumen projects is that there may
be no preceding “primary” production upon which to base improved recoveries. However, as more
steam-assisted gravity drainage projects have come on stream, the performance results in adjacent
analog reservoirs may be accepted to help underpin the booking of undeveloped reserves.
Under the PRMS, for discovered resources to be classed as Reserves, stakeholding entities
must have committed to an approved development plan, including facilities to produce, process,
and transport the products to established markets. It would be difficult to apply all classical
petroleum reserves criteria such as oil/water contacts and offset-well pressure response to
unconventional deposits like the Canadian oil sands. The appropriate assessment methods may be
a hybrid of those applied to conventional petroleum reservoirs and those applied to mining
deposits. The Canadian Oil and Gas Evaluation Handbook (SPEE 2018) provides technical
guidance in Canada’s petroleum disclosure rules and more detailed best practices for bitumen
reserves and resources assessment and classification.
10.6.2.3. Oil Shale. The organic matter in oil shale is composed chiefly of carbon, hydrogen,
oxygen, and small amounts of sulfur and nitrogen. It forms a complex macromolecular structure
that is insoluble in common organic solvents (compared to bitumen, which is soluble). Because of
its insolubility, the kerogen must be retorted at temperatures of about 500°C to convert it into
synthetic oil and gas. Oil shale differs from coal in that the organic matter in coal has a lower
293
atomic H:C ratio, and the organic matter to mineral matter ratio of coal is much greater. Oil shale
and shale oil are not interchangeable terms; in shale oil, such as in the Bakken, Niobrara, Barnett,
or Eagle Ford plays, the kerogen present in the shale has already been converted to oil and/or gas.
Oil shale in place is estimated at 6.05 trillion barrels in more than 600 known deposits in at
least 33 countries, and all of these figures are considered to be conservative (World Energy Council
2016). By their estimates, over 80% of the recoverable oil shale oil is located in the US, in
particular, in the Green River Formation of the Piceance, Green River, and Uinta Basins.
Oil shales of Estonia are used directly as fuel for power generation and in cement plants. China
and Brazil also have significant oil shale production. Brazil has developed the world’s largest
surface oil shale pyrolysis retort.
Despite very significant research investments in the Colorado Piceance Basin deposits since
the 1970s, there is no current commercial production. Initial pilot projects were based on surface
mining and associated retort facilities. Typical yields were <1 bbl of hydrocarbon liquids per tonne
of shale. Environmental issues include the disposal of large amounts of processed shale with
associated contaminants and the potential contamination of groundwater.
Due to the number of contingencies, currently there are no oil shale Reserves under the PRMS
guidelines.
10.7 References
Acuña, J. A. 2016. Analytical Pressure and Rate Transient Models for Analysis of Complex
Fracture Networks in Tight Reservoirs. Paper presented at the SPE/AAPG/SEG
Unconventional Resources Technology Conference, San Antonio, Texas, USA, 1–3 August.
URTEC-2429710-MS. https://doi.org/10.15530/URTEC-2016-2429710.
Acuña, J. A. 2017. Pressure and Rate Transient Analysis in Fracture Networks in Tight
Reservoirs Using Characteristic Flow Volume. Paper presented at the 5th Unconventional
Resources Technology Conference, Austin, Texas, USA, 24–26 July. URTEC-2667753-MS.
https://doi.org/10.15530/URTEC-2017-2667753.
Acuña, J. A., Wang, S., and Forand, D. 2018. Alternative Production Mechanisms in
Unconventional Reservoirs. Paper presented at the SPE/AAPG/SEG Unconventional
Resources Technology Conference, Houston, Texas, USA, July 2018.
https://doi.org/10.15530/URTEC-2018-2896802.
Agarwal, R. G., Gardner, D. C., Kleinsteiber, S. W., et al. 1999. Analyzing Well Production Data
using Combined-Type-Curve and Decline-Curve Analysis Concepts. SPE Res Eval & Eng 2
(5): 478–486. SPE-57916-PA. https://doi.org/10.2118/57916-PA.
Aguilera, R. 2008. Role of Natural Fractures and Slot Porosity on Tight Gas Sands. Paper
presented at the 2008 SPE Unconventional Resources Conferences, Keystone, Colorado,
10–12 February. SPE-114174-MS. https://doi.org/10.2118/114174-MS.
Aguilera, R. 2014. Flow Units: From Conventional to Tight-Gas to Shale-Gas to Tight-Oil to
Shale-Oil Reservoirs. SPE 165360-PA. SPE Res Eval & Eng 17 (02):190–208. SPE-165360-
PA. https://doi.org/10.2118/165360-PA.
Aguilera, R. ed. 2018. Unconventional Gas and Tight Oil Exploitation. Richardson, Texas, USA:
Society of Petroleum Engineers (SPE).
Aguilera, R. and Harding, T. G. 2007. State-of-the-Art of Tight Gas Sands Characterization and
Production Technology. Paper presented at the Canadian International Petroleum
Conference, Calgary, Alberta, Canada, 12–14 June. PETSOC-2007-208.
https://doi.org/10.2118/2007-208.
294
Aguilera, R. F., Harding, T., Krause, F. et al. 2008. Natural Gas Production from Tight Gas
Formations: A Global Perspective. Paper presented at the 2008 World Petroleum Congress,
Madrid, Spain, 29 June–3 July. WPC-19-2178.
Ahmed, T., Centilmen, A., and Roux, B. 2006. A Generalized Material Balance Equation for
Coalbed Methane Reservoirs. Paper presented at the SPE Annual Technical Conference and
Exhibition, San Antonio, Texas, USA, 24–27 September. SPE-102638-MS.
https://doi.org/10.2118/102638-MS.
Alberta Energy Regulator. 2020. Alberta Energy Outlook 2020. AER ST98-2020. Calgary,
Alberta, Canada: Alberta Energy Regulator. http://www1.aer.ca/st98/2020/data/executive-
summary/ST98-2020-Executive-Summary.pdf. (accessed 28 March 2021).
Anderson, D. M., Liang, P., and Okouma, V. 2012. Probabilistic Forecasting of Unconventional
Resources Using Rate Transient Analysis: Case Studies. Paper presented at the SPE
Americas Unconventional Resources Conference, Pittsburgh, Pennsylvania, 5–7 June. SPE-
Arevalo-Villagran, J. A., Wattenbarger, R. A., and Samaniego-Verduzco, F. 2006. Some Case
Histories of Long-Term Linear Flow in Tight Gas Wells. J Can Pet Technol 45 (3): 31–37.
PETSOC 06-03-01. https://doi.org/10.2118/06-03-01.
Artus, V. and Houzé, O. 2018. A Physical Decline Curve for Fractured Horizontal Wells. Paper
presented at the SPE/AAPG/SEG Unconventional Resources Technology Conference,
Houston, Texas, USA, July 2018. https://doi.org/10.15530/URTEC-2018-2856750.
Atadeger, A., Batur, E., Onur, M.et al. 2020. Comparison of the Methods for Analyzing Rate-
and Pressure-Transient Data from Multistage Hydraulically Fractured Unconventional Gas
Reservoirs. SPE Res Eval & Eng 23 (2): 627–647. SPE-196013-PA.
https://doi.org/10.2118/196013-PA.
Attanasi, E. D. and Meyer, R. F. 2010. Natural Bitumen and Extra-Heavy Oil. In 2010 Survey of
Energy Resources, 123–150. London: World Energy Council.
https://www.worldenergy.org/assets/downloads/ser_2010_report_1.pdf (accessed 28 March
2021).
Ayyalasomayajula, P. S., Silpngarmlers, N., and Kamath, J., 2005. Well Deliverability
Predictions for a Low Permeability Gas Condensate Reservoir. Paper presented at the SPE
Annual Technical Conference and Exhibition, Dallas, Texas, 9–12 October. SPE-95529-
MS. https://doi.org/10.2118/95529-MS.
Barker, G. J., 2008. Application of the PRMS to Coal Seam Gas. Paper presented at the SPE
Asia Pacific Oil & Gas Conference and Exhibition, Perth, Australia, October 2008. SPE-
Bennion, D. B., Thomas, F. B., and Bietz, R. F. 1996. Low Permeability Gas Reservoirs:
Problems, Opportunities and Solutions for Drilling, Completion, Stimulation and
Productions. SPE paper presented at the SPE Gas Technology Conference, Calgary, Alberta,
Canada, 28 April–1 May. SPE-35577-MS. https://doi.org/10.2118/35577-MS.
Berg, B. 2013. Characterizing Shale Plays: The Importance of Recognizing What You Don’t
Know. Presented as a Distinguished Lecture during the 2013–2014 season.
Billingsley, R. L. and Kuuskraa, V. 2006. Multi-Site Application of the Geomechanical
Approach for Natural Fracture Exploration. Award No. DE-RA26-99FT40720, US DOE,
Washington, DC, to Advanced Resources International (March 2006).
295
Bustin, A. A. M. and Bustin, R. M. 2009. Gas in Box: How Much Producible Gas is in the
Horseshoe Canyon. Canadian Society for Unconventional Gas (CSUG), Calgary, Alberta,
Canada (November 2009).
Bustin, A. M. M., and Bustin, R. M., 2016. Contribution of Non-Coal Facies to the Total Gas-in-
Place in Mannville Coal Measures, Central Alberta. Int J Coal Geol 144–145: 69–81.
https://doi.org/10.1016/j.coal.2015.12.002.
Bustin, R. M., and Bustin, A. M. M. 2011. Horseshoe Canyon and Belly River Coal Measures,
South Central Alberta: Part 1, Total Original Gas-in-Place. Bull Can Pet Geol 59 (03): 207–
234. https://doi.org/10.2113/gscpgbull.59.3.207.
Bustin, R. M., and Clarkson, C. R. 1999. Free Gas in Matrix Porosity: A Potentially Substantial
Resource in Low Rank Coals. Paper presented at the Coalbed Methane Symposium,
Tuscaloosa, Alabama, USA, May 1999, 197–214.
Byrnes, A. L., Cluff, R. M., and Webb, J. 2006a. Analysis of Critical Permeability, Capillary
Pressure and Electrical Properties for Mesaverde Tight Gas Sandstones from Western U.S.
Basins. Quarterly Technical Progress Report, Contract No. DE-FC26-05NT42660, US
DOE, Washington, DC (30 June 2006). https://www.kgs.ku.edu/mesaverde/reports/DE-
FC26-05NT42660%20QR063006.pdf (accessed 29 May 2022).
Byrnes, A. L., Cluff, R. M., and Webb, J. 2006b. Analysis of Critical Permeability, Capillary
Pressure and Electrical Properties for Mesaverde Tight Gas Sandstones from Western U.S.
Basins. Quarterly Technical Progress Report, Contract No. DE-FC26-05NT42660, US
DOE, Washington, DC (30 September 2006)
https://www.kgs.ku.edu/mesaverde/reports/DE-FC26-05NT42660%20QR093006.pdf
(accessed 29 May 2022).
Chan, P. B., Etherington, J., and Aguilera, R. 2012. Using the SPE/WPC/AAPG/SPEE/SEG
PRMS to Evaluate Unconventional Resources. SPE Econ & Mgmt 4 (02): 119–127. SPE-
Chu, W., Pandya, N., Flumerfelt, R. W.et al. 2017. Rate-Transient Analysis Based on Power-
Law Behavior for Permian Wells. Paper presented at the SPE Annual Technical Conference
and Exhibition, San Antonio, Texas, USA, October 2017. SPE-187180-MS.
https://doi.org/10.2118/187180-MS.
Clark, A. J., Lake, L. W., and Patzek, T. W. 2011. Production Forecasting with Logistic
Growth Models. Paper presented at the SPE Annual Technical Conference and Exhibition,
Denver, Colorado, USA, 30 October–2 November. SPE-144790-MS.
https://doi.org/10.2118/144790-MS.
Clarkson, C. R., 2013. Production Data Analysis of Unconventional Gas Wells: Workflow. Int J
Coal Geol 109–110: 147–157. https://doi.org/10.1016/j.coal.2012.11.016.
Clarkson, C. R., 2018. Coalbed Methane. In Unconventional Gas and Tight Oil Exploitation, ed.
R. Aguilera. Richardson, Texas, USA: Society of Petroleum Engineers.
Clarkson, C. R., 2021. Unconventional Gas and Light Oil Reservoir Rate-Transient Analysis.
Cambridge, Massachusetts, USA: Gulf Professional Publishing.
Clarkson, C. R., and Bustin, R. M. 2011. Coalbed Methane: Current Field-Based Evaluation
Methods. SPE Res Eval & Eng 14 (01): 60–75. SPE-131791-PA.
https://doi.org/10.2118/31791-PA.
Clarkson, C. R., Bustin, R. M., and Seidle, J. P. 2007a. Production-Data Analysis of Single-
Phase (Gas) Coalbed-Methane Wells. SPE Res Eval & Eng 10 (3): 312–331. SPE-100313-
PA. https://doi.org/10.2118/100313-PA.
296
Clarkson, C. R., Jordan, C. L., Gierhart, R. R. et al. 2007b. Production Data Analysis of CBM
Wells. Paper presented at the Rocky Mountain Oil & Gas Technology Symposium, Denver,
Colorado, USA, 16–18 April. SPE-107705-MS. https://doi.org/10.2118/107705-MS.
Clarkson, C. R., Jordan, C. L., Gierhart, R. R. et al. 2008. Production Data Analysis of Coalbed-
Methane Wells. SPE Res Eval & Eng 11 (2): 311–325. SPE-107705-PA.
https://doi.org/10.2118/107705-PA.
Clarkson, C. R. and McGovern, J. M. 2001. Study of the Potential Impact of Matrix Free Gas
Storage Upon Coalbed Gas Reserves and Production Using a New Material Balance
Equation. Paper 0113 presented at the International Coalbed Methane Symposium,
Tuscaloosa, Alabama, USA, 14–18 May.
Clarkson, C. R., and McGovern, J. M. 2005. Optimization of Coalbed Methane Reservoir
Exploration and Development Strategies Through Integration of Simulation and Economics.
SPE Res Eval & Eng 8 (6): 502–519. SPE-88843-PA. https://doi.org/10.2118/88843-PA.
Clarkson, C. R., Pan, Z., Palmer, I., and Harpalani, S. 2010. Predicting Sorption-Induced Strain
and Permeability Increase with Depletion for Coalbed-Methane Reservoirs. SPE J 15 (1):
152–159. SPE-114778-PA. https://doi.org/10.2118/114778-PA.
Collins, P. W., Fernandez-Badessich, M., and Ilk, D. 2015 Addressing Forecasting Non-
Uniqueness and Uncertainty in Unconventional Reservoir Systems Using Experimental
Design. Paper presented at the SPE Annual Technical Conference and Exhibition, Houston,
Texas, USA, 28–30 September. SPE-175139-MS. https://doi.org/10.2118/175139-MS.
Craig, D. P., Ebenhard, M. J., Odegard, C. E. et al. 2002. Permeability, Pore Pressure, and
Leakoff-Type Distributions in Rocky Mountain Basins. Paper presented at the 2002 SPE
Gas Technology Symposium, Calgary, Alberta, Canada, 30 April–2 May. SPE 75717-MS.
https://doi.org/10.2118/75717-MS.
Duong, A. N. 2011. Rate-Decline Analysis for Fracture-Dominated Shale Reservoirs. SPE Res
Eval & Eng 14 (3): 377–387. SPE-137748-PA. https://doi.org/10.2118/137748-PA.
Duong, A. N. 2014. Rate-Decline Analysis for Fracture-Dominated Shale Reservoirs: Part 2
Paper presented at the SPE Unconventional Resources Conference, Calgary, Alberta,
Canada, 30 September–2 October. SPE-171610-MS. https://doi.org/10.2118/171610-MS.
Dusseault, M. B. 2001. Comparing Venezuelan and Canadian Heavy Oil and Tar Sands. Paper
2001-061 presented at the Petroleum Society’s Canadian International Petroleum
Conference, Calgary, Alberta, Canada, 12–14 June. PETSOC-2001-061.
https://doi.org/10.2118/2001-061.
Dusseault, M. B., Zambrano, A., Barrios, J. R. et al. 2008. Estimating Technically Recoverable
Reserves in the Faja Petrolifera del Orinoco: FPO. Paper WHOC08 2008-437 presented at
World Heavy Oil Congress, Edmonton, Alberta, Canada, 10–12 March.
Etherington, J. R. and McDonald, I. R. 2004. Is Bitumen a Petroleum Reserve? Paper presented
at the 2004 SPE Annual Technical Conference and Exhibition, Houston, Texas, USA, 27–29
September. SPE-90242-MS. https://doi.org/10.2118/90242-MS.
Fetkovich, M. J. 1980. Decline Curve Analysis Using Type Curves. J Pet Technol 32 (6): 1065–
Fetkovich, M. J., Fetkovich, E. J., and Fetkovich, M. D., 1996. Useful Concepts for Decline-
Curve Forecasting, Reserve Estimation, and Analysis. SPE Res Eval & Eng 11 (01): 311–
297
Fiorillo, G., 1987. Exploration and Evaluation of the Orinoco Oil Belt. In Exploration for Heavy
Crude Oil and Natural Bitumen, ed. R. F. Meyer, 103–121. Tulsa, Oklahoma, USA: Studies
in Geology 25, American Association of Petroleum Geologists.
Firanda, E. 2011. The Development of Material Balance Equations for Coalbed Methane
Reservoirs. Paper presented at the SPE Asia Pacific Oil and Gas Conference and Exhibition,
Jakarta, Indonesia, 20–22 September. SPE-145382-MS. https://doi.org/10.2118/145382-MS.
Fulford, D., and Blasingame, T. A. 2013. Evaluation of Time-Rate Performance of Shale Wells
Using the Transient Hyperbolic Relation. Paper presented at the SPE Unconventional
Resources Conference, Calgary, Alberta, Canada, 5–7 November. SPE-167242-MS.
https://doi.org/10.2118/167242-MS.
Hamm, B. and Struyk, E. 2011. Quantifying the Results of Horizontal Multistage Development
in Tight Oil Reservoirs of The Western Canada Sedimentary Basin: Technical and
Economic Case Studies from a Reservoir Evaluator’s Perspective. Paper presented at the
Canadian Unconventional Resources Conference held in Calgary, Alberta, Canada, 15–17
November. SPE-149000-MS. https://doi.org/10.2118/149000-MS.
Haskett, W. J. and Brown, P. J. 2005. Evaluation of Unconventional Resource Plays. Paper
presented at the SPE Annual Technical Conference and Exhibition, Dallas, Texas, USA,
9–12 October. SPE-96879-MS. https://doi.org/10.2118/96879-MS.
Holditch, S. A. 2006. Tight Gas Sands. J Pet Technology 58 (6): 86–93. SPE-103356-MS.
https://doi.org/10.2118/103356-MS.
Ilk, D., Broussard, N. J., and Blasingame, T. A. 2012. Production Analysis in the Eagle Ford
Shale—Best Practices for Diagnostic Interpretations, Analysis, and Modeling. Paper
presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas,
USA, 8–12 October. SPE-160076-MS. https://doi.org/10.2118/160076-MS.
Ilk, D., Perego, A. D., Rushing, J. A. et al. 2008. Exponential vs. Hyperbolic Decline in Tight
Gas Sands—Understanding the Origin and Implications for Reserve Estimates Using
Arps’ Decline Curves. Paper presented at the SPE Annual Technical Conference and
Exhibition, Denver, Colorado, USA, 21–24 September. SPE-116731-MS.
https://doi.org/10.2118/116731-MS.
Jenkins, C. 2009. Estimating Resources and Reserves in Tight Sands: Geological Complexities
and Controversies. American Association of Petroleum Geologists Search and Discovery
Article 120020, adapted from presentation at AAPG Geoscience Technology Workshop,
“Geological Aspects of Estimating Resources and Reserves,” Houston, Texas, USA, 9–11
September.
https://www.searchanddiscovery.com/pdfz/documents/2009/120020jenkins/ndx_jenkins.pdf
.html.
Jenkins, C. 2016a. Chapter 3: Reservoir Characterization. In Estimating Ultimate Recovery of
Developed Wells in Low-Permeability Reservoirs, ed. J. Seidle. Houston, Texas, USA:
Monograph 4, Society of Petroleum Evaluation Engineers (SPEE).
Jenkins, C. 2016b. Chapter 4: Drilling, Completions and Operations. In Estimating Ultimate
Recovery of Developed Wells in Low-Permeability Reservoirs, ed. J. Seidle. Houston,
Texas, USA: Monograph 4, Society of Petroleum Evaluation Engineers (SPEE).
Jenkins, C. and McLane, M. 2019. How Not to Squander Billions on Your Next Unconventional
Venture. Paper presented at the SPE/AAPG/SEG Asia Pacific Unconventional Resources
Technology Conference, Brisbane, Australia, 18–19 November. URTEC-198318-MS.
https://doi.org/10.15530/AP-URTEC-2019-198318.
298
Jensen, D. and Smith, L. K. 1997. A Practical Approach to Coalbed Methane Reserve Prediction
Using a Modified Material Balance Technique. Paper 9765 presented at the International
Coalbed Methane Symposium, Tuscaloosa, Alabama, USA, 12–16 May.
Kazemi, H. 1982. Low-Permeability Gas Sands. J Pet Technology 34 (10): 2229–2232. SPE-
King, G. R. 1993. Material Balance Techniques for Coal Seam and Devonian Shale Gas
Reservoirs. Paper presented at the 1993 Annual Technical Conference and Exhibition, New
Orleans, Louisiana, USA, 23–26 September. SPE-20730-MS.
https://doi.org/10.2118/20730-MS.
Kolodzie, S. 1980. Analysis of Pore Throat Size and Use of the Waxman–Smits Equation to
Determine OOIP in Spindle Field, Colorado. SPE paper presented at the SPE Annual
Technical Conference and Exhibition, Dallas, Texas, USA, 21–24 September 1980. SPE-
Kuuskraa, V. A. and Ammer, J. 2004. Tight Gas Sands Development—How to Dramatically
Improve Recovery Efficiency. Gas Tips (Winter 2004): 15.
Law, B. E. 2002. Basin-Centered Gas Systems. Am Assoc Petr Geol Bull 86: 1891–1919.
Lee, W. J. 1987. Pressure-Transient Test Design in Tight Gas Formations. J Pet Technol 39 (10):
1185–1195. SPE-17088-PA. https://doi.org/10.2118/17088-PA.
Liang, X., Liu, X., Shu, H. et al. 2015. Characterization of Complex Multiscale Natural Fracture
Systems of the Silurian LongMaXi Gas Shale in the Sichuan Basin, China. Paper presented
at the SPE Asia Pacific Unconventional Resources Conference and Exhibition, Brisbane,
Australia, November 2015. SPE-176938-MS. https://doi.org/10.2118/176938-MS.
MacKenzie, W. T. 1975. Petrophysical Study of the Cardium Sand in the Pembina Field. Paper
presented at the 50th Annual Technical Meeting of SPE held in Dallas, Texas, USA, 28
September–1 October. SPE-5541-MS. https://doi.org/10.2118/5541-MS.
Masters, J. A. 1984. Elmworth—Case Study of a Deep Basin Gas Field:. Tulsa, Oklahoma, USA:
AAPG Memoir 38, American Association of Petroleum Geologists.
Mavor, M. J. 1996. Coalbed Methane Reservoir Properties. In A Guide to Coalbed Methane
Reservoir Engineering. Report GRI-94/0397, Gas Research Institute, Chicago, Illinois,
USA.
Mavor, M. J., and Nelson, C. R. 1997. Coalbed Reservoir Gas-In-Place Analysis. Report GRI-
97/0263, Gas Research Institute, Chicago, Illinois, USA.
McLennan, J. D., Schafer, P. S. and Pratt, T. J. 1995. A Guide to Determining Coalbed Gas
Content. Report GRI-94/0396, Gas Research Institute, Chicago, Illinois.
Mishra, S., Kelley, M., and Makwana, M. 2014. Dynamics of Production Decline from Shale
Gas Reservoirs: Mechanistic or Empirical Models. Paper presented at the SPE Eastern
Regional Meeting, Charleston, West Virginia, 21–23 October. SPE-171012-MS.
https://doi.org/10.2118/171012-MS.
Moore, R., Palmer, I., and Higgs, N. 2015. Anisotropic Model for Permeability Change in
Coalbed-Methane Wells. SPE Res Eval & Eng 18 (04): 456–462. SPE-169592-PA.
https://doi.org/10.2118/169592-PA.
National Energy Board of Canada. 2011. Tight Oil Developments in the Western Canada
Sedimentary Basin—Energy Briefing Note. https://www.cer-rec.gc.ca/en/data-
analysis/energy-commodities/crude-oil-petroleum-products/report/tight-oil-developments-
west-canada/tight-oil-developments-in-western-canada-sedimentary-basin-energy-briefing-
note.html (accessed 29 May 2022).
299
Palacio, J. C. and Blasingame, T. A. 1993. Decline Curve Analysis Using Type Curves—
Analysis of Gas Well Production Data. Paper presented at the SPE Joint Rocky Mountain
Regional and Low Permeability Symposium, Denver, Colorado, USA, 26–28 April. SPE-
Palmer, I. 2010. Coalbed Methane: A World View. Int J Coal Geol 35: 184–195.
https://doi.org/10.1016/j.coal.2009.12.010.
Pedersen, S. I., Randen, T., Sonneland, L., and Steen, O. 2002. Automatic 3D Interpretation by
Artificial Ants. Paper G037 presented at the EAGE Annual Conference and Exhibition,
Florence, Italy, 27–30 May.
Raghavan, R. and Chen, C.-C. 2019. Evaluation of Fractured Rocks through Anomalous
Diffusion. Paper presented at the SPE Western Regional Meeting, San Jose, California,
USA, April 2019. SPE-195305-MS. https://doi.org/10.2118/195305-MS.
Rahman, N. M. R., Pooladi Darvish, M., Santo, M. S. et al. 2006. Use of PITA for Estimating
Key Reservoir Parameters. Paper presented at the PS-CIM Canadian International
Petroleum Conference, Calgary, Alberta, Canada, 13–15 June. PETSOC-2006-172.
https://doi.org/10.2118/2006-172.
Roadifer, R. D. and Moore, T. R. 2009. Coalbed Methane Pilots—Timing, Design, and Analysis.
SPE Res Eval & Eng 12 (5): 772–782. SPE-114169-PA. https://doi.org/10.2118/114169-PA.
Rushing, J. A., Perego, A. D., and Blasingame, T. A. 2008. Applicability of the Arps Rate-Time
Relationships for Evaluating Decline Behavior and Ultimate Gas Recovery of Coalbed
Methane Wells. Paper presented at the CIPC/SPE Gas Technology Symposium, Calgary,
Alberta, Canada, 16–19 June. SPE-114514-MS. https://doi.org/10.2118/114514-MS.
Salmachi, A., Clarkson, C., Zhu, S. et al. 2018. Relative Permeability Curve Shapes in Coalbed
Methane Reservoirs. Paper presented at the SPE Asia Pacific Oil and Gas Conference and
Exhibition, Brisbane, Australia, 23–25 October. SPE-192029-MS.
https://doi.org/10.2118/192029-MS.
Salvador, A. 2005. Energy: A Historical Perspective and 21st Century Forecast. Tulsa,
Oklahoma, USA: Studies in Geology 54, American Association of Petroleum Geologists.
Schenk, C. J. and Pollastro, R. M. 2002. Natural Gas Production in the United States. Fact Sheet
FS-113-01, US Geological Survey (January 2002). http://pubs.usgs.gov/fs/fs-0113-01/
(accessed 15 March 2007).
Schmoker, J. W. 2002. Resource-Assessment Perspectives for Unconventional Gas Systems. Am
Assoc Petr Geol Bull 86 (11): 1993–1999.
Schmoker, J. W. 2005. U.S. Geological Survey Assessment Concepts for Continuous Petroleum
Accumulations. In Petroleum Systems and Geologic Assessment of Oil and Gas in the
Southwestern Wyoming Province, Wyoming, Colorado and Utah, Chap. 13, Version 1.
Reston, Virginia, USA: Data Series 69-D, US Geological Survey.
Schmoker, J. W. and Klett, T. R. 1999. U.S. Geological Survey Assessment Model for
Undiscovered Conventional Oil, Gas and NGL Resources—The Seventh Approximation.
Reston, Virginia: Bulletin 2165, US Geological Survey. https://doi.org/10.3133/b2165.
Schopf, J. M. 1956. A Definition of Coal. Econ Geol 51 (1956): 521–527.
Seidle, J. 2011. Fundamentals of Coalbed Methane Reservoir Engineering. Tulsa, Oklahoma,
USA: PennWell Books.
300
Seidle, J. P., 1999. A Modified p/Z Method for Coal Wells. Paper presented at the SPE Rocky
Mountain Regional Meeting, Gillette, Wyoming, USA, 15–18 May. SPE-55605-MS.
https://doi.org/10.2118/55605-MS.
Shahamat, M. S. and Aguilera, R. 2008. Pressure-Transient Test Design in Dual-Porosity Tight
Gas Formations. Paper presented at the SPE/CIPC Gas Technology Symposium, Calgary,
Alberta, Canada, 16–19 June. SPE-115001-MS. https://doi.org/10.2118/115001-MS.
Shanley, K., Cluff, R. M., and Robinson, J. W. 2004. Factors Controlling Prolific Gas Production
from Low-Permeability Sandstone Reservoirs: Implications for Resource Assessment,
Prospect Development and Risk Analysis. Am Assoc Petr Geol Bull 88 (08): 1083–1121.
https://doi.org/10.1306/03250403051.
Society of Petroleum Evaluation Engineers (SPEE). 2010. Guidelines for the Practical
Evaluation of Undeveloped Reserves in Resource Plays. Houston, Texas, USA: Monograph
3, Society of Petroleum Evaluation Engineers.
Society of Petroleum Evaluation Engineers (SPEE). 2016. Estimating Ultimate Recovery of
Developed Wells in Unconventional Reservoirs. Houston, Texas, USA: Monograph 4,
Society of Petroleum Evaluation Engineers (SPEE). 2018. Canadian Oil and Gas Evaluation
Handbook (COGEH), consolidated third edition (online). Calgary, Alberta, Canada: Society
of Petroleum Evaluation Engineers.
Soeder, D. J. and Randolph, P. L. 1987. Porosity, Permeability, and Pore Structure of the Tight
Mesaverde Sandstone, Piceance Basin, Colorado. SPE Form Eval 2 (02): 129–136. SPE-
Solano, N., Zambrano, L., and Aguilera, R. 2011. Cumulative-Gas-Production Distribution on
the Nikanassin Tight Gas Formation, Alberta and British Columbia, Canada. SPE Res Eval
& Eng 14 (03): 357–376. SPE-132923-PA. https://doi.org/10.2118/132923-PA.
Stalgorova, E. and Mattar, L. 2012. Practical Analytical Model to Simulate Production of
Horizontal Wells with Branch Fractures. Paper presented at the SPE Canadian
Unconventional Resources Conference, Calgary, Alberta, Canada, 30 October–1 November.
SPE-162515-MS. https://doi.org/10.2118/162515-MS.
Thararoop, P., Karpyn, Z. T., and Ertekin, T. 2015. Development of a Material Balance Equation
for Coalbed Methane Reservoirs Accounting for the Presence of Water in the Coal Matrix
and Coal Shrinkage and Swelling. J Unconv Oil Gas Res 9: 153–162.
https://doi.org/10.1016/j.juogr.2014.12.002.
Thommes, M., Kaneko, K., Neimark, A. V. et al. 2015. Physisorption of Gases, with Special
Reference to the Evaluation of Surface Area and Pore Size Distribution (IUPAC Technical
Report). Pure Appl Chem 87 (9–10): 1051–1069.
Tight Formation Gas. 1978. Code of Federal Regulations, 18 CFR §270.304
https://www.law.cornell.edu/cfr/text/18/270.304 (accessed 30 May 2022).
Valkó, P. P. 2009. Assigning Value to Stimulation in the Barnett Shale: A Simultaneous
Analysis of 7000 Plus Production Histories and Well Completion Records. Paper
presented at the SPE Hydraulic Fracturing Technology Conference, College Station,
Texas, USA, 19–21 January. SPE-119369-MS. https://doi.org/10.2118/119369-MS.
Wang, Z., Gao, J., Wang, D. et al. 2015. 3D Seismic Attributes for a Tight Gas Sand Reservoir
Characterization of the Eastern Sulige Gas Field, Ordos Basin, China. Geophysics 80(2):
B35–B43. https://doi.org/10.1190/geo2014-0362.1.
301
World Energy Council. 2016. World Energy Resources.
https://www.worldenergy.org/assets/images/imported/2016/10/World-Energy-Resources-
Full-report-2016.10.03.pdf. Downloaded 01 Apr 2021.
Zaitlin, B. A. and Moslow, T. F. 2006. A Review of Deep Basin Gas Reservoirs of the Western
Canada Sedimentary Basin. Mt Geol 43 (3): 257–262.
Zambrano, L., Ramirez Vargas, J. F., Pedersen, P. K. et al. 2016. Integrated Reservoir
Characterization and 3D Modeling of the Monteith Formation: A Case Study of Tight Gas
Sandstones in the Western Canada Sedimentary Basin, Alberta, CanadaSPE Res Eval & Eng
19 (03): 466–480. SPE-171607-PA. https://doi.org/10.2118/171607-PA.
302
Chapter 11
Production Measurement and

Operational Issues
Mohammed Al Alshaikh
11.1 Introduction
Production measurement for reserves and resources quantities is specified under the Petroleum
Resources Management System (PRMS 2018, Appendix A—Glossary) to occur at the “reference
point,” which is the “defined location within a petroleum extraction and processing operation where
quantities of produced products are measured under defined conditions before custody transfer (or
consumption). This is also called point of sale, terminal point, or custody transfer point.”
The reference point then becomes the link among the estimates of subsurface quantities, raw
production quantities, sales quantities, and product price. The PRMS provides a series of
guidelines to promote a consistent approach to measurement in all types of projects.
11.2 Background
The following discussion provides context for application of the PRMS guidelines regarding the
linkage between production measurement and resources estimates in both conventional and
unconventional resource projects. Fig. 11.1 illustrates a simplified oil and gas production operation
with local or lease processing; the historical guidance given by the Society of Petroleum Engineers
on measurement points was built around such a model. The same principles can be applied to other
production models.
Fig. 11.1—Schematic oil and gas operation.
303
A reference point must be clearly defined for each project. In many operations, the reference
point is at the exit valve of the lease separator (point 1 in Fig. 11.1). Where gas plants are involved
as part of an integrated project, the measurement point is typically at the plant outlet (point 2 in
Fig.11.1).
Volumes of oil, gas, and condensate are adjusted to a standard temperature and pressure
defined in government regulation and/or product sales contracts. Liquid production may be
reported in either volume (e.g., barrels of oil with associated density) or mass (e.g., tonnes of oil)
units. Natural gas rates are usually reported in terms of volume (e.g., cubic feet or cubic meters),
but the gas is normally sold according to heating value (e.g., Btu). Products are further specified
by their quality and composition (e.g., sweet light crude, less than X% sulfur, etc.).
There is a wide range of complexity in processing facilities. “Local plants” may range from a
simple dehydration unit to a sulfur-recovery plant to a liquefied natural gas complex or a bitumen
upgrader. The “plant” may be physically located on the producing property or may be a
considerable distance away connected by a pipeline. Table 11.1 lists the levels of processing and
example situations:
Level Description Example

Separation of condensate and natural gas liquids
Volumes undergoing purification and physical
Level 1 (NGLs) and removal of sulfur from sour gas with
separation
subsequent sale of residual dry gas
Volumes requiring more extensive treatment
Upgrading by coking, with inert gas and contaminants
Level 2 where chemical changes are induced, but no
removed
non-reservoir quantities are added
Hydrotreating that adds hydrogen using catalysts to
Volumes undergoing significant chemical change
Level 3 rechain the hydrocarbon molecules, with inert gas and
or where non-reservoir quantities are added
contaminants removed
Table 11.1—Levels of resource processing.
In Level 1 projects, the processing is primarily physical separation, and outlet quantities are
portions of the original reservoir petroleum; thus, resource measurements should be given in terms
of the outlet products (point 2 in Fig. 11.1). If natural gas is sold before extraction of liquids (i.e.,
wet gas), resource estimates are given in terms of that volume. Any further processing beyond this
reference point, including additional liquid recoveries (e.g., in “straddle plants”), are not to be
reflected in resources quantities.
Typically, product sales contracts (or pipeline constraints) set maximum limits on the non-
hydrocarbon “contaminants” content and/or minimum heating value for natural gas deliveries. The
volume sold may include some small fraction of non-hydrocarbons (H2S, CO2, or N2) as long as
that fraction meets sales specifications. Consequently, the resources volumes captured in the
PRMS categories and classifications would be estimated including the same non-hydrocarbon
content as in the sales gas.
Examples of Level 1 processes include gas oil separation plants, oil stabilization plants, and
NGL fractionation plants. All of those plants use processes that do not chemically alter the inlet
hydrocarbon but only change its physical state for further processing and/or transportation.
In the case of liquefied natural gas plants, while significant purification and associated fuel
use are involved, there is no intended chemical alteration of the gas. The process only affects the
physical properties of the gas for transportation. Non-hydrocarbon gases (inert gases and
contaminants) that need to be removed during the process should be accounted for as part of the
gas shrinkage. If condensates or NGLs are extracted during the process and reported, the gas
304
volume should be adjusted accordingly. While the output of the liquefied natural gas plant is
measured in tons, associated reservoir estimates are stated in terms of equivalent purified/shrunk
volume of gas.
Levels 2 and 3 both may be considered upstream manufacturing processes. The actual
custody transfer point in integrated upstream projects depends on the legal structure and contract
terms. Where the same corporate entity shares in both the upstream and downstream operations, it
may be necessary to establish the custody transfer point arbitrarily. Production streams should be
physically measured at the plant inlet, or quantities may be estimated from the outlet products with
accounting for shrinkage (including hydrocarbon consumed in operation as fuel) and additives.
As an example, in bitumen-upgrading operations, whereas the coking process involves
significant shrinkage, the addition of hydrogen results in a volume gain. The synthetic oil delivered
at the plant outlet is the final upstream sales product. Where the custody transfer is deemed to be
at the upgrader inlet, a virtual inlet price may be derived through a net-back calculation.
This technical analysis must be combined with royalty treatment, regulatory guidance, and
accounting to ascertain the logical measurement point for stating resources quantities. In the case
of fully integrated extraction and processing operations, transfer prices should be calculated to
value quantities correctly at the designated measurement point.
The PRMS recommends that petroleum quantities consumed as fuel in production or plant
operations before the reference point, known as lease fuel or consumed in operations (CiO), are
not included as reserves or resources. However, if included as reserves or resources, these
quantities must be stated and recorded separately from the sales (PRMS § 1.1.0.6.A.2, also § 3.2.2).
A further issue is the treatment of the non-hydrocarbons—whether they are contaminants
(with disposal cost and/or no net sales value) or byproducts (e.g., sulfur or helium) that can be sold
to produce additional income. There is general industry agreement that these non-hydrocarbons,
in excess of sales specifications, are not included in resources quantity estimates; however, the
volume should be included as part of the production volume under the PRMS guidelines to
preserve the mass balance. Income generated by the sale of the non-hydrocarbons can be used to
offset expenses incurred to extract and process the associated hydrocarbons (subject to applicable
regulatory guidance) when determining economic producibility for the PRMS classifications.
Accordingly, the cost to dispose of contaminants in excess of sales specifications should also be
considered when determining economic producibility.
Disclosure under some jurisdictions may require separate reporting of heavy oil from
light/medium crude oil. The granularity of reporting by the oil and gas industry is not prescribed here.
11.3 Reference Point

The reference point is a defined location within a petroleum extraction and processing operation
where the produced quantities are measured or assessed (PRMS § 3.2.1.1). A reference point is
typically the point of sale to third parties or where custody is transferred to the entity’s midstream
or downstream operations. Sales production is normally measured and reported in terms of
produced quantities crossing this point over identified time periods.
The reference point may be defined by relevant accounting regulations to ensure that the
reference point is the same for both the measurement of reported sales quantities and for the
accounting treatment of sales revenues. This ensures that sales quantities are stated according to
the delivery specification at a defined price. In integrated projects, the appropriate price at the
reference point may need to be determined using a net-back calculation.
305
Sales quantities are equal to raw production less non-sales quantities (those quantities
produced at the wellhead but not marketable or available for sales at the reference point). Non-
sales quantities include petroleum consumed as lease fuel, flared, or lost in processing plus non-
hydrocarbons that must be removed before sale. Sales quantities may need to be adjusted to
exclude components added in processing but not derived from raw production. Raw production
measurements are necessary and form the basis of many engineering calculations (e.g., material
balance and production performance analysis) based on total reservoir voidage. Substances added
to the production stream for various reasons, such as diluent added to enhance flow properties, are
not to be counted as production, sales quantities, reserves, or resources.
As an example of the way in which the reference point affects the reporting of reserves and
resources, Fig. 11.2 illustrates one such process for a nonassociated gas system. Factors outlined
in Table 11.2 are the expected quantities to be reported depending on the reference point.
NGL
Well Head Separator Gas Plant
Fractionation
A B C
Conditioning
Separator Gas
Slug Catcher &
Separator Ethane
Sales Gas
Well Stream Condensate
Propane
Fluid
Stabilizer NGLs
Gas Treatment Butane
Gas
Condensate
Sweetening Pentane +
Dehydration Non-
Hydrocarbon
Sulfur Recovery Acid Gas
NGL Recovery
Separator
Condensate
Fig. 11.2—Reference points in a typical gas operation.
Reference Point Reported Quantities

Point A Wet gas + condensate
Natural gas + natural gas liquids (NGLs) + condensate – (lease fuel + flared
Point B
quantities + system losses)
Natural gas + ethane + propane + butane + pentane + condensate – (lease fuel +

Point C
flared quantities + system losses)
Table 11.2—Reference points and typical reporting requirements (CiO not reported as reserves).
11.4 Consumed in Operations (CiO)

Lease fuel, also known as fuel “consumed in operations” or CiO, refers to that portion of produced
petroleum consumed as fuel for power generation or other field/plant operations before the
reference point.
When gas (or crude oil) consumed in lease operations is treated as shrinkage and excluded
from sales quantities, it should not be included in reserves and resources estimates under the
PRMS. Many international and national companies do, however, report CiO as part of their
306
“marketable gas” reserves. Although reserves are recommended (PRMS §1.1.0.6.A.2) to be sales
quantities, the CiO quantities may be included as reserves or resources but must be stated and
recorded separately from the sales portion. Entitlement rights for the fuel usage must be established
before recognizing CiO as reserves. Flared gas and oil and other petroleum losses must not be
included in either product sales or reserves, but once produced, they are included in produced
quantities to account for total reservoir voidage. Third-party gas obtained under a long-term
purchase, supply, or similar agreement is excluded from reserves.
The CiO quantities must not be included in the project economics because there is neither a
cost incurred for purchase nor a revenue stream to recognize a sales quantity. The CiO fuel replaces
the requirement to purchase fuel from external parties and results in lower operating cost. All
actual costs for facilities-related equipment, the costs of the operations, and any purchased fuel
must be included as an operating expense in the project economics.
11.5 Associated Non-Hydrocarbon Components

In the event that non-hydrocarbon components are associated with production, the quantities
reported as reserves or resources should honor is the values measured at the reference point,
provided those quantities adhere to specifications in sales contracts and/or by regulatory
authorities. Hence, if gas (as produced) includes a proportion of CO2, then the pipeline may accept
sales gas with a limited CO2 content.
Reported Quantities (when CiO is excluded)

= Raw Quantities − Extracted Nonhydrocarbons − CiO − Flared Gas
− System Losses.
For example, if produced gas has 4 mol% H2S, and the development plan is to utilize an existing
pipeline that will only accept 2 mol% H2S, then the following factors need to be considered:
• Additional facilities are required to extract the additional H2S to meet the pipeline
requirement.
• The cost of the additional facilities needs to be added to the total cost of the project.
• The additional gas consumed in operations required for the additional processing needs
to be excluded from total reserves or resources.
• If the additional cost takes the project from commercial to sub-commercial, then
alternatives need to be considered.
In this situation, the sales gas volume would include 2% H2S, and natural gas reserves
dedicated to that pipeline would be estimated including 2% H2S. For high concentrations of H2S
(concentrations as high as 90% have been known), the H2S gas may be separated and converted to
sulfur, which can then be sold. In such cases, the natural gas reserves exclude the H2S volumes,
and the sulfur volume may be quoted separately. At times, prices for sulfur can be low, and
stockpiling for future sale is not uncommon.
Even if the non-hydrocarbon components are removed before the reference point, accurate
records of the raw quantities (voidage extraction quantities) are required for reporting production
volumes and engineering calculations.
When the non-hydrocarbon components extracted generate revenue (e.g., sulfur and helium),
they still cannot be classified as hydrocarbon reserves or resources under the PRMS but must be
reported separately. Revenues realized from the sales can be considered in the economics of the
307
project, which, in return, can result in additional hydrocarbon reserves as a result of a lower
economic limit.
11.6 Natural Gas Reinjection

Gas can be reinjected into the same or different reservoirs for many reasons, including gas cycling,
pressure maintenance, miscible injection, or other enhanced oil recovery processes. Classification
of the injected gas depends on the reference point.
If the gas was produced and reinjected before the reporting reference point (i.e., no transfer of
ownership), then the reinjected gas can be reported as reserves if it meets all the reserves criteria.
The reinjected gas needs to be technically and commercially mature, as well as estimated to be
eventually recoverable (marketable, with a distribution/export system available nearby). In the
case of miscible injection or other enhanced recovery processes, due allowance needs to be made
for any gas not available for eventual recovery as a result of losses associated with the
inefficiencies inherent in the corresponding process. Normally, these volumes are not included in
any PRMS reserves category. In some cases, the objective of gas injection in a reservoir can be
efficient disposal of the gas; consequently, no gas volume should be allocated to reserves.
If the gas injected was produced after the transfer of ownership (perhaps purchased or
transferred from another field or reservoir), then the injected gas can be classified as inventory (for
the reservoir into which it was injected) but not as reserves or resources. When produced, the gas
would not contribute toward field production or sales. If the injected gas was produced from
another field or reservoir (perhaps in which the recipient was the same producer) after the reference
point, then the produced gas would have been part of the reserves from the donor reservoir, but
the recipient reservoir could only consider the gas as part of inventory but not reserves. Otherwise,
this would constitute “double-booking” of reserves. When produced, the gas would not contribute
toward field production or sales.
Typically, under such circumstances, the field would then contain gas that is part of the
original in-place volumes as well as injected gas held in inventory. On commencing gas production
from the field, the last-in/first-out principle is recommended; hence, the inventory gas would be
produced first and not counted toward field production. Once the inventory gas had been re-
produced, further gas production would be drawn against the reserves and recorded as production.
The above methodology ensures that the uncertainty with respect to the original field volumes
remains with the gas reserves and not the inventory. An exception to this could occur if the gas is
acquired through a production payment. In this situation, the volumes acquired could be considered
as reserves.
11.7 Underground Natural Gas Storage

One case of gas reinjection is gas storage, where gas is injected into a reservoir to be recovered later,
perhaps to meet a higher market demand. Quantities reinjected should not be classified as reserves
if the gas was transferred from one lease or field to another with sale or custody transfer between the
points. In other words, if the gas passed the reference point pre-injection, then it cannot be classified
as reserves but rather as inventory. At this point, produced quantities are accounted for as production
from the original reservoir and then as inventory after accounting for system losses.
Commencement of gas production from the field will be handled in the same manner as
described for gas reinjection above (Section 11.6).
308
There may be occasions in which gas is transferred from one lease or field to another without
a sale or custody transfer occurring. In such cases, the reinjected gas could be included with the
native reservoir gas as reserves.
11.8 Production Balancing

11.8.1 Production Imbalances (Overlift/Underlift). Production overlift or underlift can occur in
annual records because of the necessity for companies to lift their entitlement in parcel sizes to
suit the available shipping schedules as agreed among the parties. At any given financial yearend,
a company will be in an overlift or an underlift situation. Based on the production matching of the
company’s accounts, production should be reported in accord with and equal to the liftings actually
made by the company during the year, and not on the production entitlement for the year.
For companies with small equity interests, where liftings occur at infrequent intervals (perhaps
greater than 1 year), the option remains to record production as entitlement on an accrual basis.
11.8.2 Gas Balancing. In gas production operations involving multiple working interest owners, an
imbalance in gas deliveries can occur that must be accounted for. Such imbalances result from the
owners having different operating or marketing arrangements that prevent the gas volumes sold from
being equal to the ownership share. One or more parties then become over/underproduced. For
example, one owner may be selling gas to a different purchaser from the others and may be waiting
on a gas contract or pipeline installation. That owner will become underproduced, while the other
owners sell their gas and become overproduced. These imbalances must be monitored over time and
eventually balanced in accordance with accepted accounting procedures.
Some points to consider in gas-balancing arrangements:
• In gas swaps, early production from one field may be traded with later production from
another field.
• Take-or-pay gas means that the production has to be paid for even if it is not “taken” (i.e.,
produced).
There are two methods of recording revenue to the owners’ accounts. The “entitlement” basis
of accounting credits each owner with a working interest share of the total production rather than
the actual sales. An account is maintained of the revenue due to the owner from the overproduced
owners. The “sales” basis of accounting credits each owner with actual gas sales, and an account
is maintained to record the over- and underproduced volumes (relative to the actual ownership).
The production volumes recorded by the owners will be different in the two cases. The reserves
estimator also must consider the method of accounting used, the current imbalances, and the
manner of balancing the accounts.
11.9 Shared Processing Facilities

It is not uncommon in gas production operations for several fields to be grouped to supply gas to
a central processing facility (gas plant) to remove non-hydrocarbons and recover liquids. Where a
company has an equity interest in one or more of the contributing gas fields and also in the
processing facility, the allocation of dry gas and NGLs back to the fields (and reservoirs) for
estimation of reserves can be complex. While not addressed specifically in the PRMS, the basic
principle that reserves estimates must be linked to sales products applies. Thus, by measuring the
volumes and components of the gas stream leaving each lease and the equity share in the lease, the
company can calculate its share of the sales products for purposes of reserves. This share is not
affected by the company’s actual equity interest in the gas plant as long as it is greater than zero.
309
If the company has no equity interest in the facility, then it is treated as a straddle plant, and
reserves are estimated in terms of the wet gas and the non-hydrocarbon content accepted at the
lease outlet. The allocation of revenues is subject to the contractual agreement among the lease
and plant owners.
When the plant ownership and lease working interest are different, booking may be an issue.
This can be highly complex, but some general points are captured in the following:
1. If the plant is associated with unit production and is unit owned (see Chapter 12—
Resources Entitlement and Recognition for discussion of unit agreements), book residual
plus liquids.
2. If the plant is 100% owned by the company sending produced volumes to the facility,
then that company books the volumes processed by the plant as residual plus liquids.
3. If the contract directly stipulates the retention, by the producer, of products through plant
processing, then the volumes are booked according to contract.
4. If plant ownership and lease ownership interests are different, and existing contracts do
not conclusively specify product allocation, then the issues may be complex. In this case,
where the trail is not clear, the booking of wet gas is recommended. The asset team
responsible for handling the produced stream is afforded, however, the opportunity to
present information that describes a specific instance in which the booking of residual
plus liquids is reasonable and adheres to applicable contract terms. Where processed
volumes are significant, this reconciliation is required.
11.10 Hydrocarbon Equivalence Issues

11.10.1 Gas Conversion to Oil Equivalent. Gas quantities often are converted to barrels of oil
equivalent (BOE) to facilitate like-for-like comparison as the term allows for consolidation of
multiple product types into a single equivalent type. That said, the conversion is usually performed
based on the gross heating content or calorific value of the gas compared to the heating value of a
reference barrel of oil.
It is important to note that, when several projects are being evaluated, the gas volumes first
must be converted to the same temperature and pressure. It is customary to convert to standard
conditions of temperature and pressure (STP) associated with the system of units being used.
One way of calculating the gross heating value (GHV) of a given gas relies on the use of a
representative gas composition. As there are industry-recognized standards for reporting GHV for
each component (j) at STP, the GHV will be the aggregate of the mole fraction of a component
multiplied by the GHV of that component.
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐺𝐺𝐺𝐺𝐺𝐺 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉 = ∑𝑗𝑗(𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑗𝑗 × 𝐺𝐺𝐺𝐺𝐺𝐺𝑗𝑗 ). ........................... (11.1)
The calorific value obtained using these formulas can be cross-checked by taking actual
calorific value measurements of some gas samples from the sales point.
To further simplify the process, some rules of thumb are used in the industry in the absence
of a GHV computation. One very common rule of thumb is 1 barrel of oil equivalent (BOE) equals
5.8 thousand standard cubic feet (Mscf) at standard conditions (15°C and 1 atm). (The conversion
is sometimes rounded to 6.0 Mscf per BOE for ease of calculation.) This rule of thumb assumes
that the gas has a GHV of 1,000 Btu/scf, and the oil has a GHV of 5.800 million Btu/bbl (based
on a 35 °API oil; see Fig. 11.3). The 5.8 Mscf/BOE conversion is recommended when the gas is
dry at the point of sale.
310
Another rule of thumb that is common when using the metric system is 5.62 Mscf/BOE. It is
based on 1,000 standard m3 of gas per 1 standard m3 of oil. A useful formula for changing calorific
value from Imperial (customary) to metric units at STP (15°C and 1 atm) is 1 MJ/m3 = 1 Btu/scf
× 35.3 scf/m3 × 1 kJ/0.948 Btu × 1 MJ/1000 kJ.
The calorific value obtained by the process described above can be used for estimating BOE
with a more customized approach for an entity’s portfolio by taking into consideration the crude oil
characteristics of each reservoir and/or where the gas has a calorific value other than 1,000 Btu/scf.
This will enhance the reporting of gas in terms of oil equivalent, but it is more data intensive.
If calorific values of gas volumes are not available at the gas reference point, then multistage
pressure-volume-temperature (PVT) experimental data on the gas-liberation process at different
conditions mimicking field-gathering facilities may be used. This is done by calculating the
calorific value of the total gas output from all separation stages based on the composition. The
process is outlined as follows:
• Calculate the gross calorific value (CV) for each stage using the gas composition at that
stage.
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝐶𝐶𝐶𝐶 = ∑ 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝑀𝑀𝑀𝑀𝑀𝑀% × 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝐶𝐶𝐶𝐶. ..................................(11.2)
• Calculate the weighted average calorific value using the gas/oil ratio (GOR) as a
weighting factor.
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐺𝐺𝐺𝐺𝐺𝐺 = ∑𝑛𝑛𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆=1 𝐺𝐺𝐺𝐺𝐺𝐺𝑛𝑛 . ......................................................................................(11.3)
∑𝑛𝑛
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆=1 𝐶𝐶𝐶𝐶𝑛𝑛 × 𝐺𝐺𝐺𝐺𝐺𝐺𝑛𝑛
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐺𝐺𝑎𝑎𝑠𝑠 𝐶𝐶𝐶𝐶 = . ...............................................................................(11.4)
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐺𝐺𝐺𝐺𝐺𝐺
• The calorific value obtained using these formulas can be cross-checked by taking actual
calorific value measurements of some gas samples from the sales point.
11.10.2 Liquid Conversion to Oil Equivalent. Regulatory reporting usually stipulates that liquid
and gas hydrocarbon reserves volumes must be reported separately, with liquids being the sum of
the crude oil, condensate, and NGLs. For internal company reporting purposes and often for
intercompany analysis, the combined volumes for crude oil, condensate, NGL, and gas as an oil-
equivalent value offer a convenient method for comparison.
The arithmetic sum of crude oil, condensate, and NGL reserves volumes can provide an oil-
equivalent volume when one product dominates, and the other streams are not material in
comparison. This method will not be satisfactory for certain regulatory disclosures because the
products have different market values. A more correct, but imperfect, method in terms of value
involves taking into account the different densities of the fluids. Further improvement in
combining crude oil, condensate, and NGL can be achieved by considering the heating equivalent
of the three fluids and combining them accordingly.
The correlation between the heat content (Btu/bbl) of crudes, condensates, fuel oils, and
paraffins is illustrated in Fig. 11.3.
311
Fig. 11.3—Heat content (Btu) of crudes, condensates, fuel oils, and paraffins. *
11.11 Example
A reservoir with 100 million STB of 35 °API oil reserves has the multistage pressure-volume-
temperature experimental data listed in Table 11.3. The process to convert the reserves to oil
equivalent is as follows:
Stage 1 Stage 2 Calorific Value

Hydrocarbon constituents (mol%) (mol%) (Btu/ft3)
N2 2.00 0.43 0
CO2 1.41 1.71 0
H2S 1.33 0.28 0
C1 78.95 51.85 1010
C2 10.79 23.41 1769.7
C3 3.57 13.41 2516.1
i-C4 0.48 2.18 3251.9
n-C4 0.92 4.33 3262.3
i-C5 0.21 1.01 4000.9
n-C5 0.2 0.91 4008.7
C6 0.14 0.48 4755.9
C7+ 0.00 0.00
Total 100% 100%
Gas/oil ratio, scf/STB 750 150
Table 11.3—Example multistage separation pressure-volume-temperature composition.
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝐶𝐶𝐶𝐶 = ∑ 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝑀𝑀𝑀𝑀𝑀𝑀% × 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝐶𝐶𝐶𝐶.
Stage 1 𝐶𝐶𝐶𝐶 = 1,147 Btu/ft 3 .
Stage 2 𝐶𝐶𝐶𝐶 = 1,587 Btu/ft 3 .
*
Personal communication from Cronquist in McMichael and Spencer (2001).
312
∑𝑛𝑛𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆=1 𝐶𝐶𝐶𝐶𝑛𝑛 × 𝐺𝐺𝐺𝐺𝐺𝐺𝑛𝑛 (1,147 × 750) + (1,587 × 150)
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐺𝐺𝐺𝐺𝐺𝐺 𝐶𝐶𝐶𝐶 = = = 1,220 Btu/ft 3 .
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐺𝐺𝐺𝐺𝐺𝐺 (750 + 150)
scf
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐺𝐺𝐺𝐺𝐺𝐺 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 = 𝑂𝑂𝑂𝑂𝑂𝑂 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 × 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐺𝐺𝐺𝐺𝐺𝐺 = 100 million STB × 900
STB
= 90,000 million scf = 90 Bscf.
Using the caloric value of 35 °API oil based on Fig. 11.3 of 5,800 million Btu/bbl, now it is
possible to convert the total gas reserves to barrels of oil equivalent (BOE).
90 Bscf × 1,220 Btu/ft 3

𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐺𝐺𝐺𝐺𝐺𝐺 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 (𝐵𝐵𝐵𝐵𝐵𝐵) = = 19 million BOE.
5,800 million Btu/bbl
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 (𝐵𝐵𝐵𝐵𝐵𝐵) = 𝑂𝑂𝑂𝑂𝑂𝑂 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 + 𝐺𝐺𝐺𝐺𝐺𝐺 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅

= 100 million bbl + 19 million BOE = 119 million BOE.
Much of this chapter builds on the work of Satinder Purewal in the 2011 version of the Guidelines
for Application of the PRMS. Helpful insight also was provided by Monica Clapauch Motta.
11.13 References
McMichael, C. L. and Spencer, A. 2001. Operational Issues. In Guidelines for the Evaluation of
Petroleum Reserves and Resources, Chap. 3, 25–34. Richardson, Texas, USA: Society of
Petroleum Engineers.
Petroleum Resources Management System (PRMS), Version 1.01. 2018. Richardson, Texas, USA:
313
Chapter 12
Resources Entitlement
and Recognition
Monica Clapauch Motta (Chair)
Elliott Young and Regnald A. (Reggie) Boles
12.1 Foreword
This chapter is about resources entitlement, which refers to the ability of an entity to claim quantities
(net to the entity) of reserves and resources in its reports. Entitlement is also referred to as
recognition, as the entitled quantities may be recognized in the entity’s reserves and resources
reports. The term “recognized quantities” is frequently referred to as the “booked quantities.” This
chapter text is an update to Chapter 10 of the Guidelines for Application of the Petroleum Resources
Management System published in 2011. Drawing heavily on the prior text, the content has been
updated to reflect refinements in generally accepted industry practices commonly used when
determining entitlement to production and recognizable quantities of reserves and resources under a
range of agreement types and fiscal terms, as well as aligning with the Petroleum Resources
Management System (PRMS 2018). It is not the intent of the Society of Petroleum Engineers, or the
cosponsors of the PRMS, to comment on the individual disclosure regulations promulgated by
specific government agencies regarding entitlement to production or the ability to report reserves.
As a consequence, emphasis has been placed on principles for reserves and resources recognition
under the PRMS and determination of net quantities, rather than specific government regulations,
financial reporting guidelines, or the classification of Reserves, Contingent Resources, and
Prospective Resources into the various certainty categories of the PRMS.
12.2 Introduction
The ability to discover, develop, and commercially produce hydrocarbons in accordance with all
environmental and safety regulations is the primary goal of the upstream petroleum industry.
Aggressive competition, government policies, ever-sharpening scrutiny by the investment
community, and volatility in product prices drive companies to search for attractive new exploration
and producing venture opportunities that will add the greatest value for a given investment. As a
consequence, contracts and agreements for these opportunities are becoming increasingly complex,
further increasing the focus on the ability to recognize reserves and resources.
Production-sharing and other nontraditional agreements have become popular because of the
flexibility they provide host countries in tailoring fiscal terms to fit their sovereign needs while
enabling contracting companies to recover their costs and achieve a desired rate of return. Actual
agreement terms, including those that relate to royalties or some royalty payments, cost recovery,
profit sharing, and taxes, may have a significant influence on the ability of an entity to recognize
and report hydrocarbon reserves and resources. The terms “reserves” and “resources,” as used in
this chapter, are defined in the context of the PRMS. This chapter focuses on reserves and resources
314
entitlement (quantities an entity may recognize and report as its reserves and resources) under the
more common fiscal systems being used throughout the industry. Various types of production-
sharing, service, and other types of common contracts are reviewed to illustrate their impact on
recognition and reporting of oil and gas reserves and resources in the context of the PRMS
framework. This chapter also introduces other agreements and concerns that can impact a given
entity’s reserves and resources entitlement, in the context of the PRMS, such as conveyances and
unitization agreements.
Oil and gas reserves and resources are the fundamental assets of producing companies and
host countries alike. They have been literally the fuel that drives economic growth and prosperity.
When produced and sold, they provide cash flows and the crucial funding for future exploration
and development projects. With the increased focus of the investment community on reserves and
resources inventories and the value of externally reported, project-related reserves that are added
each year, many companies are reluctant to undertake a project that does not provide the
opportunity to eventually report reserves.
12.3 Regulations, Standards, and Definitions

When reporting reserves and resources, it is important to distinguish between the specific
regulations governing their external reporting (such as financial reporting to a stock exchange or
to a host country regulatory agency), and internal company use for technical and business-planning
purposes. In any assessment, the basis used, assumptions, and purpose for which reserves and
contingent resources are recognized and reported must be defined. Fig. 12.1 summarizes the
PRMS reserves and discovered resources categories relative to categories that many government
regulatory agencies and/or stock exchanges allow in required disclosures.
Reporting to National Regulators
PRMS Class
and
Categories
Generally Proved Probable Possible
Allowed in Reserves Reserves Reserves
Regulatory
Reporting
C1 C2 C3
Resources Resources Resources
Fig. 12.1—PRMS (2018) discovered categories generally allowed in required disclosures.
Numerous national regulatory bodies have developed regulations and standards for reporting
oil and gas reserves (and, in some cases, resources) within their respective countries. These standards
provide detailed descriptions of the categories of reserves and resources to be reported, required
supporting information, and the format to be used for the disclosures. Those reports will typically
require information on the complete (100%) accumulation (e.g., petroleum initially in place,
reserves, resources), as national regulatory bodies are interested to know the resources of the country.
National regulatory standards and reporting rules do not generally provide guidance on the type or
315
extent of rights to the underlying resource or production that is required for reporting entitled
reserves and resources. Stock exchanges will typically require that companies report their entitled
quantities and related information in financial reporting. A company’s internal planning process will
also typically require the entitled quantities. For some unique types of agreements, it may not be
clear whether a company is even entitled to report the related reserves and resources. This is
particularly the case with agreements in which production ownership and control reside, by law, with
the host country rather than with the contracting party. Analysis of the key elements and fiscal terms
of these contracts and comparison to those in more widespread use are good approaches to determine
whether reserves and resources can be recognized and subsequently reported.
The PRMS acknowledges the concept of an economic interest as the basis for recognizing and
reporting reserves and resources. To determine when an economic interest exists, many have
referred to the US Securities and Exchange Commission Regulation S-X, Rule 4-10b, “Successful
Efforts Method” (US Securities and Exchange Commission 2011) and Financial Accounting
Standards Board (FASB) Accounting Standards Codification® (ASC) Topic 932, “Extractive
Activities—Oil and Gas” (FASB 2010).
12.4 Reserves and Resources Recognition Under the PRMS

This section describes the basic elements used for determining reserves and resources entitlement.
12.4.1 Reserves and Resources Entitlement Elements. US Securities and Exchange

Commission Regulation S-X (US Securities and Exchange Commission 2017) and FASB ASC
Topic 932 (FASB 2010) are points of reference for entitlement elements, as shown in Fig. 12.2.
These references outline fundamental principles of resources entitlement and provide a useful
framework and set of criteria for establishing when an interest in a property exists, and guidance
on when reserves and resources can be recognized under the PRMS and government regulations.
Fig. 12.2—Text from FASB ASC Topic 932 “Extractive Activities—Oil and Gas.”
These references state that, for recognizing reserves, there must exist (or there must be a
reasonable expectation that there will exist) the legal right to produce or a revenue interest in the
316
production, which can be summarized into elements that support and establish an economic
interest and the ability to recognize reserves and resources. These include the following:
• The right to produce oil or gas
• The right to take produced volumes in kind or share in the proceeds from their sale
• Exposure to market risk and technical risk
• The opportunity for reward through participation in exploration, appraisal, development,
and producing activities
In addition, the regulation establishes specific elements that do not support an economic
interest and preclude the recognition of reserves and resources. These include the following:
• Participation that is limited only to the right to purchase volumes
• Supply or brokerage arrangements
• Agreements for services or funding that do not contain aspects of risk and reward or
convey an interest in the minerals
Note that the FASB Topic 932 permits reporting of Proved Reserves received under long-
term supply agreements with governments provided that the enterprise wishing to report the
reserves participates in the operation or otherwise serves as the operator. Under the PRMS,
applying the principle to this type of agreement, recoverable amounts could be considered for
classification as reserves or resources depending on project maturity and technical certainty.
The right to extract hydrocarbons and the exposure to elements of risk and the opportunity for
reward are key elements that provide the basis for recognizing reserves and resources. Many
companies use these elements to differentiate between agreements that would allow reserves and
resources to be recognized and reported to regulatory agencies from those purely for services that
would not allow recognition of reserves and resources. Risks and rewards associated with oil and
gas production activities stem primarily from the variation in revenues from technical and
economic risks. Technical risk affects a company’s ability to physically extract and recover
hydrocarbons, and it is usually dependent on a number of technical parameters. Economic risk is
a function of the success of a project and is critically dependent on the ability to economically
recover the in-place hydrocarbons. It is highly dependent on the economic environment over the
life of the project and fluctuates with the prevailing price and cost structures. It should be noted
that risk associated with variations in operating cost alone is not generally sufficient to fulfill the
requirements of risk and reward and allow reserves to be reported. It should also be noted that the
ability or obligation to report reserves to regulatory agencies does not necessarily imply ownership
of the underlying resources. Reserves and resources entitlement under the PRMS are reported
according to economic interest, after deducting royalties, when appropriate, and any interests
owned by others.
Some contracts or agreements may not be satisfactorily clear enough to draw a definitive
conclusion about reserves and resources entitlement. In such contracts, the responsibility for
payments of royalties and production taxes to the government and contract clauses that may refer
to reserves or resources ownership must be analyzed to clarify entitlement.
12.4.2 Working Interest and Economic Interest. An economic interest in minerals (sometimes
referred as mineral interest) may result from a working interest (WI) or a non-working interest.
Typically, a WI is created when a mineral rights owner, who owns minerals beneath the surface,
leases or otherwise contracts another party to access the subsurface and conduct exploration,
development, and production activities. The party that conducts such activities will typically have
certain cost obligations, which are based on the WI, and may have conveyed to it an economic
317
interest, which is related to the WI, but not necessarily numerically equivalent. The economic
interest resulting from a WI is typically less than the numerical WI, being reduced by non-working
interests held by other parties (e.g., non-operating WI owners, royalty owners, overriding royalty
interest owners, etc.). In this instance, the economic interest may be referred to as the “net
economic interest” or “net revenue interest” or even just “net interest.” Note that for the purpose
of this chapter, the term economic interest is used. As companies undertake joint operations,
typically to share costs and risks, WI may be shared among companies. A non-working interest
may have a reduced or no-cost obligation, which would include royalties and similar interests
(discussed herein). For example, the mineral rights owner may retain a non-operating interest
(royalty interest, in this case) when conveying a WI. In such a case, the economic interest of a non-
working interest may be equivalent to the interest held. Contracts and agreements may have several
types of operating and non-operating interest.
In a joint venture, partners typically align on a best estimate production forecast for project
planning purposes. However, other estimates in the uncertainty range are seldom shared or agreed
upon due to the fact that each partner will likely have its own perception of particular uncertainties.
For example, as each partner estimates 1P, 2P, and 3P reserves as a function of its perceived
uncertainties (which may consider the partner’s technical experience, its project approval process,
its forecast prices and costs associated with the production forecast, and, in the case of non-
operator partners, availability of data and time dedicated to the project), differences in estimated
reserves may exist between partners. The low estimate is usually more reflective of such
differences due to the accentuated effect of varied assumptions and knowledge on lesser quantities.
Partners will frequently consider their estimated reserves as strategic information, in consideration
of partner transactions such as farmouts (assignment of all or part of the WI, including exploration
and/or production rights and obligations, to another party in return for agreed-upon compensation),
trades, and exchanges. As such, partners will be less likely to share anything other than the best
estimate approximating the 2P reserves. The low estimate (indicative of the Proved Reserves) is
particularly protected due to the implications of possibly significant differences, as noted. This
does not imply that the economic interest is ignored, but rather it is a recognition that each partner
may have a different basis for uncertainties and some economic and commercial elements.
12.4.2.1 Example: Difference Between Working Interest and Economic Interest in a
Concession Contract. A government has subsurface mineral rights and leases a property to two
companies, retaining 15% of production, free and clear of any costs of exploring, developing, or
operating the property. According to the lease contract, Company A has 60% interest, and
Company B has 40% interest. Company A is designated as operator.
In this case, the government will have a WI of 0% and a 15% economic interest (a non-
working interest). Company A will have a 60% WI, and Company B will have 40% WI. As the
operator, Company A will typically manage the day-to-day operations, with Company B being
responsible for its proportionate share of expenses, according to a joint operating agreement
between both companies. Company A’s economic interest in the production and entitled reserves
or resources will be 51%, while Company B’s economic interest will be 34% (both having their
WI shares proportionately reduced by the 15% non-working interest of the government).
In many types of contracts and agreements, the associated economic interest may be less
simple to determine. Production-sharing and similar contracts have more complex calculations for
economic interest and entitled reserves and resources, involving cost recovery and other
mechanisms, which may be variable with time. Additionally, economic interest may be conveyed
to other companies.
318
12.4.3 Mineral Property Conveyances. A mineral (economic) interest in a property may be
conveyed to others to raise revenue, to spread economic and technical risks, to obtain financing, to
improve operating efficiency, or to accrue tax benefits. Some types of conveyances are essentially
financial arrangements or loans and do not carry with them the ability to recognize or report reserves
or resources (discussed herein as including no economic interest). The forms that transfer an
economic interest would include WIs (operating and non-operating) and non-working interests. A
common conveyance transfers all or a part of the rights and responsibilities of operating a property
(an operating WI) and the ability to recognize reserves or resources. Alternatively, a party may
receive a transfer of a portion of the WI that does not include direct operatorship (a non-operating
WI). The transferor may or may not retain an interest in the oil and gas produced that is free of the
responsibilities and costs of operating the property (a non-working interest), and/or the transferor
may convey an interest to another party and retain a portion of the WI.
Farmout agreements, overriding royalty agreements, or volumetric production payments
agreements would typically convey reserves and resources entitlement. The FASB ASC 932-360-
55-2 to 932-360-55-14 text (FASB 2010) provides useful guidance on when reserves and resources
associated with transfer of interest may be recognized in the PRMS categories. The FASB ASC
932-470-25-1 text (FASB 2010) provides useful guidance on arrangements involving reserves and
resources that are, in substance, borrowings and may not be recognized in the PRMS.
As mentioned in Section 12.4.1, in some contracts or agreements, it may not be easy to draw
a definitive conclusion about reserves or resources entitlement. In reporting entitled quantities,
opinions regarding the responsibility for payments of royalties and production taxes to the
government, as well as contract clauses that may be material to reserves ownership, must be
properly disclosed.
12.5 Agreements and Contracts

Agreements and contracts cover a wide spectrum of legal arrangements typically established by
host countries to best meet their sovereign needs, including conveying licenses, rights, and/or
opportunity to potentially profit from petroleum resources under their purview, as well as some
other common agreements among entities. While certain approaches have been routinely applied
for determining when reserves or resources can be recognized under these contracts and
agreements, there is no sanctioned established system of practice for doing so. The purpose of this
section is to expand on the text contained in PRMS § 3.3.2 by providing more detailed information
for frequently encountered agreement types and to promote consistency in the recognition of
reserves and resources under them, based on the essence of the contract. The focus is on the
specific elements of the agreements that enable recognition of reserves and resources but not on
the specific PRMS classes and categories.
Presently, mineral ownership in most onshore properties in the US belongs to the private
landowner (or subsurface mineral owner), who has the authority to execute a lease agreement with
a production and development company. Globally, host countries will hold mineral ownership,
and companies may be entitled to reserves/resources through what can be described broadly as
concessionary fiscal systems or contractual fiscal systems. In a concessionary fiscal system
(described in Section 12.5.1), the mineral right owners transfer title of resources produced to
companies (contractors). In a contractual fiscal system, typically the mineral rights owners retain
ownership of resources produced and grant companies the right to receive a share of production or
its revenue. The most frequent oil and gas contract types (including frequent types of conveyances)
are described in Sections 12.5.2 through 12.5.12.
319
This section follows the classification system, modified from the template originally proposed
by Johnston (1994) as shown in Fig. 12.3, with the broad descriptions of concessionary fiscal
systems or contractual fiscal systems. In some fiscal systems, an economic interest may be
recognized by the contractor or lessee, which will then be entitled to reserves/resources.
Contractors and lessees may also arrange for other contracts or agreements (e.g., purchase
agreements, production payments, loan agreements). Some of those contracts and agreements may
convey an economic interest, initially granted by a fiscal system.
In some types of contracts or agreements, such as concessions, production-sharing contracts,
and revenue-sharing contracts (described and discussed in Sections 12.5.1, 12.5.2, and 12.5.3,
respectively), considering the host country’s regulation, reserves and resources entitlement may
be clear. In other cases, careful examination of contract and agreement terms, along with the host
country’s regulation, is needed to determine reserves and resources entitlement. Furthermore,
countries and the oil industry may refer to some contractual arrangement as one contract model,
when it is fundamentally another contract model. The expanded template of fiscal systems and
some other agreement types along with their potential to recognize reserves and resources and
report them under the PRMS is shown in Fig. 12.4 (modified from the figure originally proposed
by McMichael and Young 1997). The greater participation a contractor has in exploration,
development, and production activities and exposure to risk and physical production ownership,
the stronger is the probability of reserves and resources entitlement.
12.5.1 Concessions, Mineral Leases, and Permits. Historically, concessionary systems, such as
leases and concessions, have been the most commonly used agreements between oil companies
and governments or mineral owners. In such agreements, typically, the host government or mineral
owner grants the producing company the right to explore for, develop, produce, transport, and
market hydrocarbons or minerals within a defined area for a specific amount of time. The
production and sale of hydrocarbons from the concession are then typically subject to rentals,
royalties, bonuses, and taxes. Under these types of agreements, the company usually bears all risks
and costs for exploration, development, and production and generally would hold title to all
resources that will be produced while the agreement is in effect. Reserves and resources consistent
with the net working interest (after deduction of any royalties owned by others) that can be
recovered during the term of the agreement are typically recognized by the producing/development
company. Ownership of the reserves or resources producible over the term of the agreement is
normally taken by the company. However, as described in PRMS § 3.3.3 and in Section 12.6.4
herein, volumes recoverable after the term of the contract could potentially be recognized as
Contingent Resources, i.e., contingent on the successful negotiation of an agreement extension.
The project(s) under the terms of the agreement would not result in the “split classification”
violation noted in PRMS § 2.2.0.4 by having, for example, 2P Reserves with 2C Contingent
Resources applicable during a postconcession term. Funding and development of the resources
during the postconcession term in this context are considered to be a separate project. If the
agreement contained provisions for an automatic extension (or there is an established track record
for extensions), and the likelihood of extension was judged to have a reasonable expectation of
being achieved, additional reserves could potentially be recognized for the length of the extension
period, provided requirements for project commitment, funding, economic viability, and other
commerciality requirements were satisfied.
320
Petroleum Fiscal Arrangements
Concessionary Systems Contractual Systems
Service Production-Sharing
Contracts Contracts
Pure-Service Risked-Service Revenue-Sharing

Contracts Contracts Contracts
Fig. 12.3—Classification of petroleum fiscal systems.
Fig. 12.4—Reserves/resources reporting potential.
12.5.2 Production-Sharing Contracts. In a production-sharing contract (PSC; also called

production-sharing agreement, or PSA) between a contractor and a host government, the contractor
typically bears all risks and costs for exploration, development, and production. In return, if
exploration is successful, the contractor is given the opportunity to recover the investment from
production (cost hydrocarbons), subject to specific limits and terms. The contractor also receives
a stipulated share of the production remaining after cost recovery (profit hydrocarbons).
Ownership of the underlying resource is almost always retained by the host government. However,
the contractor normally receives title to the prescribed share of the quantities (an economic
interest), based on the cost and profit revenue divided by the average product price, as they are
produced or sold. Subject to technical certainty, reserves in one or more of the PRMS categories
based on cost recovery plus a profit element for hydrocarbons that are recoverable under the terms
321
of the contract are typically recognized by the contractor. Resources may also be recognized for
future development phases where project maturity is not sufficiently advanced or for possible
extensions to the contract term where this would not be a matter of course.
As in the case of a concession, volumes recoverable after the term of the contract ends will
normally be classified as Contingent Resources unless the contract contained provisions for
extension, and there is reasonable expectation that an extension, a renewal, or a new contract will
be granted (refer to previous Section 12.5.1 and see also Section 12.6.4). Costs considered for
recovery by the contractor are commonly specified in the contract and may be limited to
percentages of production or revenues. They are commonly subject to the host country approval
(via its regulator or national oil company [NOC]). Under a production-sharing contract, the
contractor’s entitlement to production generally decreases with increasing prices because a smaller
share of production is required to recover investments and costs. This may cause price-related
volatility in annual reserves estimates, particularly in cases using a constant price. These
agreements may also contain terms that reduce contractor entitlement as production rate
(production tranches) and/or cumulative production increases. In many host countries, entitlement
will vary according to some sort of specifically defined relation between revenue and costs (termed
R factors). The R-factor value (e.g., R = cumulative revenue ÷ cumulative capital expenditures, or
R = cumulative revenue ÷ cumulative expenses) and the corresponding reduction applied to
entitlement are typically defined in a tranche table, where increasing R values correspond to
decreasing entitlement. Fig. 12.5a is a schematic indicating the distribution of yearly project
production between contractor and government. Fig. 12.5b shows examples of profit split between
the contractor and host government (indicated by NOC in the figure), based on (a) average daily
production and (b) an R-factor representing the ratio between the contract cumulative revenue and
cumulative expenses. Note that “revenue” and “expenses” are defined in the contract and may be
gross, net, or some specific description unique to the contract.
12.5.2.1 Example: Resources Estimates in a Production-Sharing Agreement. This example
is for an oil project, but references can apply to gas or barrel of oil equivalent quantities. In a PSA,
the profit oil and cost oil for each period are calculated as:
Profit Oil (STB) = [Sales (as defined in the agreement) Oil Production Net of Royalty (STB) –
Recoverable Costs (USD)]/Price (USD/STB).
Cost Oil (STB) = Recoverable Costs (USD)/price (USD/STB), where Recoverable Costs (“Cost
Recovery”) are those costs that are allowed to be reimbursed in oil during the
period. Those recoverable costs may arise from costs incurred in the period as
well as from past unrecovered investments and operating costs.
In this contract example, there is a limit on the amount of product that the contractor may recover
during each period: Cost Oil Limit each period = 70% of Gross Revenue before royalty
deduction for each period.
In this contract example, Royalties are paid in kind as:
Royalty (STB) = 15% of sales production (STB).
322
(a) 100 Barrels
Contractor Government
Royalty
Cost
Barrels Cost Recovery
% Profit (100%) Profit

Barrels Profit Split Barrels
Mechanism
Total Entitlement
Volume
Taxes
(b) Profit Oil based on R-Factor

R-Factor Contractor NOC
Profit Oil based on average daily production R<1.5 50% 50%
Average Daily Production (bbl) Contractor NOC 1.5 ≤r< 2 45% 55%
0-30,000 40% 60% 2 ≤r< 2.5 40% 60%
30,000-200,000 30% 70% 2.5 ≤r< 3.5 35% 65%
Above 200,000 20% 80% 3.5 ≤r< 4.5 30% 70%
R≥4.5 25% 75%
Fig. 12.5—(a) Example production-sharing contract. (b) Profit-split examples. NOC stands for National Oil Company.
In this contract example, profit oil and cost oil are based on sales quantities. In some contracts,
profit oil or gas and cost oil or gas may be based on produced quantities. Each contract will have
its own specifications for profit oil or gas, cost-recovery limits, royalty percentages, and other
definitions for royalties and cost recovery. The contractor is entitled to a share of the profit oil,
based on an R-factor variable table (part of the contract), and the contractor is also responsible for
income taxes in the host country. The project is already in production, and the best estimate
forecast for next year is:
Oil production = 100 thousand STB.
Contractor’s forecast price = USD 50/STB.
Recoverable Cost of the period (next year), including unrecovered cost from previous years
and forecast cost of year 1 allowed to be recovered: USD 2,500 thousand.
Profit split of the year based on the R-factor table for this specific contract:
40% (attributed to the host government).
Fig. 12.6 shows a schematic of how to estimate entitlement resources for next year.
Fig. 12.6—Entitled resources.
323
The contractor may assume it will recover all recoverable costs (USD 2,500 thousand) next
year because the limit of recoverable cost of 70% of gross revenue before royalty deduction (100
thousand STB × USD 50 /STB × 70% = USD 3,500 thousand) is not exceeded.
The contractor best estimate entitled resources corresponding to next year are calculated under
this example as:
Cost Oil + [(Production Net of Royalty) – (Cost Oil)] × (1 − NOC Profit Split)
= (USD 2,500 ÷ USD 50/STB) + [ (100 − 15) − 50 ] × 60% ]
= 50 thousand STB + 21 thousand STB = 71 thousand STB.
The calculation of entitled reserves/resources that will be produced in the remaining future
years will be based on each year’s forecast of production, royalty rate, costs to be recovered as
cost oil, and forecast profit split based on the R-factor table. 2P Reserves associated with this
project will be estimated by adding the entitled best estimate resources of each year until the
economic limit, if there is no other contingency that would prevent commercial production to that
date. The actual economic limit calculation will consider revenues from the contractor’s entitled
quantities and the contractor’s expenses (which may differ from the recovered costs of the period).
1P and 3P Reserves will be estimated based on the same logic used for 2P Reserves.
Note that entitled reserves/resources related to cost oil (50 thousand STB) would be higher if
the forecast price were less than USD 50/STB. If the forecast price was USD 40/STB, cost oil
would be 2,500 thousand/40 = 62.5 thousand STB, and the recalculated profit oil would be ([85
thousand STB − 62.5 thousand STB] × 60%) = 22.5 thousand STB, where 13.5 thousand STB is
entitled to the contractor. This would bring the contractor entitlement reserves/resources
corresponding to next year production to 76 thousand STB. Conversely, reserves/resources related
to cost oil (50 thousand STB) would be lower if the forecast price were higher than USD 50/STB.
PSCs normally have limits and specifications for cost oil in each period that might be based
on production revenues. Therefore, when prices get very low, it might be more frequent for
incurred expenses to be recovered in the following year. Each contract and host government will
have its own cost-recovery limits and specifications, profit split tables, prices used for cost oil,
royalty volumes determination, and other information to be used in the estimation of
reserves/resources entitlement. As noted, recoverable costs are usually subject to NOC approval,
so estimated future cost oil estimates should consider historic approval level.
Table 12.1 shows an example of estimated entitled quantities, contractor cash flow, and
reserves estimates for the remaining years for the best estimate, considering a forecast production
decline, as shown in line (B) of the table, forecast price as shown in line (A), and a fixed profit oil
split of 40%. For this table construction, sales price was assumed to be the same price used for
cost oil calculation. The company’s forecast costs displayed in line (G) represent costs allowed to
be recovered under this specific PSC contract, while those displayed in line (O) represent costs
that will incur but are not allowed to be recovered under this specific PSC. (These are only
examples for this table construction.) In this example, the economic limit would occur in year 5
(line P). Estimated 2P Reserves are the accumulated forecast quantities from years 1 to 5 (260.56
thousand STB), if all remaining commerciality requirements are met.
324
REF Year 1 2 3 4 5 6
(A) Price Forecast (USD/STB) 50 45 35 35 35 35
(B) Gross Oil Production (thousand STB) 100 83 69 58 48 40
(C) Oil Production net of Royalty (thousand STB): 85% X (B) 85,00 70,55 58,65 49,30 40,80 34,00
(D) Gross Revenue before royalty deduction (thousand USD): (A) X (B) 5000 3735 2415 2030 1680 1400
Cost Oil Calculation
(E) Limit to Recover Cost Oil (thousand USD): 70% X (D) 3500 2615 1691 1421 1176 980
(F) Unrecovered Cost from prior periods (thousand USD) 950 15
(G) Capex + Opex of the period allowed to recover (thousand USD) 1550 2629 1147 1004 874 770
(H) Balance to recover (thousand USD): (F) + (G) 2500 2629 1162 1004 874 770
(I) Cost Recovered in the period (thousand USD): (H), limited by (E) 2500 2615 1162 1004 874 770
(J) Cost Oil (STB): (I)/(A) 50,00 58,10 33,19 28,69 24,97 22,00
Profit Oil Calculation
(K) Oil to share (thousand STB): (C) - (J) 35,00 12,45 25,46 20,61 15,83 12,00
(L) Contractor Share (thousand STB): 60% X (K) 21,00 7,47 15,28 12,37 9,50 7,20
(M) Total Contractor Share (thousand STB): (L) + (J) 71,00 65,57 48,46 41,05 34,47 29,20
Contractor Cash Flow
(N) Revenue (thousand USD): (A) X (M) 3550 2951 1696 1437 1206 1022
(O) Non recoverable project costs of the period 350 333 319 308 298 290
(P) Net Cash Flow (thousand USD): (N) - (G) - (O) 1650 -11 230 125 34 -38
Table 12.1—Estimated entitled quantities and reserves estimates. Rounding may cause distortion in numbers.
12.5.3 Revenue-Sharing Contracts. Revenue-sharing contracts are very similar to the PSCs
described earlier, with the exception of contractor remuneration. As in a PSA, the contractor
typically bears all risks and costs for exploration, development, and production. In return, the
contractor is given the opportunity to recover the investment plus obtain a return on the investment
from an agreed sharing of the net revenue. There is typically no explicit cost recovery, but rather
the contractor is expected to pay its costs with its share of the revenue. The contractor normally
receives title to the prescribed share of the volumes as they are produced (just as in a PSA). Subject
to technical certainty, reserves/resources in one or more of the PRMS categories based on the
revenue share that are recoverable under the terms of the contract are typically recognized by the
contractor. Revenue-sharing agreements are commonly referred to as risked-service contracts, or
even production-share contracts, in the industry. Fig. 12.7 is a schematic of the distribution of
yearly project revenue between contractor and government.
100 Barrels
Royalty In-Kind or
Monetary
Cost Cost Recovery

Recovery
Revenue
% Profit (100−%) Profit

Revenue Revenue
Revenue Split
Split Revenue
Mechanism
Mechanism
Total
Revenue
Received Taxes
Fig. 12.7—Example revenue-sharing contract.
325
12.5.4 Risked-Service Contracts. With a risked-service contract, the contractor usually receives
a defined remuneration fee, fixed or variable, plus cost recovery (introducing risk/reward due to
uncertainty and pace of payout). The contractor has an economic or revenue interest in the
production and hence can recognize reserves and resources. As in the PSC, the contractor provides
the capital and technical expertise, at its sole risk, required for development. The contractor can
recover those costs (as prescribed in the contract) from sales revenues. The remuneration fee
typically requires achieving some target of production or incremental production (this adds to the
element of risk/reward). Similar to a PSC, resources may be recognized for future development
phases or possible extensions to the contract term. Volumes recoverable after the term of the
contract ends will normally be classified as Contingent Resources unless the contract contains
provisions for extension, and there is reasonable expectation that an extension, a renewal, or a new
contract will be granted (refer to previous Section 12.5.1 and see Section 12.6.4).
Fig. 12.8 is a schematic of the distribution of yearly project revenue between contractor and
government. This type of agreement is also often used where the contracting party provides
expertise and capital to develop and enhance operations in an existing field and has rights and
obligations and bears risks similar to those in the previously noted agreement types.
100 Barrels
Contractor Payment
Mechanism
In-Kind Investment
Investment
Payment Component
Component
And / Or Expense
Expense Remainder
Component After Contractor
% of Component
Payment
Production
Revenue Performance
Performance
Component
Component
Total In-Kind
and/or Monetary
Payment Received
Taxes
Fig. 12.8—Example risked-service contract.
Reserves and resources recognized under the PRMS and those reported to regulatory agencies
would be based on the economic interest held or the financial benefit received. Depending on the
specific contractual terms, the reserves and resources equivalent to the value of the cost-recovery-
plus-remuneration fee (fixed and/or variable) are normally reported by the contractor. In the
absence of a cost-recovery mechanism for this type of risk-service agreement, a fixed remuneration
fee alone (e.g., USD 100/year or USD 10/STB) would not typically characterize an economic
interest in the production (see Section 12.5.5—Pure-Service Contracts). If an incentive that
depends on meeting and/or exceeding a baseline of production is part of the agreement
(incorporating only a remuneration fee), then there is the potential for an economic interest to be
realized. In this circumstance, a common approach is to recognize the incentivized portion of the
production as entitled reserves/resources, in accordance with its risk and reward.
12.5.4.1 Example: Risked-Service Contracts. A host country has granted a 6-year, risked-
service contract to Company One. The terms of the contract state that Company One will assume
management of a producing oil field (called Blue Field) and will be tasked with maintaining and
increasing production from the Blue Field. A baseline of production that must be maintained is
provided as part of the contract (see Table 12.2). As long as average production for any given year
326
meets or exceeds that baseline production, Company One will be paid a remuneration fee equal to
USD 1.50 for every whole barrel produced above the baseline production. Company One will be
allowed to request 100% repayment of annual approved (by the host country) operating expenses
not to exceed USD 20 per barrel of the projected baseline production (or 50% of the prevailing
sales price per barrel, whichever is lower) for the life of the project. Capital costs can also be
recovered based on pre-approved (by the host country) budgeted capital for Blue Field. Table 12.2
illustrates how the estimate of reserves in the Blue Field would be apportioned to Company One,
assuming an estimated constant oil price of USD 50/STB.
A B C D E F G H=(E+F+G)/USD 50
Estimated Gross Baseline Incremental Approved Pre-Approved Remuneration Net

Year Oil Production Production Production Over Operating Costs Capital Costs Fee Entitlement
(STB) (STB) Baseline (STB) (USD) (USD) Revenue (USD) (STB)
1 40.000 40.000 0 200.000 50.000 0 5.000
2 38.000 36.000 2.000 195.000 250.000 3.000 8.960
3 38.000 32.400 5.600 195.000 0 8.400 4.068
4 40.000 29.200 10.800 200.000 0 16.200 4.324
5 36.000 26.200 9.800 190.000 0 14.700 4.094
6 32.400 23.600 8.800 181.000 0 13.200 3.884
Table 12.2—Company One production forecast, contract production baseline, and estimated entitled quantities.
The entitlement quantity is calculated by summing the sources of revenue (cost recovery plus
remuneration fee) for the contractor (Company One) and dividing by the prevailing oil price
(forecasted as a constant USD 50/STB in this example). Company One is exposed to both market
risk and technical risk. For example, if for some reason, Company One’s production forecast
dropped significantly, then Company One’s entitlement would be affected by both the reduced
forecast and by the possibility of not being reimbursed for all operating costs (particularly if costs
are firm) or investments. There is also the possibility that the reduced production would preclude
achievement of an incremental production above the baseline. Price fluctuations may also affect
entitlement significantly, as entitled quantities are based on cost plus investment divided by price,
with a sales price/barrel limit for operating cost reimbursement. Pre-approved investments are
intended to be reimbursed, but with low prices or low production, there might be a possibility of
not being fully reimbursed.
12.5.5 Pure-Service Contracts. A pure-service contract is an agreement between a contractor and

a host government that typically covers a defined technical service to be provided or completed
during a specified period of time. The service company investment is typically limited to the value
of equipment, tools, and personnel used to perform the service. In most cases, the service
contractor’s reimbursement is fixed by the terms of the contract with little exposure to either
project performance or market factors. Payment for services is normally based on daily or hourly
rates, a fixed turnkey rate, or some other specified amount. Payments may be made at specified
intervals or at the completion of the service. Payments, in some cases, may be tied to the field
performance, operating cost reductions, or other important metrics. In many cases, payments are
made from government general revenue accounts to avoid a direct linkage with field operations.
Risks undertaken by the service company under this type of contract are usually limited to
nonrecoverable cost overruns, losses owing to client breach of contract or default, or contract
dispute. These agreements generally do not have exposure to petroleum exploration, development,
327
production volume, or market price risks or other risks; consequently, reserves and resources
entitlement is not (typically) recognized under this type of agreement. The service company may,
however, have an obligation to report gross (total working interest basis) reserves and resources to
the regulatory agencies of the host country, but this would not imply any ownership or entitlement
to reserves. These agreements can be very complex, and an evaluator should carefully analyze the
contract terms before characterizing them as pure service and making a decision on whether
entitlement of reserves can be claimed. Fig. 12.9 is a schematic of the distribution of yearly project
revenue between contractor and government.
100 Barrels
Contractor Payment
Mechanism
Expense
Expense
Component
Component
Monetary Investment
Investment Remainder
Payment not Component
Component After Contractor
related to Payment
production Performance
Performance
Component
Component
Total Payment
Taxes
Received
Fig. 12.9—Example pure-service contract.
12.5.6 Loan Agreements. A loan agreement is typically used by a bank, other financial investor,
or partner to finance all or part of an oil and gas project. Compensation for funds advanced is
typically limited to a specified interest rate. The lender does not participate in profits earned by
the project above this interest rate. There is normally a fixed repayment schedule for the amount
advanced, and repayment of the obligation is usually made before any return to equity investors.
Risk is limited to default by the borrower or failure of the project. Variations in production, market
prices, and sales do not normally affect compensation. Reserves and resources would not be
recognized in any PRMS categories by the lender under this type of agreement, although they
would be reported by the contractor.
12.5.7 Production Loans, Forward Sales, and Similar Arrangements. There are a variety of
forms of transactions that involve the advance of funds to the owner of an interest in an oil and gas
property in exchange for the right to receive the cash proceeds of production, or the production
itself, arising from the future operation of the property. In such transactions, the original interest
owner almost invariably has a future performance obligation, the outcome of which is uncertain to
some degree. Determination of whether the transaction represents a sale of interest or financing
rests on the particular circumstances of each case. If the risks associated with future production,
particularly those related to ultimate recovery and price, remain primarily with the original interest
owner, then the transaction should be considered as financing or contingent financing. In such
circumstances, the repayment obligation will normally be defined in monetary terms and would
not enable recognition of reserves and resources by the funding entity under the PRMS.
328
In cases where funds advanced for exploration are repayable by offset against purchases of
oil or gas discovered, or in cash if insufficient oil or gas is produced by a specified date, such
arrangements are considered borrowings, and entitlement of reserves and resources remain with
the production seller. When funds are advanced to an operator for development of a property (or
for increased operating activities) and are repayable in cash out of the proceeds from a specified
share of future production from the producing property, until the amount advanced plus interest
at a specified or determinable rate is paid in full, these arrangements are typically considered a
borrowing, and reserves and resources entitlement remains with the original producer. Such
transactions, as well as those described in Section 12.5.11, are commonly referred to as
production payments.
If the risks associated with future production, particularly those related to ultimate recovery
and price, rest primarily with the funding entity, then the transaction should be considered either a
contingent sale or a disposal of fixed assets (and sale of interest). In this case, reserves and
resources would be recognized under the PRMS by the funding entity. The ability/duty to report
reserves to applicable government agencies may be permissible, depending on their regulation.
Volume-denominated production payments (see Section 12.5.11) typify such sales of interest.
For the companies’ financial statements, the specific accounting standards for the jurisdiction
should be consulted for appropriate treatment. In some regulatory environments, merely holding
such a quantity deemed as purchased may not meet the standard of the purchaser as an oil and gas
producing entity, which may preclude reporting such volumes as reserves or resources.
12.5.8 Carried Interests. Carried interests cover a broad spectrum of arrangements where one
party (the carrying party) agrees to pay for a portion or all of the costs of another party (the carried
party) on a license or contract in which both own a portion of the WI. The following discussion is
intended to illustrate the basic concepts of carried interests and the reporting of resources under
them with a simple type of reversionary agreement. In this example arrangement, the carrying
party agrees to pay for the preproduction costs of the carried party. This arises when the carried
party is either unwilling to bear the risk of exploration or is unable to fund the cost of exploration
or development directly. Owners may enter into carried-interest arrangements with existing or
incoming joint venture partners at the exploration stage, the development stage, or both.
If the property becomes productive, then the carrying party will be reimbursed either (a) in
cash out of the proceeds of the share of production attributable to the carried party or (b) by
receiving a disproportionately high share of the production until the carried costs have been
recovered. In the case of mechanism (b), the adjusted share of production held by the parties during
the recovery of the carried costs period (“payout”) may be referred to as “interest before payout.”
At the point of recovery of the carried costs (or a multiple thereof), the parties will revert to each
respective original share of production, referred to as “interest after payout” or as a “reversionary
interest.” The carrying party normally recognizes the additional future production received in one
or more of the PRMS reserves and resources categories. The estimate of reserves and resources
for all the parties would be based on the future projection of production and revenue, to which the
interest before payout would be applied prior to the point of payout and the interest after payout
(or reversionary interest) would be applied for the remaining period of production. If project
maturity is not sufficient to classify the amounts as reserves, then the PRMS resources classes and
categories would be used according to the agreed reimbursement terms.
If reimbursements are guaranteed at some level of monetary payment, with the potential for
any shortfall from the entity of interest to be compensated from another production stream or just
329
paid in cash, then this is typically considered as removal of the carrying party from the requisite
risk of production, and, in that case, quantities should not be booked as reserves or resources
entitlement for the carrying party.
12.5.9 Entitlement with Interest Before Payout and Interest After Payout Example. Consider
a producing project with two partners (Oil Company A and Oil Company B), each holding 50%
of the rights to production (and the same responsibility for costs), where all forecast quantities are
categorized as Proved Reserves by Company A. A new infill well is proposed, and drilling costs
will be USD 1,000,000. Company B does not have the capital to fund the drilling well, so
Company A agrees to pay the full cost of the well in exchange for a higher (25% more) portion of
production (and related costs) until the share owed by Company B (USD 500,000, which it was
supposed to pay for the well) is fully repaid. The agreed interest before payout is Company A =
75% and Company B = 25%. After payout, the interests will revert to 50% each (interest after
payout, also called reversionary interest). Table 12.3 illustrates how Proved Reserves quantities
would be apportioned to each party, considering Company A gross oil, cost, and revenue low
estimate forecast, for a period of 6 years to show calculation during and immediately after the
payout period. (For this example, there are no royalties due.)
Company A Company B
Gross Revenue
Gross Revenue Year End
Less Costs from
Gross Oil Carried Cost Less Costs from Balance of Net Entitled Net Entitled
Year Operations for WI (%) WI (%)
(STB) (USD) Operations Payout Gross Oil (STB) Gross Oil (STB)
Payout: 25%
(USD) (USD)
(USD)
1 20.000 500.000 600.000 150.000 350.000 15.000 75% 5.000 25%
2 16.000 440.000 110.000 240.000 12.000 75% 4.000 25%

3 14.000 360.000 90.000 150.000 10.500 75% 3.500 25%
4 13.000 320.000 80.000 70.000 9.750 75% 3.250 25%
5 12.000 280.000 70.000 - 9.000 75% 3.000 25%
6 10.000 248.000 - - 5.000 50% 5.000 50%
Total 85.000 500.000 2.248.000 500.000 61.250 23.750
Table 12.3—Apportioned quantities during interest before payout and interest after payout periods using Company A
forecast for 6 years.
If the license to produce ended at year 6, then Company A Proved Reserves would equal the
total entitled quantities (61,250 STB). Forecasts for the best estimate could be used in a similar
logic for 2P Reserves estimates. Considering the remaining forecast years and no expiration date
for the license to produce, Company A would determine the economic limit for entitled reserves
estimates based on its perceived uncertainties and the same forecast prices, costs, and share of
production. Company B would use similar logic based in its perceived uncertainties, forecast
prices, and costs, and its share of production to estimate reserves.
In this example, there were no royalties due. If royalties were due, then quantities to settle
carried costs would usually be based on quantities and revenues after royalties. The actual payout
terms will usually be described in an agreement among partners (e.g., joint operating agreement).
12.5.10 Purchase Contracts. A contract to purchase oil and gas provides the right to purchase a
specified volume at an agreed-upon price for a defined term. Under purchase contracts, exposure
to technical and market risks is borne by the production seller. While a purchase or supply contract
330
can provide long-term access to reserves and resources through production, it does not convey the
right to extract, nor does it convey an economic interest in, the reserves. Consequently, reserves
and resources would not be recognized under the PRMS for this type of agreement.
12.5.11 Volume-Denominated Production Payments. Unlike the agreements described in

Section 12.5.7, which are essentially loans or borrowings, there are arrangements that result in the
transfer of an interest in reserves from one party to another. These types of arrangements are
volume-denominated production payments, and they specify terms where assets (typically cash)
are transferred between participants in return for the right to take a specified volume from the
production stream of a project. In common practice, volume-denominated production payments
are also commonly referred to as “production payments,” making the distinction between the two
different types of production payments less clear. Production payments have been widely used as
a hedging vehicle in periods of price volatility.
In volume-denominated production payments, as referenced in Section 12.4.3, reserves and
resources may be recognized by the purchaser (entity who takes the production), thus reducing the
entitlement of the seller, for quantities associated with a production payment. In such an
arrangement, the production seller’s obligation is not expressed in monetary terms but rather as an
obligation to deliver, free and clear of all expenses associated with operation of the property, a
specified quantity of oil or gas to the purchaser out of a specified share of future production. In
order to be considered a recognizable quantity for reserves and resources, the essential
requirements of Section 12.4.1 must be met. The specific accounting standards for the jurisdiction
should be consulted for appropriate accounting treatment of any transaction. Fig. 12.10 gives an
example of a typical production payment arrangement.
100 Barrels
Volume: Specified Portion of

Production and Duration
Purchaser of Seller
Purchaser of Sellerofof
Production
Production Production
Payment Production
Payment
Payment Payment
Value: Agreed Payment
Fig. 12.10—Example conveyance-production payment.
12.5.12 Other Contracts and Agreements. From time to time, evaluators can encounter unique
agreements that have elements of the descriptions in Section 12.5.1 through 12.5.11 but are not
exactly as described here. These types of arrangements and agreements must be considered in the
context of the key principles outlined above, including economic interest and risk exposure.
Examples of other agreements that are encountered on occasion include net profits interests,
interest swaps, and storage and transfer agreements. Net profits interests are most often
encountered in North America where the holder of the interest holds a stipulated portion of the
proceeds from a well or other specific field entity. Swap agreements are binding contracts that
convey ownership in production from oil and gas properties linked to production in another
separate oil and gas property. Storage and transfer agreements do not typically meet the criteria
for reserves and resources recognition; however, in certain circumstances, agreements can
encompass risk exposure with an economic interest.
331
12.6 Other Concerns of Resources Entitlement and Recognition
Royalties, taxes, unitization agreements, and contract expiration are commonly present in any
contract type. This section discloses considerations that may influence the entitlement of resources.
12.6.1 Taxes and Reserves/Resources. In general, reserves and resources are recognized in
situations where there is an economic interest, and after deduction of any royalty (see Section
12.6.2) owed to others. Production sharing or other types of operating agreements lay out the
conditions and formulas for calculating the share of produced volumes to which a contracting
company will be entitled. As explained above, these volumes are normally divided into cost
recovery and profit volume components. The summation of the cost and profit volumes that the
contractor will receive through the term of the contract represents the reserves and resources to
which the contractor is entitled. In many instances, these agreements may also contain clauses that
provide that host country income taxes will be paid to the government by the NOC on behalf of
the contractor. While details on the specific hydrocarbons produced and revenues that are used to
fund the payments may not usually be specified in the agreement, they are inferred to come from
the government’s share of production (under the agreement). In such cases, the contractor derives
a benefit from the government’s share of hydrocarbons used to fund the contractor’s tax payments,
representing an economic interest in such a share, which may be considered as contractor’s
reserves or resources. By virtue of the economic interest that the contractor has in these additional
volumes, common practice is to include the related quantities in the contractor’s share. This also
typically requires reporting of the value related to the tax payment that is received in the financial
reporting statements. Section 12.7.1 shows an example where volumes related to taxes are included
in reserves or resources estimates.
12.6.2 Royalties, Overriding Royalty, and Government Fixed Volume Entitlement and
Reserves/Resources. A royalty is commonly retained by a mineral owner (lessor/host) when
granting rights to a producer (lessee/contractor) to develop and produce the resources. Royalties
are a non-working (and non-operating) type of entitlement interest in resources that is free and
clear of the costs and expenses of development and production to the royalty interest owner, as
opposed to a WI, where an entity has cost exposure. Royalties are generally a fixed percentage or
may have some form of a sliding scale basis. Depending on the specific terms defining the royalty,
the payment obligation may be expressed in monetary terms as a portion of the proceeds of
production in cash or as a right to take a portion of the volume of production (royalties in kind).
The royalty terms may also provide the option to switch between forms of payment at the discretion
of the royalty owner. In either case, royalty quantities must be deducted from the lessee’s or
contractor’s resources entitlement so that only economic interest quantities are recognized.
Conversely, if an entity owns a royalty or equivalent interest of any type in a project, then the
related quantities should be included in that entity’s resources entitlements and should not be
included in entitlements of others. Royalty interest owners are typically responsible for any
severance or production taxes assessed on their share of production or proceeds of production.
While the royalty description and treatment with respect to entitlement may appear
straightforward, it is not uncommon to see the term “royalty” applied to cash payments that are not,
in fact, a royalty as defined above (an interest retained by the mineral right owner). In some
agreements or fiscal systems, there may not be a clear distinction between production or other form
of taxes imposed by the host government and other payment obligations that may be referred to as
royalties. These other payment obligations are expressed in monetary terms with no alternative to
332
take equivalent quantities in kind. They are typically linked to any or all of the following: production
rates, quantities produced, cost recovery, or the value of production (price sensitive). In many cases,
they can also be offset or reduced by taxes, tariffs, and other obligations, which is contrary to the
definition of a royalty noted above. These payment obligations may be referred to as royalties, but
they do not represent an interest retained by the lessor/host and are effectively an additional form of
tax, and they are therefore not a true royalty. The production and underlying resources are controlled
by the lessee/contractor, who may (subject to contractual terms and/or regulatory guidance) elect to
report these payment obligations as a tax. Such payments referred to as royalties (but more correctly
referred to as a production or other form of tax) must be included as an expense in the resources cash
flows used for economic limit and economic determinations, and all quantities should be recognized
as reserves or resources by the lessee/contractor without a corresponding reduction in
lessor/contractor’s entitlement related to such so-termed royalty payments.
In some cases, government legislation, regulations, or contract agreement terms may contain
clauses that help to establish if the obligation is a true royalty or not, and how the royalty is to be
satisfied in monetary terms. In other cases, judgement must be carefully applied to determine if a
payment obligation is a true royalty or actually a form of production or other kind of tax. While
there are no published standards to differentiate between royalties and taxes, examination of the
specific attributes and the intent of the payment or obligation in comparison to other established
royalties and taxes as recognized in the finance and accounting practices of a given country (e.g.,
as in the US and Canada) is one approach often used to make the distinction. Examples where a
case may be made that the obligation referred to as a royalty may in fact be a tax (production or
other form) rather than a true royalty could include any of the following:
• The obligation is linked to project profitability rather than a defined interest.
• Costs are deductible from the obligation.
• Cash royalty payments are not based on sales quantities.
• Prices used for cash royalties are arbitrated by the government.
• Provisions only allow for cash payment without an option to take in-kind production
quantities.
• The clause describing the obligation is grouped with clauses detailing production or other
taxes and tariffs that are due.
Where the payment is not a royalty interest as defined, then the related reserves and resources are
included in entitlement quantities recognized by the lessee/contractor, and expenses (including the
payment) are recognized in the lessee’s/contractor’s reserves cash flows. Fig. 12.11 shows a
schematic figure for reserves and resources entitlement when royalties are applied in kind or in
cash in a concession contract. Fig. 12.12 is a schematic figure for reserves entitlement when the
term “royalty” is applied to monetary payments that are considered to be production taxes.
It is important to note that a government fixed volume supply (e.g., 100 million CFD) is not
considered to be a royalty but must also be excluded from a lessee’s/contractor’s entitlement.
Similar to royalties, an overriding royalty interest is a non-working interest that is free and
clear of the costs and expenses of development and production, as opposed to a WI, where an
entity has cost exposure. Whereas royalty interests are derived from an ownership in the minerals,
overriding royalty interests are a stated percentage (or they may have some form of a sliding scale)
that is typically carved out of the WI share. An entity who owns an overriding royalty interest may
claim entitlement of resources that correspond to its proceeds from sales. Conversely, if an entity
has WI in projects that also have an overriding royalty interest, the related quantities must be
excluded from the entity’s entitlement.
333
Fig. 12.11—Example of entitlement with royalties in a concession contract for an oil-producing field.
Fig. 12.12—Example of entitlement with payments termed royalties that are considered to be
production taxes in a concession contract for an oil-producing field.
12.6.2.1 Example: Reserves Estimates in a Concession Contract with Royalties. SUN OIL
Co. entered into an onshore concession contract to explore, develop, and produce hydrocarbons in
Block A in Country ABC in a government bid, with 100% WI. Block A is located close to several
other producing fields. In Country ABC, hydrocarbon resources are owned by the ABC Federal
Government, which grants producing companies the right to explore, develop, and produce
hydrocarbons under concession contracts for 25 years. The contractor bears all risk and costs for
334
exploration, development, and production and holds title to all produced resources during the
concession period.
Over a 2-year period, SUN OIL Co. explored Block A and drilled two exploratory wells,
which discovered oil in commercial quantities. The company conceived and approved a project to
produce the already drilled wells and two additional wells within the proved area, and the company
included this project in the field development plan, attending to all commercial considerations.
Production will start in year 3 and will be transferred to a nearby facility, from which it will be
sold to local market. In this simplified example case, SUN OIL Co. oil and gas forecast prices are
USD 45/STB and USD 10 /Mcf for all future years, respectively.
12.6.2.1.1 Royalties Are Paid in Cash, with Tax Attributes. The contractor is subject to a cash-
only termed royalty payment to the government, which has tax attributes. For this contract, those
termed royalties are calculated as 17% of total produced hydrocarbons (including gas consumed in
operations and any gas loss) multiplied by a government-established tax price. SUN OIL Co. oil and
gas forecast prices for those payments are USD 60/STB and USD 14/Mcf for all future years,
respectively. Table 12.4 shows SUN OIL Co. sales production forecast and cash flow for the best
estimate case, using SUN OIL Co. forecast prices. Since the termed royalty payments have tax
attributes, Sun OIL Co. does not exclude quantities related to such payments from its forecast, and
it includes the related payments as an expense in the cash flow. (Note that the gas volume used for
“Future Payments Termed Royalties” includes quantities consumed in operation and losses, which
are reflected in the “Gas for Payments Termed Royalty” column used to calculate the royalties.)
BEST ESTIMATE
Future Gas for
Future Operating
Oil and Future Payments Cumulative Payments
Sales Gas Revenue Cost Future NCF
Year Condensate Investment termed Future NCF termed Royalty
(million CF) (million USD) (million USD) (million USD)
(thousand STB) (million USD) Royalties (million USD) Calculation
(*)
(million USD) (million CF)
1 4,50 -4,50 -4,50
2 0,50 -0,50 -5,00
3 230,53 162,20 12,00 3,27 2,89 5,84 0,84 227,07
4 119,67 84,20 6,23 1,94 1,50 2,79 3,63 117,88
5 72,24 50,82 3,76 1,37 0,91 1,49 5,11 71,15
6 47,82 33,65 2,49 1,07 0,60 0,81 5,93 47,11
7 33,16 23,34 1,73 0,90 0,42 0,41 6,34 32,68
8 23,63 16,63 1,23 0,78 0,30 0,15 6,49 23,29
9 17,35 12,20 0,90 0,71 0,22 -0,02 6,47 17,08
10 12,74 8,97 0,66 0,65 0,16 -0,15 6,31 12,56
11 8,62 6,06 0,45 0,59 0,11 -0,25 6,06 8,48
12 6,63 4,66 0,35 0,58 0,08 -0,32 5,74 6,53
Table 12.4—SUN Oil Co. sales production forecast and cash flow calculation, where the future operating cost (*) column
includes applicable severance and ad valorem taxes amounts, and NCF indicates net cash flow.
The economic limit for 2P Reserves is in year 8, at the maximum undiscounted cumulative
cash flow, considering incremental costs, as shown in Fig. 12.13.
As SUN OIL Co. does not include gas consumed in operations in its reserves estimates, best
estimate (2P) reserves are estimated by adding forecast oil and sales gas production from year 3 to
year 8, as shown on Table 12.5. If SUN OIL included fuel gas as reserves, forecast quantities of
fuel gas from years 3 to 8 would also be included as gas reserves.
The 1P and 3P reserves are estimated similarly. In this example, the royalty payments are
considered to be similar to production taxes, and they are not considered to be royalty interest.
Therefore, they do not reduce the yearly contractor entitlement, which is 100% WI. Such cash
royalty reduces the reserves cash flow of SUN OIL Co., but entitlement and revenue of all sales
production remain with SUN OIL Co. If those cash payments were considered to be a true royalty
335
interest, then reserves entitlement would be estimated similarly to the method described in Section
12.6.2.1.2 (following section).
Fig. 12.13—SUN Oil Co. cash flow and economic limit for the best estimate, where cash royalty has tax attributes, and NCF
indicates net cash flow.
YEAR 1 2 3 4 5 6 7 8 2P RESERVES
Oil and Condensate

0,00 0,00 230,53 119,67 72,24 47,82 33,16 23,63 527,06
(thousand STB)
Sales Gas
0,00 0,00 162,20 84,20 50,82 33,65 23,34 16,63 370,85
(thousand CF)
Table 12.5—2P Reserves.
As mentioned in Section 12.6.2, cases of cash royalties should be carefully analyzed to

determine the nature of the payment. Each situation will need to be carefully analyzed.
12.6.2.1.2 Royalties Are Paid in Kind. In this case, the contractor bears all risk and costs for
exploration, development, and production and holds title to all produced resources during the
concession period, except for royalties sent to the NOC. Royalty shares are specified in each
contract, and in the case of Block A, the royalty is calculated as 17% of sales quantities.
Table 12.6 shows SUN OIL Co. gross and net-of-royalties sales production forecast for the best
estimate, using SUN OIL Co. forecast prices.
Revenue is calculated based on net-of-royalties quantities because SUN OIL Co. is not
entitled to the royalties share of quantities. In this case, there is no royalty expense in the illustrated
cash flow. The economic limit for 2P Reserves is in year 9, at the maximum undiscounted
cumulative cash flow, considering incremental project costs, as shown in Fig. 12.14. The economic
limit is different from the situation in which the royalties were paid in cash and had production
taxes characteristics, which was in year 8.
As SUN OIL Co. does not include gas consumed in operation in reserves estimates, 2P
Reserves are estimated by adding forecast oil and sales gas production from year 3 to year 9, as
shown on Table 12.7. If SUN OIL included fuel gas as reserves, then forecast quantities of fuel
gas from years 3 to 9 would also be included as gas reserves.
336
BEST ESTIMATE
GROSS QUANTITIES NET QUANTITIES
Future
Future Operating Cumulative
Oil and Oil and Revenue Future NCF
Sales Gas Sales Gas Investment Cost Future NCF
Year Condensate Condensate (million USD) (million USD)
(million CF) (million CF) (million USD) (million USD) (million USD)
(thousand STB) (thousand STB)
(*)
1 4,50 -4,50 -4,50

2 0,50 -0,50 -5,00
3 230,53 162,20 191,34 134,62 9,96 3,27 6,69 1,69
4 119,67 84,20 99,33 69,89 5,17 1,94 3,23 4,92
5 72,24 50,82 59,96 42,18 3,12 1,37 1,75 6,68
6 47,82 33,65 39,69 27,93 2,07 1,07 0,99 7,67
7 33,16 23,34 27,53 19,37 1,43 0,90 0,53 8,20
8 23,63 16,63 19,62 13,81 1,02 0,78 0,24 8,44
9 17,35 12,20 14,40 10,13 0,75 0,71 0,04 8,48
10 12,74 8,97 10,57 7,45 0,55 0,65 -0,10 8,38
11 8,62 6,06 7,15 5,03 0,37 0,59 -0,22 8,15
12 6,63 4,66 5,51 3,87 0,29 0,58 -0,29 7,86
Table 12.6—SUN Oil Co. sales production forecast and cash flow calculation, where future operating cost (*) column
includes applicable severance and ad valorem taxes amounts, and NCF indicates net cash flow.
Fig. 12.14—SUN Oil Co. cash flow and economic limit for the best estimate, with royalties paid in kind, where
NCF is net cash flow.
YEAR 1 2 3 4 5 6 7 8 9 2P RESERVES
Oil and Condensate

0,00 0,00 191,34 99,33 59,96 39,69 27,53 19,62 14,40 451,86
(thousand STB)
Sales Gas
0,00 0,00 134,62 69,89 42,18 27,93 19,37 13,81 10,13 317,93
(million CF)
Table 12.7—2P Reserves.
The 1P and 3P reserves are estimated similarly. In this example, royalties are paid in kind to
the ABC Federal Government and are excluded from the contractor’s reserves quantities.
12.6.3 Unitization Agreements. Petroleum accumulations straddling block contracts or license

boundaries are common. Unitization is the process by which licensees of hydrocarbon leases or
contracts pool their individual interest in return for an interest in the overall unit. Each participant
involved in the unitization can have a different equity interest. Fig. 12.15 shows an example of an
accumulation that straddles block contracts.
337
Fig. 12.15—Accumulation straddling block contracts.
Unitization has its origins in the US, where hydrocarbon accumulations can extend across
multiple privately owned land parcels or mineral rights boundaries/leases related to those lands. In
the early days of the petroleum industry, the existing “rule of capture” doctrine acknowledged
ownership of production by the landowner/mineral owner where the petroleum was produced,
resulting in a rush to drill on adjacent spaces to prevent one’s hydrocarbons from being drained
from under one’s property/ownership. This certainly had a profound effect on the pace of drilling
in the US, including hydrocarbon drainage, production inefficiencies, and reservoir damage
through overproduction. Subsequent development of US state-level legislation often required
mineral rights owners to form a unit in petroleum development to promote “efficient and effective”
production from the reservoir, avoid those harmful effects, and yield a fair outcome. Unitization
is now applied throughout the world where similar circumstances prevail, while in some
countries/areas, the rule of capture may still prevail. Generally, tracts to be unitized have proved
the existence of the underlying petroleum resources, by drilling a well.
The act of forming this unit, i.e., unitization, is usually formalized by creating a unitization
agreement (UA) or unit operating agreement between the mineral rights holders or mineral interest
holders. This legal structure sets out the boundaries of the unit, describes the separate ownership
by area (“tracts”), sets allocated ownership, and establishes how production operations are to be
managed. The agreement typically contains clauses naming the party who will operate the unit
(unit operator) and describing how hydrocarbon quantities and expenses will be shared among
tracts making up the unit (termed tract participation or TP). Note that each tract may have multiple
owners, and the TP would be apportioned according to ownership level in the original tract.
Treatment of prior costs, provisions in the case of redetermination of TP, and other key elements
of cooperation and dispute may also be addressed in the agreement.
Countries or states may have laws and regulations that set rules for the terms that should be
followed by all parties, including “triggering” events, such as cash calls for development capital
and required approvals for implementation. Unitization can also occur in accumulations that cross
international borders, invoking use of international laws and agreements. Model contracts can also
contain unitization clauses. Depending on the country and circumstances, more than one document
may be needed to regulate the unit operation agreements
Arriving at the point of signing a UA or unit operating agreement may require significant data
and technical study by the TPs, in addition to legal and regulatory activities, and the underlying
uncertainties can lead to a long period of negotiations. Typically, an agreement will provide for an
initial TP, to be followed by one or more redeterminations. As tracts may have different participant
entities, a unit interest (for an ownership entity) is usually determined by multiplying the party’s
participating interest in each tract by that tract’s interest (TP) in the unit.
338
Common parameters used for determining TP are: petroleum initially in place, hydrocarbon
pore volume, estimated ultimate recovery, or net present value, but they can also include other
factors, such as well productivity, well density, cumulative production, and/or surface acreage. The
idea is to allocate the holdings in the unit by a reasonable facsimile of value or economic interest.
Reserves and resources entitlement under the PRMS is based on economic interest in the
forecast production. The reserves and resources that are associated with each participant and party
to the unitization are not necessarily just the product of the production forecast of the accumulation
and determined TP (which shall represent the future split of production and expenses related to the
unitized reservoir). Reserves and resources estimates will consider other factors, such as each
tract’s contract model, status, and expiration date. As for any other project, the entitled future
production is subject to commerciality determination to be classified as reserves. Forecast prices,
estimated development and operating costs, market infrastructure, and other factors may result in
different commerciality considerations among participating entities in the unitized reservoir. Of
course, each entity may have its own perception of uncertainties, which may result in different
resources categorization among the participating entities. Additionally, the existence of other non-
unitized reservoirs in a given tract would be appropriately considered in reserves estimates by
those entities having an interest in them.
When estimating reserves and resources for a reservoir that may become unitized, an evaluator
must also consider the unitization approval and effective dates because they affect the forecast,
which may also be impacted by the venue regulatory framework, specific UA status and clauses,
required outstanding approvals, history of similar unitizations, and any other document related to
the unitization process. There may be a need for assumptions of entitlement before any agreement
is signed or approved, and such an estimate of unit entitlement would be based on such
considerations. Of course, entitlement of reserves and resources would be modified and adjusted
to reflect actual unitized conditions when known.
Where costs have been incurred by one or more parties to the unitization prior to the effective
date of unitization, the handling of those costs are typically subject to make-up provisions included
in the UA. These costs may be recovered out of future production, or there may be an agreed-upon
monetary settlement. In cases where cost reconciliation is received from future production, the
entities on each side of the reconciliation may choose to report the future volume as a reduction
(those that did not incur the pre-unitization costs) or an addition (those that incurred the pre-
unitization costs and are being reimbursed) to their respective reserves or resources entitlement, if
contracts, agreements, and regulatory framework permit. This scenario may include unitization of
not-yet-contracted areas or a tract held by the government, whose costs are carried by other
participants, which may be partly or wholly settled with future production to the contractor.
When a unitization or a redetermination results in adjustments to compensate for pre-event
splits (including provisional unitization splits) that are changed, parties on each side of the
agreement may receive or grant future production allocations to achieve balance between entities.
In this case, the receiving entity must add reserves/resources entitlement in the appropriate amount.
Accordingly, where part of an entity’s future production is reduced to provide make-up volumes
to other participants, that entity must reduce its reported entitlement reserves or resources by the
amount by which the future production is reduced.
In certain cases, the tract that paid for past expenses may be under a PSC or similar agreement,
and contracts, agreements, and regulatory framework may allow such expenses to be included as
cost petroleum. Depending on contracts and regulation, care would be necessary to ensure that the
entity did not receive double credit for booking cost petroleum (associated with the past costs) and
339
then booking future production offsets. If the tract under a PSC or similar agreement did not pay
for past expenses, attributed past expenses may be included as cost petroleum, depending on
contracts and regulation.
12.6.3.1 Example: Reserves Estimates with Unitization of Accumulation in Two
Concession Areas after a Unitization Agreement. SUN OIL Co. and MOON OIL Co. entered into
an onshore concession contract with a local government to explore, develop, and produce
hydrocarbons in Block 36. Each company has 50% interest in the block. After a 2-year exploration
period, which included two exploratory wells and the acquisition of seismic data, they discovered
oil in commercial quantities and that the accumulation extended to Block 37, recently leased from
the local government by STAR OIL Co. (80% WI) and COMET OIL Co. (20% WI), as shown in
Fig. 12.16. According to government regulations, the companies in Blocks 36 and 37 are
responsible to pay cash payments termed royalties to the government, which have production tax
characteristics. Those cash payments are calculated as 10% of all produced oil and gas, multiplied
by a tax price established by the government. In this simplified example case, SUN OIL Co. oil
and gas forecast sales prices are USD 50/STB and USD 8/MCF for all future years, and forecast
royalty prices are USD 55/STB and USD 10/MCF for all future years.
BLOCK 37 BLOCK 36
80% STAR OIL Co 50% SUN OIL Co
20% COMET OIL Co 50% MOON OIL Co
Fig. 12.16—Accumulation that extends across Block 36 and Block 37.
As the accumulation straddles the concession contract boundaries, after negotiation, all
companies signed a UA in which they defined the TP based on estimated total petroleum initially
in place from a P50 model of the accumulation in each lease area, with a possibility of future
redetermination. The resulting TP was 70% (Block 36) and 30% (Block 37). Unit interests (UIs)
were obtained by multiplying each company’s WI in the block by the block TP, as shown in
Table 12.8.
Block Company WI (%) TP (%) UI (%)

SUN OIL Co 50% 35%
36 70%
MOON OIL Co 50% 35%
STAR OIL Co 80% 24%
37 30%
COMET OIL Co 20% 6%
Total 100% 100%
Table 12.8—Tract participation and unit participation for the accumulation.
The UA included the development plan for the accumulation, which was approved by the
local government entity. SUN OIL Co. was specified as the operator of the accumulation. The
development plan included a project with drilling of two additional wells within the proved area,
as shown in Fig. 12.17, which was approved by all four companies and attended to all commercial
considerations.
340
BLOCK 37 BLOCK 36
80% STAR OIL Co 50% SUN OIL Co
20% COMET OIL Co 50% MOON OIL Co
Fig. 12.17—Development wells in unitized accumulation.
Table 12.9 shows the SUN OIL Co. production forecast based on sales quantities for the low
estimate case, with each company share according to the respective approved UIs. In this example,
there are no royalties that diminish the reserves entitlement because the termed cash royalties were
interpreted as production taxes. In the cases where royalties are payable or otherwise excluded
from the contractor’s entitlement, the split among participants of the unitized accumulation would
probably be based on gross quantities (before deducting royalties), unless specified differently by
unitization agreements.
LOW ESTIMATE FORECAST
SALES QUANTITIES SUN OIL Co (35%) MOON OIL Co (35%) STAR OIL Co (24%) COMET OIL Co (6%)
Oil Sales Gas Oil Sales Gas Oil Sales Gas Oil Sales Gas Oil Sales Gas
Year
(thousand STB) (million CF) (thousand STB) (million CF) (thousand STB) (million CF) (thousand STB) (million CF) (thousand STB) (million CF)
3
4
5 461,07 324,39 161,37 113,54 161,37 113,54 110,66 77,85 27,66 19,46
6 239,35 168,41 83,77 58,94 83,77 58,94 57,44 40,42 14,36 10,10
7 144,47 101,64 50,57 35,57 50,57 35,57 34,67 24,39 8,67 6,10
8 95,64 67,30 33,47 23,56 33,47 23,56 22,95 16,15 5,74 4,04
9 66,33 46,68 23,21 16,34 23,21 16,34 15,92 11,20 3,98 2,80
10 47,27 33,27 16,54 11,64 16,54 11,64 11,34 7,98 2,84 2,00
11 34,70 24,40 12,15 8,54 12,15 8,54 8,33 5,86 2,08 1,46
12 25,47 17,95 8,92 6,28 8,92 6,28 6,11 4,31 1,53 1,08
13 17,23 12,12 6,03 4,24 6,03 4,24 4,14 2,91 1,03 0,73
14 13,27 9,33 4,64 3,26 4,64 3,26 3,18 2,24 0,80 0,56
. 15 10,43 7,34 3,65 2,57 3,65 2,57 2,50 1,76 0,63 0,44
Table 12.9—SUN OIL Co. low estimate production forecast of the accumulation.
In general, in actual situations, each company will apply its own uncertainties for production
forecasts in the low, best, and high cases and will consider them to classify and categorize reserves
and resources. For example, this would mean that MOON OIL Co., STAR OIL Co., and COMET
OIL Co. would estimate low, best, and high case forecasts differently from SUN OIL Co., and they
would multiply them by the approved UI to obtain that company’s low, best, and high case forecast
share, resulting in different forecast quantities from those estimated by SUN OIL Co. as their
forecast share. For example, MOON OIL Co.’s low estimate forecast will probably be different
from the value shown as the MOON OIL Co. share column in Table 12.9. Each company will most
likely use its own forecast to estimate reserves and resources and apply other commerciality
constraints to its share (i.e., prices, royalty rates, contract constraints, etc.). According to the UA,
companies will share costs based on their TP and UI, and costs incurred before the unitization
agreement was effective will be settled with future sales production also based on their TP and UI,
with a provision to limit this settlement to 25% of Block 37 contractor annual revenue.
Table 12.10 shows the SUN OIL Co. forecast of investment and operating costs for the low
estimate case according to each company’s UI.
341
LOW ESTIMATE
SUN OIL (35%) MOON OIL (35%) STAR OIL (24%) COMET OIL (6%)
FORECAST
Investment Operating Cost Investment Operating Cost Investment Operating Cost Investment Operating Cost Investment Operating Cost
Year
(million USD) (million USD) (million USD) (million USD) (million USD) (million USD) (million USD) (million USD) (million USD) (million USD)
3 9,00 3,15 3,15 2,16 0,54

4 1,00 0,35 0,35 0,24 0,06
5 6,53 2,29 2,29 1,57 0,39
6 3,87 1,36 1,36 0,93 0,23
7 2,73 0,96 0,96 0,66 0,16
8 2,15 0,75 0,75 0,52 0,13
9 1,80 0,63 0,63 0,43 0,11
10 1,57 0,55 0,55 0,38 0,09
11 1,42 0,50 0,50 0,34 0,08
12 1,31 0,46 0,46 0,31 0,08
13 1,19 0,42 0,42 0,28 0,07
14 1,16 0,41 0,41 0,28 0,07
15 1,13 0,39 0,39 0,27 0,07
Table 12.10—Low estimate SUN OIL Co. forecast investments and operating costs.
According to the UA, the Block 36 contractor (i.e., SUN OIL Co. and MOON OIL Co.) is
entitled to extra produced oil and sales gas to settle past costs of USD 8 million, which they paid
entirely. Because the Block 36 TP is 70%, its contractor will receive 30% of this amount (USD
2.4 million, representing Block 37 TP percentage in past costs) out of future sales production from
the Block 37 contractors, which will be added to their Block 36 entitled resources. According to
the UA, production costs and production taxes (including the cash payments termed royalties)
related to the quantities used for settlement of past costs remain with the Block 37 contractors.
Table 12.11 shows Block 37 oil and gas sales production that will be used to settle past costs,
according to the SUN OIL Co. low case forecast.
LOW ESTIMATE
Block 37 (30% of accumulation)
Oil to settle Sales Gas to
Limit for Recovered Unrecovered
Sales Gas Total past settle past
Oil Sales Gas Oil Revenue Recovery Cost past cost
Year Revenue Revenue investment investment
(thousand STB) (million CF) (million USD) (25% of (million USD) (million USD) (thousand STB) (million CF)
(million USD) (million USD)
Revenue)
3 2,40
4 2,40
5 138,32 97,32 6,92 0,78 7,69 1,92 1,92 0,48 34,58 24,33
6 71,80 50,52 3,59 0,40 3,99 1,00 0,48 8,56 6,03
7 43,34 30,49 2,17 0,24 2,41 0,60
Table 12.11—SUN OIL Co. low estimate of Block 37 production that will be used to settle costs incurred before the
unitization agreement became effective.
Settlement will begin in year 5, when production starts. Past costs attributed to Block 37
contractors will be settled with production in years 5 and 6. In year 5, part of entitled Block 37
contractor sales production is estimated to be used to settle past costs, limited to 25% of annual
Block 37 revenue (i.e., limited to 25% of 7.69 = USD 1.92 million). Therefore, SUN OIL Co.
estimates that 25% of Block 37 sales quantities in year 5 (25% of 138.32 thousand STB = 34.58
thousand STB and 25% of 97.32 MMCF of gas = 24.33 MMCF) will be used to settle past costs.
Since the total cost to be settled is USD 2.4 million, USD 0.48 million (USD 2.4 million – USD
1.92 million = USD 0.48 million) remains to be settled in future years. This amount may be
estimated to be settled entirely in year 6, because it does not reach the settlement limit (25% of
annual income of USD 3.99 million = USD 1.00 million, which is greater than USD 0.48 million).
342
Quantities used to settle past costs were estimated as a percentage of annual revenues that
correspond to cost recovered (0.48/3.99), applied to sales quantities. Therefore, SUN OIL Co.
estimates that 8.56 thousand STB of oil (71.80 × 0.48/3.99) and 6.03 MMCF of sales gas (50.52 ×
0.48/3.99) in year 6 will be used by Block 37 to settle past costs. SUN OIL Co. and MOON OIL
Co. are each entitled to 50% of oil and gas estimates shown in Table 12.11.
Table 12.12 shows SUN OIL Co.’s share of the sales production forecast for the low estimate
case, as well as the related future cash flow using SUN OIL Co. forecast prices, including entitled
production received to settle past costs. There are no royalties due (and no royalty interest quantities
excluded from the contractor), as the termed royalty cash payments have production tax
characteristics. SUN OIL Co. is responsible for such cash payments related to its entitled quantities,
and it includes them as an expense in the cash flow. Production costs and production taxes (including
the cash payments termed royalties) related to the quantities used for settlement of past costs are not
expensed in the SUN OIL Co. cash flow, as they remain with the Block 37 contractors.
SUN OIL Co LOW ESTIMATE
PRODUCTION ENTITLED
TOTAL ENTITLEMENT FUTURE CASH FLOW (million USD)
ENTITLED FROM UI (35%) FROM SETTLEMENT
Cash Payments
Oil Sales Gas Oil Sales Gas Oil Sales Gas Operating termed Cumulative
Year Revenue Investment NCF
(thousand STB) (million CF) (thousand STB) (million CF) (thousand STB) (million CF) Cost Royalties & Ad NCF
Valorem Taxes
3 3,15 -3,15 -3,15
4 0,35 -0,35 -3,50
5 161,37 113,54 17,29 12,16 178,66 125,70 9,94 2,29 1,02 6,63 3,13
6 83,77 58,94 4,28 3,01 88,05 61,95 4,90 1,36 0,53 3,01 6,14
7 50,57 35,57 50,57 35,57 2,81 0,96 0,32 1,54 7,68
8 33,47 23,56 33,47 23,56 1,86 0,75 0,21 0,90 8,58
9 23,21 16,34 23,21 16,34 1,29 0,63 0,15 0,52 9,09
10 16,54 11,64 16,54 11,64 0,92 0,55 0,11 0,27 9,36
11 12,15 8,54 12,15 8,54 0,68 0,50 0,08 0,10 9,46
12 8,92 6,28 8,92 6,28 0,50 0,46 0,06 -0,02 9,44
13 6,03 4,24 6,03 4,24 0,34 0,42 0,04 -0,12 9,33
14 4,64 3,26 4,64 3,26 0,26 0,41 0,03 -0,18 9,15
15 3,65 2,57 3,65 2,57 0,20 0,39 0,02 -0,21 8,93
Table 12.12—SUN OIL Co. share of sales production forecast and cash flow for the low estimate, where NCF is net cash flow.
SUN OIL Co. revenue is calculated by multiplying oil and gas prices by its total entitled sales
production forecast (i.e., its UI share plus production received to settle past costs). The economic
limit for 1P Reserves is in year 11, at the maximum undiscounted cumulative cash flow,
considering incremental project costs. As SUN OIL Co. does not include gas consumed in
operation in its reserves estimates, 1P Reserves are estimated by adding sales production forecast
gas from year 5 to year 11, as shown in Table 12.13.
YEAR 3 4 5 6 7 8 9 10 11 1P RESERVES
Oil and Condensate (thousand STB) 0,00 0,00 178,66 88,05 50,57 33,47 23,21 16,54 12,15 402,66
Sales Gas (million CF) 0,00 0,00 125,70 61,95 35,57 23,56 16,34 11,64 8,54 283,31
Table 12.13—SUN OIL Co. 1P Reserves estimates.
SUN OIL Co. 2P and 3P Reserves are estimated similarly, considering best estimate and high
estimate forecasts. MOON OIL Co. reserves entitlement will be estimated with a similar cash flow
calculation, but it will probably use its own defined conditions (i.e., prices and costs) and other
constraints for estimated quantities to settle past cost, commerciality determination, economic
determination, and economic limit determination for the project, which might result in different
reserves and resources quantities, even having the same UI as SUN OIL Co. Both companies may
343
also consider different uncertainties for the low, best, and high cases, which would also result in
different reserves and resources quantities for 1P, 2P, and 3P Reserves and 1C, 2C, and 3C
Contingent Resources.
STAR OIL Co. and COMET OIL Co. will exclude forecast sales quantities used to settle past
costs from their cash flows and reserves and resources entitlement. They may also have different
contract models and their own defined conditions and uncertainties with which to estimate reserves
and resources quantities for 1P, 2P, 3P, 1C, 2C, and 3C. In addition, their contract (Block 37
contract) may have different royalties, production taxes, or other conditions to consider for their
entitled reserves and resources quantities. In an actual situation, Blocks 36 and 37 might also
include other non-unitized reservoirs, which would be considered in the resources estimates. Each
situation would need to be carefully analyzed to evaluate each contract model, the UA terms, the
possibility to share costs with other accumulations or specific block-related issues, and each
company’s defined conditions and uncertainties for reserves and resources estimates.
12.6.4 Contract Extensions or Renewals. As described in the PRMS § 3.3.3.1, reserves cannot
be claimed for those quantities that will be produced beyond the expiration date of the current
production, concession, or license agreement unless there is reasonable expectation that an
extension, a renewal, or a new contract will be granted. In the absence of such a reasonable
expectation, estimated future production beyond the contract term may be classified as Contingent
Resources, if the possibility exists that the terms of the agreement can be extended/renewed. This
does not result in a split classification (PRMS § 2.2.0.4), because the Contingent Resources
associated with the Reserves constitute a separate project in this context. As a separate project, a
decision is likely to have to be made and approved. This may include decisions about a substantial
bid amount to obtain/continue a concessionary arrangement and required investments that would
be made, before or after the present contract expires, to produce beyond the present contract
expiration date. When no extension or renewal of an existing agreement nor an assurance of a new
agreement appears to be possible, estimated future production should be classified as
unrecoverable resources by the contractor.
An additional uncertainty to be considered when future estimated quantities are being
evaluated is whether it is reasonable to assume that the fiscal terms in a negotiated
extension/renewal will be similar to existing terms. Differences in the agreed-upon terms could
impact gross and net entitlement quantities.
Support for the reasonable expectation of extension or renewal would typically be based on
the (1) intent of the interest holder and host government, (2) status of renewal negotiations, (3)
historical treatment of similar transactions by the license-issuing jurisdiction, (4) legislation, and
(5) the specific contract terms, which may also contain clauses about the possibility and
requirements for contract extension or renewal. Reserves associated with production beyond the
expiration date of the current agreement must comply with the other commerciality requirements
of the PRMS § 2.1.2, commensurate with the project to produce those quantities.
Similar logic should be applied for gas sales agreements where market demand is not
sufficient to support commerciality for all producible volumes. Reserves should not be claimed for
estimated future production quantities that would be produced beyond those amounts specified in
the current agreement or that do not have a reasonable expectation to be included in either contract
renewals or future agreements.
12.6.4.1 Example: Considering Recovery Estimates After Contract Expiration Date. A
contractor produces a field under a concession contract that is initially granted for 20 years, which
344
will expire in 6 years. The concession contract has clauses specifying the requirements to extend
the contract for 20 years from the original expiration date. The contractor has formally requested
the extension and has already attended to its requirements. In similar requests, the local license-
issuing jurisdiction has granted extension, without undue delay. There are no known contingencies
that would prevent the contract extension.
The contractor’s recoverable technical forecast of existing wells is expected to last 23 years
from the effective date of the evaluation, and the project’s economic limit will be reached in 16
years from the effective date (i.e., 10 years into the extension), if the extension is granted.
The contractor, under these circumstances, has a reasonable expectation that an extension will
be granted, because there are no contingencies for the extension, and the local license-issuing
jurisdiction historically granted the extension. The production forecast associated with the
producing wells may be classified as Reserves, for 16 years (until the economic limit), if the other
commerciality requirements of the PRMS § 2.1.2 are met. Production after the economic limit in
year 16 may be classified as Contingent Resources (as noted previously in Section 12.5.1).
Fig 12.18 displays the contractor’s resources classification for the field.
Fig. 12.18—Contractor’s resource classification.
12.6.5 Appropriate Date on Which to Recognize Entitlement for Reserves Estimates in

Specific Cases. There are situations where there is uncertainty about the appropriate date on which
to consider and potentially recognize entitlement of reserves. For example, in situations where
contracts or agreements are signed well in advance of the project’s effective date, when should
entitlement be applied? Farm-in and farmout agreements, unitization agreements, and other
pending circumstances also may be unclear as to the particular date of entitlement application.
For reserves estimates, there must be evidence that legal, contractual, environmental, and
government approvals are in place or will be forthcoming, together with solution of any social and
economic concerns (see PRMS § 2.1.2). There must exist a reasonable expectation for all such
approvals, which is a requirement for all activities to implement the project and begin production.
When contracts or agreements are signed before the effective date begins, and there are no
further required approvals or circumstances that might prevent it from being effective at the signed
terms and dates (e.g., environmental approvals), entitlement of reserves may be recognized when
the contract is signed (if other commerciality requirements are met). The reserves estimated would
include those quantities that will be produced from the contract effective date and term.
When external approvals or specific circumstances are required to exist or come to pass, historical
approvals and comparison with similar transactions may be considered to determine the reasonable
345
expectation for entitlement recognition date. Required approvals or circumstances that might
represent a meaningful possibility that the terms of the contract or agreement may change
(circumstances not matching the terms signed or that might postpone the effective date
significantly) may suggest caution in assuming the anticipated terms and dates will be applicable.
In such a circumstance, it might be appropriate to wait for clearance of the approvals or
circumstance milestones before recognizing the impact on reserves estimates.
This same logic may apply for farm-in and farmout arrangements. There is often a need for
regulatory approvals after a farm-in or farmout contract is signed, which may prevent the
transaction from actually occurring or may postpone the effective date significantly. In cases where
such approvals or other circumstances are required, depending on historical similar transactions,
it may be appropriate to wait for all necessary approvals and for the specific outstanding matters
to be cleared up in order for the grantor to reduce its reserves and for the grantee to book reserves.
Similar issues may occur when estimating reserves for unitized accumulations where the TP
is not yet finally agreed upon or approved. Depending on government regulation and
circumstances, it is sometimes desirable (for more accurate depiction of future reserves) to
estimate the TP or the effective date to begin the production split. In such cases, care should be
taken to prevent recognizing significant (particularly Proved) reserves that might be subject to later
reduction due to the actual terms and effective date of the unitization agreement once fixed by
formal approval and agreement.
Judgement must also be applied when more than one related transaction involving reserves
exists (e.g., pooling of assets). In such cases, the effects on reserves entitlement should represent
the related transactions applied simultaneously.
12.7 Example Case

12.7.1 Reserves in a PSC. The following example illustrates the approach used to estimate
reserves and resources under a PSC. In this example, the contractor develops and operates the field
and is entitled to a share of production that is based on cost recovery and profit-share components.
The contractor takes their share of product in kind. The contractor does not have ownership of the
underlying resources being produced but does earn an economic interest by virtue of the exposure
to technical, financial, and operational risks and is therefore able to recognize reserves and
resources for the project under the PRMS. Due to the difficulty in predicting prices, this example
uses a constant forecast oil price of USD 60/STB. Sensitivity cases of USD 10/STB above and
below this price are also shown in the example. Although unlikely to represent the actual forecast
prices in effect, they do provide a good illustration of how entitlement and contract terms respond
to price changes.
12.7.1.1 Example Description. An oil field has a technically recoverable resource estimate of
500 million STB, of which 400 million STB are estimated to be produced during the PSC term,
and, for the purposes of this example, the project is considered to be commercial, and all quantities
are reflected in the Proved Reserves category during the PSC term, assuming a discovery. The
contract provides for an initial exploration period, with the contract term lasting 20 years from the
start of production. The general field data are summarized in Table 12.14.
The production forecast and full-life cost summary for the PSC term are shown in Table 12.15.
The remaining 100 million STB are classified as Contingent Resources, assuming a discovery,
related to a potential contract extension. In this simplified example, no additional drilling is required,
and Probable and Possible Reserves are not shown.
346
Table 12.14—Example general field data, where TRR indicates technically recoverable resources, and EUR is estimated
ultimate recovery.
Table 12.15—Project production cost and cost schedule. Figures may not match due to rounding; Devt is development,
and OP is operating.
Production startup is midyear in the second year of the project and builds to a peak rate of
95,000 BOPD (34.7 million STB annualized) in the eighth year.
12.7.1.2 PSC Terms. The example contract contains many common contractual terms
affecting the industry today. These include royalty payments, limitations of revenue available for
cost recovery, a fixed profit-share split, and income taxes. For simplicity, the example does not
consider a variable profit share. It is a typical PSC in which the contractor is responsible for the
field development and all exploration and development expenses. In return, the contractor recovers
investments and operating expenses out of the gross production stream after royalty deduction and
is entitled to a share of the remaining profit oil. The contractor receives payment in oil production
and is exposed to both technical and market risks.
347
Fig. 12.19 shows the general terms of the contract, whereby taxes are paid by the contractor.
The contract is for a 20-year production term with the possibility of an extension. The terms
include a royalty payment in kind on gross production of 15%. Yearly cost recovery is limited to
a maximum of 50% of the annual gross revenue after royalty deduction, with the remaining cost
carried forward to be recovered in future years. The contractor’s profit share is based on a simple
split: 20% to the contractor and 80% to the host government.
Crude Price USD 60 /STB
100 Barrels
Contractor Governm ent
Royalty 15% 15.0
22.6** Cost Recovery (50% Lim it )
12.5 Profit Oil Split (20% / 80%) 49.9

35.1 Subtotal 64.9
Tax Paid (50%)*

$$$
* Figure shows terms when taxes are not paid by the gov ernment, NOC or other gov ernment related company in behalf of
the contractor
** Contract av erage Cost Oil at USD 60/STB: Contract forecast recoverable cost/crude price = USD 5,420 million / U$D 60
= 90.33 million STB. This represents 22,58% (90/400) of EUR under the contract, or an average of 22.6 STB for each
100STB
Fig. 12.19—Production sharing contract terms whereby taxes are paid by the contractor, where EUR is estimated ultimate
recovery.
The contractor’s share of reserves and resources will be estimated in the following sections,
followed by sensitivity cases for the effects of price and alternative tax treatment on entitled reserves.
12.7.1.3 Contractor Entitlement and Reserves Estimates. The terms of a PSC determine the
contractor’s yearly entitlement or share of the project production based on the yearly cost recovery
and profit split. Table 12.15 shows the forecast production, investment, and cost profiles for the
project. The calculation of the contractor’s entitlement for each year at the forecast price of USD
60/STB is shown in Table 12.16. Looking at year 8 as an example, the gross revenue from 34.7
million STB produced is USD 2,083 million. At a royalty rate of 15%, the government would
receive as royalty 5.2 million STB valued at USD 312 million. The remaining USD 1,771 million
would remain for cost recovery and profit split according to the terms of the contract. In this PSC,
revenue available for cost recovery is limited to 50% after royalty amount, or USD 886 million.
Costs and expenses for the year total USD 248 million (assumed to be fully recoverable under the
agreement), and there are no costs carried forward from previous years. These costs are forecast
to be fully recovered in the same period. In the case of unrecovered costs, they would be carried
forward by the contractor for recovery in future years. The remaining revenue after royalty and
cost recovery is shared by the contractor and government according to the contract profit split. In
348
this case, the contractor’s profit share is USD 305 million, or 20%. The contractor’s total entitled
revenue is the sum of the contractor’s cost recovery and profit share.
Table 12.16—Project cost and profit-share schedule. Figures may not match due to rounding.
The estimated average contractor cost plus profit share value in year 8 is USD 553 million,
about 27% of the project gross revenue. This 27% share is converted to an equivalent volume of
the production of 9.2 million STB, representing the contactor’s entitlement in year 8. Because the
cost and revenue vary yearly, each estimated entitlement percentage applies only to the year in
question. This calculation provides only the contractor’s share of the annual production for the
year in question. After discovery, as the project is commercial, and all example production forecast
is categorized as Proved, the contractor’s Proved Reserves are estimated by the summation of the
estimated annual volume entitlements over the production forecast years of the project, up to the
contract term limit (economic and technical limit would occur after that date). Table 12.16 shows
the forecast entitlements from project initiation to the end of the contract term. They were
estimated with the forecasted production, exploration, development, and operating expenses and
schedules through the term of the agreement. After discovery and considering a commercial
project, the contractor’s Proved Reserves are estimated at 140 million STB, or 35% of the total
project Proved Reserves of 400 million STB.
The contractor is obligated to pay income tax on its profits, which in the example case
amounts to USD 152 million at the tax rate of 50% of its profit share in year 8 (50% of USD 553
million − USD 248 million). Income taxes were calculated as a percentage of yearly project
revenues minus project costs. In many actual situations, income tax calculations will consider other
deductions, percentages, or other data.
In this example, prices and profit splitting were held constant over the period, and the effect
of the recovery of initial capital investments on the effective net entitlement interest can be seen.
At the onset of production, entitlement (economic) interest is approximately 51% and declines
over the next several years to a low of 27% in year 8. The entitlement interest then increases to
37% by the end of the term. This increase is due to the natural decline in the production rate and
349
the need to have a greater portion of the production reimburse fixed operating costs. In general,
PSC entitlement percentages are highest at the point of first production and tend to decrease as a
project becomes cost current. Entitlements tend to increase as costs increase and prices decline;
however, many agreements contain “R” terms and/or stepwise tranches that tend to reduce the
profit share allocation to the contractor over time. These take many different forms, but they
generally tend to be related to cumulative production, cumulative reimbursements, or higher
production rates.
For the example entitlement calculation, price was considered to be constant during the year.
In actual situations, the crude price may vary over the year, and the method for calculating the
price for each settlement period is normally defined in the agreement.
In an actual field development, there will frequently exist low, best, and high cases with
corresponding production forecasts. Entitlement volumes will be estimated for each, which will
correspond to 1P, 2P, and 3P Reserves if entitled volumes are economic and the project satisfies
commerciality requirements. Those 2P or 3P entitled volumes may be sourced from portions of
the reservoir that are not considered Proved at the time of classification. Depending on technical
information, those quantities may be included in 1P Reserves in later assessments.
12.7.1.4 Crude-Price Sensitivity. In a PSC, contractor reserves are sensitive to the assumed
production schedule, price projections, and cost forecasts. Frequently, the most volatile of these
factors is the price. Table 12.17 demonstrates the inverse relationship between crude price and
contractor reserves in this example. For a USD 10/STB increase in crude price, the contractor’s
reserves decrease from 140 to 130 million STB. (It may be noted that, although the reserves
entitlement decreases as the price increases, the net production income per barrel also increases.)
Such swings in reserves can be expected when prices are volatile. Several other commonly used
financial metrics have also been included in Table 12.17 to illustrate how they also change with
price. Subject to specific pricing requirements in the PSC agreement, the ability to use average
prices over a year, as provided by the PRMS, can help to dampen price-related reserves changes.
The contractor’s actual ultimate recovery and resources will, however, be determined by prices
over the project life.
USD 50 Oil Price USD 60 Oil Price USD 70 Oil Price
Normal TPOB Normal TPOB Normal TPOB

Parameter Measured Tax Tax Tax
Reserv es (million STB) 155 178 140 165 130 156

Cost of Finding & Dev. (USD/STB) 11.63 10.12 12.83 10.89 13.85 11.52
Profit/STB (USD/STB) 14.97 19.53 21.36 27.20 28.29 35.30
Production Costs (USD/STB) 23.40 20.35 25.81 21.91 27.86 23.18
Net Production Income (USD/STB) 7.48 13.02 10.68 18.13 14.14 23.53
Table 12.17—Examples of financial metrics, reserves sensitivity to oil prices, and taxes [normal tax or tax paid on behalf
(TPOB)], where Dev is development.
12.7.1.5 Income Tax Paid on Behalf. In the normal case, as shown in Sections 12.7.1.2 to
12.7.1.4, the contractor is obligated to pay income tax out of their share of the project profit. In
such cases, the contractor’s tax obligation affects the project’s economics but has no impact on the
350
reserves estimates because reserves are calculated on a before-tax cash-flow basis. In some PSCs,
however, the government or state-owned oil company agrees to pay tax on behalf of the contractor.
If the tax payment is a purely financial arrangement, and the payments cannot be attributed to a
portion of the government’s production revenues, an economic interest would not exist; therefore,
no additional reserves would be recognized by the contractor. In this case, the carried tax reserves
will equal those obtained in the normal tax case shown in Table 12.17.
As mentioned in Section 12.6.1, if, under the terms of the contract, the contractor derives a
benefit from and has an economic interest in the government’s share of hydrocarbon volumes used
to fund the tax payments, those volumes may be considered as the contractor’s reserves. Fig. 12.20
shows the general terms of the contract if taxes are paid by the host government on behalf of the
contractor. Table 12.17 shows this impact on both the project financial indicators and the reserves.
The contractor’s cost recovery and profit share are computed in the standard fashion, but
entitlement of reserves would now include the economic benefit related to the taxes paid on behalf
of the contractor. With a tax-paid-on-behalf arrangement, the contractor’s Proved Reserves (at
USD 60/STB forecast price) would increase by 25 to 165 million STB. In an actual field
development, 2P and 3P Reserves would also have estimates associated with cost recovery, profit
share, and benefit related to the taxes paid on behalf of the contractor. In this example, all quantities
are considered to be commercial up to the PSC contract limit.
Crude Price USD 60 /STB

100 Barrels
Royalty 15% 15.0
22.6** Cost Recovery (50% Limit )
12.5 Profit Oil Split (20% / 80%) 49.9

35.1 Subtotal 64.9
6.2 Tax Paid on Behalf (50%)* 6.2

41.3 58.7
* Taxes may be paid by government on behalf of Contractors in some PSCs. Depending on specific terms, the payments
may be treated as a tax credit or a revenue gross-up. This figure shows when terms when taxes are part of profit share,
because the government pays in the name of the company
** Contract average Cost Oil at USD 60/STB: Contract forecast recoverable cost/crude price = USD 5,420 million/USD 60
= 90.33 million STB. This represents 22,58% (90/400) of EUR under the contract, or an average of 22.6 STB for each
100STB
Fig. 12.20—PSC terms when taxes are paid by the host government on behalf of the contractor, where EUR is
estimated ultimate recovery.
12.7.1.6 Common Procedure to Assess Reserves Sensitivity. The preceding reserves

estimates illustrate the general approach that can be used for PSCs and revenue-sharing contracts
at all levels of project maturity (risked-service contracts might use similar calculations). It accounts
for varying yearly investment levels and the relative relationship between project costs and project
revenue. In a mature project, with relatively stable prices and the relationship between project costs
351
and project revenues relatively constant, some companies simplify the process by assuming that
the reserves share is equal to an average entitlement percentage. In general, this approach is
believed to be sufficiently accurate, and corrections would be applied when accounts were adjusted
for actual production and realizations on the regular intervals prescribed in the agreement.
12.7.1.7 Assessing Other Classes and Categories of Reserves and Resources. In actual
field development, a portion of the production forecast would likely be categorized in the
Probable (and Possible) PRMS reserves category(ies), depending on supporting information and
technical certainty. Each project would normally have Proved, Probable, and Possible Reserves
components. For example, Probable (or Possible) Reserves may be estimated for better-than-
expected recovery or perhaps for undrilled areas where technical certainty was not sufficient to
classify the reserves as Proved. In this instance, modeling two cases, one for the Proved plus
Probable flow streams and a separate model for the Proved-only case, would give the Probable
Reserves entitlement by difference.
In this PSC example, 100 million STB were noted to be related to the potential extension of
the original contract agreement. If significant additional new investments were required to produce
this volume, and/or there was some doubt that the agreement would be extended, then the related
volume would most likely be categorized as Contingent Resource in one or more of the 1C, 2C, or
3C scenarios, depending on the level of technical certainty. There may also be a question of
whether the same or different terms would apply to the extension. Consequently, judgment must
be used when estimating the entitlement interest that will be applied to determine the net share of
the PRMS resources potentially available to the contractor.
In a different example scenario, if the 100 million STB were related to potentially higher
recovery efficiency from the reservoir within the original term, and no additional debottlenecking
or development investments were required, then the volume could be classified as Probable (and/or
Possible) Reserves (assuming appropriate technical certainty). To determine the effective net
interest for this Probable increment, a two-step process is commonly used. In the first step, the
Proved flow stream is evaluated using the PSC model described in the preceding sections. In the
second step, the forecast 2P (Proved plus Probable) flow stream and related costs are then
evaluated with the PSC model, and the results from the Proved case are subtracted. This provides
the entitlement and revenues related to the discrete Probable component. This approach can be
used with multiple categories and in cases where additional investments or operating costs may
also be required. It may also be used where there are multiple fields being developed within the
same PSC ring fence.
12.8 Summary
Reserves and resources disclosures are not only a matter of forecast production and costs.
Contractors must focus on their entitled resources, which are usually required for disclosure.
Concessions, production-sharing, revenue-sharing, risked-service, and other related contracts offer
the host country and the contractor alike considerable flexibility in tailoring agreement terms to
best meet sovereign and corporate requirements, which can have a variety of competing and
aligned factors. The entitlement of reserves and resources under the PRMS is based on economic
interest and will depend on the host fiscal system and other contracts and agreements involved.
When considering projects, each fiscal system, contract, or agreement must be reviewed on a
case-by-case basis to determine whether there is an opportunity to recognize reserves and resources
for internal use, regulatory reporting, or public disclosure. Particular care should be taken to ensure
352
that the contractual terms satisfy the company’s business objectives and that the impact of
alternative agreement structures is understood and considered.
The US Securities and Exchange Commission Section S-X, Rule 4-10b, “Successful Efforts
Method” (US Securities and Exchange Commission 2011) and FASB ASC 932 Topic “Extractive
Activities—Oil and Gas” (FASB 2010) provide guidelines and a useful framework for determining
when a mineral (economic) interest in hydrocarbon reserves and resources exists. These guidelines
may be used to supplement PRMS to help determine when an economic interest in hydrocarbons
exists, allowing reserves and resources to be recognized and reported. However, the distinction
between recognition and non-recognition of reserves and resources under many service-type
contracts may not be clear and may be highly dependent on subtle aspects of contract structure
and wording.
In some fiscal systems, such as concessions, it may be easier to determine the economic
interest and the entitled reserves and resources, while it may be difficult in many other contracts.
Cost-recovery terms in production-sharing, risked-service, and other related contracts provide a
more complex calculation for economic interest and typically reduce the production entitlement
(and hence reserves) obtained by a contractor in periods of high price and increase the volumes in
periods of low price. While this ensures cost recovery, the effect on investment metrics may be
counterintuitive. The treatment of taxes and royalties used can also have a very significant effect
on the reserves and resources recognized and production reported from these contracts.
Given the complexity of these types of agreements, determination of the net company share
of hydrocarbons recognized for each PRMS classification requires economic modeling of the flow
streams with the related costs and investments for each cumulative PRMS categorization and/or
classification (1P, 2P, 3P and 1C, 2C, 3C). The net amount for each discrete PRMS category can
then be determined by difference from the model results (e.g., net Probable Reserves = 2P Reserves
– 1P Reserves).
Important feedback and editorial effort have been provided by Monica Regina Bucholdz Monteiro,
Mariana de Azevedo B. Pereira da Hora, Aline Barreto Oliveira, Charles Vanorsdale, Bernard
Seiller, Xavier Troussaut, Dan DiLuzio, and Danilo Bandiziol.
12.10 References
Financial Accounting Standards Board (FASB). 2010. Accounting for Extractive Activities
(ASC)—Oil and Gas (Topic 932)—Oil and Gas Estimation and Disclosures, 2010-14, April
2010 and No. 2010-03, January 2010. Norwalk, Connecticut, USA: Financial Accounting
Standards Board.
Johnston, D. 1994. Petroleum Fiscal Systems and Production Sharing Contracts. Tulsa,
Oklahoma, USA: PennWell.
McMichael, C. L. and Young, E. D. 1997. Effect of Production Sharing and Service Contracts on
Reserves Reporting. Paper presented at the SPE Hydrocarbon Economics and Evaluation
Symposium, Dallas, Texas, USA, 16–18 March. SPE-37959-MS.
https://doi.org/10.2118/37959-MS.
McMichael, C. L. and Young, E. D. 2001. Reserve Recognition Under Production-Sharing and
Other Nontraditional Agreements. In Guidelines for the Evaluation of Petroleum Reserves
and Resources, Chap. 9. Richardson, Texas, USA: Society of Petroleum Engineers.
https://www.spe.org/industry/docs/Guidelines-Evaluation-Reserves-Resources-2001.pdf.
353
US Securities Exchange Commission (SEC). 2017. Regulation S-X 17 CFR. Washington, DC:
US Securities Exchange Commission. https://www.ecfr.gov/cgi-bin/text-
idx?amp;node=17:3.0.1.1.8&rgn=div5 (accessed 27 August 2020).
US Securities Exchange Commission (SEC). 2011. Regulation S-X, Rule 4-10b. Washington,
DC: US Securities Exchange Commission. https://www.ecfr.gov/current/title-17/chapter-
II/part-210 (accessed 1 June 2022).
Young, E. D. 2012. Reporting Reserves Under Modern Fiscal Arrangements. Paper presented at
the SPE Hydrocarbon Economics and Evaluation Symposium, Calgary, Alberta, Canada,
24–25 September. SPE 162708-MS. https://doi.org/10.2118/162708-MS.
12.11 Bibliography
Brock, H. R., Carnes, M. Z., and Justice, R. 2007. Petroleum Accounting, Principles,
Procedures, & Issues. Denton, Texas, USA: Professional Development Institute
Ernst & Young. 2019. Global Oil and Gas Tax Guide. London, UK: Ernst & Young.
Ernst & Young. 2019. US Oil and Gas Reserves Study. London, UK: Ernst & Young.
Financial Reporting in the Oil and Gas Industry. 2017. International Financial Reporting
Standards (July 2017). https://www.pwc.com/gx/en/services/audit-assurance/assets/pwc-
financial-reporting-in-the-oil-and-gas-industry-2017.pdf (accessed 29 August 2020).
Guidelines for Application of Petroleum Reserves Definitions. 1998. Houston, Texas, USA:
Guidelines for Reporting Oil and Gas Reserves. 1995. Canberra, Australia: Australian Petroleum
Production and Exploration Association.
Johnston, D. 1995. Different Fiscal Systems Complicate Reserves Values. Oil & Gas J 93 (22):
29 May 1995.
Martinez, A. R., Ion, D. C., DeSorcy, G. J., et al. 1987. Classification and Nomenclature Systems
for Petroleum and Petroleum Reserves. Study Group Report. Paper presented at the 11th
World Petroleum Congress, Houston, Texas, USA, 28 August –2 September. WPC-20134.
Peacock, D. 2018. Reserves Assessment in Petroleum Accumulations Straddling Boundaries:
The Relationship between Reserves & Resources and Unitisation. Paper presented at the
SPE Asia Pacific Oil and Gas Conference and Exhibition, Brisbane, Australia, 23–25
October. SPE-192144-MS. https://doi.org/10.2118/192144-MS.
Seba, R. D. 2008. Economics of Worldwide Petroleum Production. Tulsa, Oklahoma, USA: Oil
& Gas Consultants International.
Standards Pertaining to the Estimating and Auditing of Oil and Gas Reserve Information. 1979.
Richardson, Texas, USA: Society of Petroleum Engineers (December 1979).
Statement of Financial Accounting Standards 19, Paragraph 11. 1977. Norwalk, Connecticut,
USA: Financial Accounting Standards Board.
Statement of Recommended Practices—Fourth in a Series of SORPs. 1991. UK Oil Industry
Accounting Committee (January 1991).
Weaver, J. L. and Asmus, D. F. 2006. Unitizing Oil and Gas Fields Around the World: A
Comparative Analysis of National Laws and Private Contracts. Houst J Int Law 28: 3–197.
Wright, C. J. 2008. Fundamentals of Oil & Gas Accounting. Tulsa, Oklahoma, USA: PennWell.
354
Glossary
This Glossary provides further definition of terms used within the Guidelines for Application of
the PRMS and the Chapter and subsections in which they appear (e.g., 12.4.2 refers to section 4.2
in Chapter 12). References in numerous chapters are identified as “General,” while multiple
references within a given chapter may be identified as “Ch. X—General.”
TERM USED IN DEFINITION

THESE
GUIDELINES
1C 2.1 Denotes low estimate scenario of Contingent Resources.

2C 2.1 Denotes best estimate scenario of Contingent Resources.
3C 2.1 Denotes high estimate scenario of Contingent Resources.
1P 2.1 Denotes low estimate of Reserves (i.e., Proved Reserves).
Equal to P1.
2P 2.1 Denotes best estimate of Reserves. The sum of Proved plus
Probable Reserves.
3P 2.1 Denotes high estimate of reserves. The sum of Proved plus
Probable plus Possible Reserves.
1U 2.1 Denotes the unrisked low estimate qualifying as Prospective
Resources.
2U 2.1 Denotes the unrisked best estimate qualifying as Prospective
Resources.
3U 2.1 Denotes the unrisked high estimate qualifying as Prospective
Resources.
Abandonment, 9.3.2 The process (and associated costs) of returning part or all of a
Decommissioning, project to a safe and environmentally compliant condition
and Restoration when operations cease. Examples include, but are not limited
(ADR) to, the removal of surface facilities, wellbore plugging
procedures, and environmental remediation. In some
instances, there may be salvage value associated with the
equipment removed from the project. ADR costs are presumed
to be without consideration of any salvage value, unless
presented as “ADR net of salvage.”
Accumulation General An individual body of naturally occurring petroleum in a
reservoir.
Aggregation 8.1 The process of summing well, reservoir, or project-level
estimates of resources quantities to higher levels or
combinations, such as field, country, or company totals.
Arithmetic summation of incremental categories may yield
different results from probabilistic aggregation of distributions.
Amplitude Variation 3.4.1 The variation in the amplitude of a seismic reflection with angle
with Offset/Angle of incidence or source-geophone distance(depends on
(AVO/AVA) changes in velocity, density, and Poisson's ratio). Often used
as a hydrocarbon gas indicator because gas generally
decreases Poisson’s ratio and often increases amplitude with
incident angle/offset, although other conditions can produce
similar effects. (Source: SEG Wiki)
355
THESE
GUIDELINES
Analogous 4.3 Reservoirs that have similar rock properties (e.g.,
Reservoir petrophysical, lithological, depositional, diagenetic, and
structural), fluid properties (e.g., type, composition, density,
and viscosity), reservoir conditions (e.g., depth, temperature,
and pressure), and drive mechanisms, but are typically at a
more advanced stage of development than the reservoir of
interest and thus may provide insight and comparative data to
assist in estimation of recoverable resources.
Approved for 2.7, 2.8 All necessary approvals have been obtained; capital funds
Development have been committed, and implementation of the development
project is ready to begin or is underway. A project maturity
subclass that reflects the decision to start investing capital in
the construction of production facilities and/or drilling
development wells.
Assessment General See Evaluation.
Barrels of Oil 11.10.1, 11.10.2 The term allows for a single value to represent the sum of all
Equivalent (BOE) the hydrocarbon products that are forecast as resources.
Typically, condensate, oil, bitumen, and synthetic crude
barrels are taken to be equal (1 bbl = 1 BOE). Gas and NGL
quantities are converted to an oil equivalent based on a
conversion factor that is recommended to be based on a
nominal heating content or calorific value equivalent to a barrel
of oil.
Basin-Centered Gas 10.2.1, 10.2.3 An unconventional natural gas accumulation that is regionally
pervasive and characterized by low permeability, abnormal
pressure, gas saturated reservoirs, and lack of a downdip
water leg.
Best Estimate 2.1 With respect to resources categorization, the most realistic
assessment of recoverable quantities if only a single result
were reported. If probabilistic methods are used, there should
be at least a 50% probability (P50) that the quantities actually
recovered will equal or exceed the best estimate.
Bitumen 10.6.1, 10.6.2.1, See Natural Bitumen.
10.6.2.2
Bright Spot 3.4.1 A spot with a local increase of amplitude associated with
hydrocarbon accumulations. In its original implementation, the
phrase referred to amplitudes of the Common Mid-Point (CMP)
stack; modern implementations are pre-stack. These
amplitude highs can be caused by an increase of reflection
coefficient by gas in the pore space. It is used to identify the
increase in amplitude rather than the presence of
hydrocarbons, and it is usually greater in unconsolidated
clastic rocks. Acoustic impedance is lower in sands than in
shales, where pore space may be filled with water. As
hydrocarbons are added to pore spaces, the velocity and
density of the sand decreases. Due to this, the impedance
contrast at the top of the sand increases, making the reflection
stronger and more negative.
C1 2.1 Denotes low estimate of Contingent Resources. C1 is equal to
1C.
356
THESE
GUIDELINES
C2 2.1 Denotes Contingent Resources of same technical confidence
as Probable, but not commercially matured to Reserves.
C3 2.1 Denotes Contingent Resources of same technical confidence
as Possible, but not commercially matured to Reserves.
Carried Interest 12.5.8 A carried interest is an agreement under which one party (the
carrying party) agrees to pay for a portion or all of the
preproduction costs of another party (the carried party) on a
license in which both own a portion of the working interest.
Chance 8.1.1 Chance equals 1-risk. Generally synonymous with likelihood.
(See Risk) As used in Chapter 8 herein, “Chance” refers to the
elements necessary to establish a hydrocarbon accumulation
(e.g., reservoir seal, source, migration path).
Chance of 2.6 The estimated probability that the project will achieve
Commerciality commercial maturity to be developed. For Prospective
Resources, this is the product of the chance of geologic
discovery and the chance of development. For Contingent
Resources and Reserves, it is equal to the chance of
development.
Chance of 2.6 The estimated probability that a known accumulation, once
Development discovered, will be commercially developed.
Chance of Geologic 2.6 The estimated probability that exploration activities will confirm
Discovery the existence of a significant accumulation of potentially
recoverable petroleum.
Coalbed Methane 10.5 Natural gas contained in coal deposits. Coalbed gas, although
(CBM) usually mostly methane, may be produced with variable
amounts of inert or even non-inert gases. [Also called coal-
seam gas (CSG) or natural gas from coal (NGC)].
Commercial 9.2, 9.5, 9.7 A project is commercial when there is evidence of a firm
intention to proceed with development within a reasonable
time-frame. Typically, this requires that the best estimate case
meet or exceed the minimum evaluation decision criteria (e.g.,
rate of return, investment payout time). There must be a
reasonable expectation that all required internal and external
approvals will be forthcoming. Also, there must be evidence of
a technically mature, feasible development plan and the
essential social, environmental, economic, political, legal,
regulatory, decision criteria, and contractual conditions are
met.
Committed Project General Project that the entity has a firm intention to develop in a
reasonable time-frame. Intent is demonstrated with
funding/financial plans, but FID has not yet been declared.
(See also Final Investment Decision.)
Common Mid-Point 3.4.2, 3.5.1, Having the same mid-point between source and detector. The
(CMP) 3.5.3 CMP method is a recording/processing method where each
source is recorded at a number of geophone locations and
each geophone location is used to record from a number of
source locations.
Completion Interval General The specific reservoir interval(s) that is (are) open to the
borehole and connected to the surface facilities for production
or injection, or reservoir intervals open to the wellbore and
each other for injection purposes.
357
THESE
GUIDELINES
Concession 9.3.6, 12.5.1, A grant of access for a defined area and time period that
12.6.2.1, transfers certain entitlements to produced hydrocarbons from
12.6.3.1 the host country to an entity. The entity is generally
responsible for exploration, development, production, and sale
of hydrocarbons that may be discovered. Typically granted
under a legislated fiscal system where the host country
collects taxes, fees, and sometimes royalty on profits earned.
(Also called a license.)
Conditions 2.9, 2.10, 9.2, The economic, marketing, legal, environmental, social, and
9.3.3, 9.3.4, 9.5, governmental factors forecast to exist and impact the project
9.6, 9.7, 9.10 during the time period being evaluated (also termed
Contingencies).
Constant Case 9.3.4 A descriptor applied to the economic evaluation of resources
estimates. Constant case estimates are based on current
economic conditions being those conditions (including costs
and product prices) that are fixed at the evaluation date and
held constant, with no inflation or deflation made to costs or
prices throughout the remainder of the project life other than
those permitted contractually.
Consumed in 9.5, 11.4 That portion of produced petroleum consumed as fuel in
Operations (CiO) production or lease plant operations before delivery to the
market at the reference point. (Also called lease fuel.)
Contingency 2.6, 2.7, 2.11 A condition that must be satisfied for a project in Contingent
Resources to be reclassified as Reserves. Resolution of
contingencies for projects in Development Pending is
expected to be achieved within a reasonable time period.
Contingent 2.3 Those quantities of petroleum estimated, as of a given date, to
Resources be potentially recoverable from known accumulations by
application of development projects, but which are not
currently considered to be commercially recoverable owing to
one or more contingencies.
Continuous-Type 10.2.1, 10.2.2, A petroleum accumulation that is pervasive throughout a large
Deposit 10.2.3, 10.5.1, area and that generally lacks well-defined OWC or GWC.
10.6.1 Such accumulations are included in unconventional resources.
Examples of such deposits include “basin-centered” gas, tight
gas, tight oil, gas hydrates, natural bitumen, and oil shale
(kerogen) accumulations.
Conveyance 12.4.3, 12.5.11 Certain transactions that are in substance borrowings
repayable in cash or its equivalent and shall be accounted for
as borrowings and may not qualify for the recognition and
reporting of oil and gas reserves.
Correlation 8.1.1 A statistical measure by which the linear relationship between
two variables is described and/or quantified (through the use
of a correlation coefficient).
Correlation 8.2.3 Numerical measure relating the degree of correlation between
Coefficient two variables, ranging from an inverse relationship (-1) to a
direct relationship (+1). A value of (0) indicates that the two
variables have no relationship (i.e., they are independent of
each other).
358
THESE
GUIDELINES
Cost Recovery 12.5.2, 12.5.2.1, Under a typical production-sharing agreement, the contractor
12.5.4.1, 12.6.1, is responsible for the field development and all exploration and
12.7.1, 12.7.1.2, development expenses. In return, the contractor recovers
12.7.1.3, costs (investments and operating expenses) out of the
12.7.1.5 production stream. The contractor normally receives an
entitlement interest share in the petroleum production and is
exposed to both technical and market risks.
Cumulative 2.10 The sum of petroleum quantities that have been produced at a
Production given date. (See also Production.) Production is measured
under defined conditions to allow for the computation of both
reservoir voidage and sales quantities and for the purpose of
voidage also includes non-petroleum quantities.
Current Economic 9.3.4 Economic conditions based on relevant historical petroleum
Conditions prices and associated costs averaged over a specified period.
The default period is 12 months. However, in the event that a
step change has occurred within the previous 12-month
period, the use of a shorter period reflecting the step change
must be justified and used as the basis of constant-case
resources estimates and associated project cash flows.
Defined Conditions 2.10, 9.3.3, Forecast of conditions to exist and impact the project during
9.3.5, 9.6 the time period being evaluated. Forecasts should account for
issues that impact the commerciality, such as economics (e.g.,
hurdle rates and commodity price); operating and capital
costs; and technical, marketing, sales route, legal,
environmental, social, and governmental factors.
Dependency 8.1.1, 8.2.4 A statistical measure of the causative effect of one variable
upon another.
Depth Migration 3.5.3 A step in seismic processing in which reflections in seismic
data are moved to their correct locations in space, including
position relative to shotpoints, in areas where there are
significant and rapid lateral or vertical changes in velocity that
distort the time image. This requires an accurate knowledge of
vertical and horizontal seismic velocity variations.
Deterministic 2.5, 4.1 An assessment method based on defining discrete parts or
Incremental Method segments of the accumulation that reflect high, moderate, and
low confidence regarding the estimates of recoverable
quantities under the defined development plan.
Deterministic 2.5, 4.1 An assessment method based on discrete estimate(s) made
Method based on available geoscience, engineering, and economic
data and corresponding to a given level of certainty.
Deterministic 2.5, 4.1 Method where the evaluator provides three deterministic
Scenario Method estimates of the quantities to be recovered from the project
being applied to the accumulation. Estimates consider the full
range of values for each input parameter based on available
engineering and geoscience data, but one set is selected that
is most appropriate for the corresponding resources
confidence category. A single outcome of recoverable
quantities is derived for each scenario.
Developed 2.8 Reserves that are expected to be recovered from existing
Reserves wells and facilities. Developed Reserves may be further sub-
classified as Producing or Non-Producing.
359
THESE
GUIDELINES
Developed 2.8 Developed Reserves that are expected to be recovered from
Producing Reserves completion intervals that are open and producing at the
effective date. Improved recovery reserves are considered
producing only after the improved recovery project is in
operation.
Developed Non- 2.8 Developed Reserves that are either shut-in or behind-pipe.
producing Reserves
Development Not 2.7, 2.9 A discovered accumulation for which there are contingencies
Viable resulting in there being no current plans to develop or to
acquire additional data at the time due to limited commercial
potential. A project maturity sub-class of Contingent
Resources.
Development 2.7, 2.9 A discovered accumulation where project activities are
Pending ongoing to justify commercial development in the foreseeable
future. A project maturity sub-class of Contingent Resources.
Development Plan 2.2 The design specifications, timing, and cost estimates of the
appraisal and development project(s) that are planned in a
field or group of fields. The plan will include, but is not limited
to, well locations, completion techniques, drilling methods,
processing facilities, transportation, regulations, and
marketing. The plan is often executed in phases when
involving large, complex, sequential recovery and/or extensive
areas.
Development on 2.7, 2.9 A discovered accumulation where project activities are on hold
Hold and/or where justification as a commercial development may
be subject to significant delay.
Development 2.7, 2.9 A discovered accumulation where project activities are under
Unclarified evaluation and where justification as a commercial
development is unknown based on available information. This
sub-class requires appraisal or study and should not be
maintained without a plan for future evaluation. The sub-class
should reflect the actions required to move a project toward
commercial maturity. A project maturity sub-class of
Contingent Resources.
Dim Spot 3.4.1 A local low amplitude seismic attribute anomaly that can
indicate the presence of hydrocarbons and is therefore known
as a direct hydrocarbon indicator. It’s caused by highly
consolidated sands with a much greater acoustic impedance
than the overlying shale. It primarily results from the decrease
in acoustic impedance contrast when a hydrocarbon (with a
low acoustic impedance) replaces the brine-saturated zone
(with a high acoustic impedance) that underlies a shale (with
the lowest acoustic impedance of the three), decreasing the
reflection coefficient.
Discount Rate 9.3.5 The rate used in converting an undiscounted cash flow to a
cash flow that reflects a present value to the entitled
interest(s). The rate used will generally represent the entity’s
weighted average cost of capital (WACC) or its minimum
attractive rate of return (MARR).
360
THESE
GUIDELINES
Discovered/ 2.3 A petroleum accumulation where one or several exploratory
Discovery wells through testing, sampling, and/or logging have
demonstrated the existence of a significant quantity of
potentially recoverable hydrocarbons and thus have
established a known accumulation. In this context, “significant”
implies that there is evidence of a sufficient quantity of
petroleum to justify estimating the in-place volume
demonstrated by the well(s) and for evaluating the potential for
commercial recovery. (See also Known Accumulation.)
Discovered General Quantity of petroleum that is estimated, as of a given date, to
Petroleum Initially- be contained in known accumulations before production.
in-Place (PIIP) Discovered PIIP may be subdivided into commercial, sub-
commercial, and the portion remaining in the reservoir as
Unrecoverable.
Discovered 2.3, 2.7 Discovered petroleum in-place resources that are evaluated,
Unrecoverable as of a given date, as not able to be recovered by the
commercial and sub-commercial projects envisioned.
Economic 9.2, 9.7 A project is economic when it has a positive undiscounted
cumulative cash flow from the effective date of the evaluation,
the net revenue exceeds the net cost of operation (i.e.,
positive cumulative net cash flow at discount rate greater than
or equal to zero percent).
Economic Interest 9.3.6, 12.3, Interest that is possessed when an entity has acquired an
12.4.2, 12.4.2.1, interest in the minerals in-place or a license and secures, by
12.4.3 any form of legal relationship, revenue derived from the
extraction of the mineral to which they must look for a return.
Economic Limit 9.3.3 Defined as the time when the maximum cumulative net cash
flow (see Entitlement) occurs for a project.
Economic Limit Test 9.2 The process of identifying the Economic Limit from a forecast
(ELT) of revenues and expenses attributable to an entity’s interests.
Economically Not 2.9 Those quantities for which development projects are not
Viable Contingent expected to yield positive cash flows under reasonable
Resources forecast conditions. May also be subject to additional
unsatisfied contingencies.
Economically Viable 2.9 Those quantities associated with technically feasible projects
Contingent where cash flows are positive under reasonable forecast
Resources conditions but are not Reserves because it does not meet the
other commercial criteria.
Effective Date General Resource estimates of remaining quantities are “as of the
given date” (Effective Date) of the evaluation. The evaluation
must take into account all data related to the period before the
“as of date.”
Entitlement 12.1 That portion of future production (and thus resources) legally
accruing to an entity under the terms of the development and
production contract or license.
Entity General A legal construct capable of bearing legal rights and
obligations. In resources evaluations, this typically refers to the
lessee or contractor, which is some form of legal corporation
(or consortium of corporations). In a broader sense, an entity
can be an organization of any form and may include
governments or their agencies.
361
THESE
GUIDELINES
Established 2.3, 2.11 Methods of recovery or processing that have proved to be
Technology successful in commercial applications.
Estimated Ultimate 2.11 Those quantities of petroleum estimated, as of a given date, to
Recovery (EUR) be potentially recoverable plus those quantities that have
already been produced. For clarity, EUR must reference the
associated technical and commercial conditions for the
resources; for example, proved EUR is Proved Reserves plus
prior production.
Evaluation General The geosciences, engineering, and associated studies,
including economic analyses, conducted on a petroleum
exploration, development, or producing project resulting in
estimates of the quantities that can be recovered and sold and
the associated cash flow under defined forward conditions.
(Also called Assessment.)
Evaluator General The person or group of persons responsible for performing an
evaluation of a project. These may be employees of the
entities that have an economic interest in the project or
independent consultants contracted for reviews and audits. In
all cases, the entity accepting the evaluation takes
responsibility for the results, including its resources and
attributed value estimates.
Farm-in, Farm-out 12.4.2 Farm-in and Farm-out typically refer to an arrangement in
which the owner of a working interest (the farmor) assigns of
all or part of the working interest (including exploration and/or
production rights and obligations) to another party (the farmee)
in return for agreed compensation which can be monetary, in-
kind, or as relief from an obligation associated with the
assigned working interest. The farmor may or may not retain
any of a variety of types of interest.
Flat Spot 3.4.1, 3.4.2, Represents a hydrocarbon contact seismic response where it is
3.4.4 apparently flat. Such contact may be between gas and oil, oil
and water, or gas and water. The hydrocarbon reservoir must
be thicker than the vertical resolution in order to represent a flat
spot. Flat spots are often difficult to find; the edge or base of
channels, low angle faults, or processing artifacts can often be
misconceived as flat spots. Flat spots may also be caused by
low saturated gas in a reservoir.
Final Investment General Project approval stage when the participating companies have
Decision (FID) firmly agreed to the project and the required capital funding.
Forecast Case 9.3.4 A descriptor applied to a scenario when production and
associated cash-flow estimates are based on those conditions
(including costs and product price schedules, inflation indexes,
and market factors) forecast by the evaluator to reasonably
exist throughout the evaluation life (i.e., defined conditions).
Inflation or deflation adjustments are made to costs and
revenues over the evaluation period.
Gas Balance 11.8.2 In gas production operations involving multiple working interest
owners, maintaining a statement of volumes attributed to each,
depending on each owner’s portion received. Imbalances may
occur that must be monitored over time and eventually
balanced in accordance with accepted accounting procedures.
362
THESE
GUIDELINES
Gas Hydrates 10.1 Naturally occurring crystalline substances composed of water
and gas, in which a solid water lattice accommodates gas
molecules in a cage-like structure or clathrate. At conditions of
standard temperature and pressure, one volume of saturated
methane hydrate will contain as much as 164 volumes of
methane gas. Gas hydrates are included in unconventional
resources, but the technology to support commercial maturity
has yet to be developed.
Geostatistical 2.5 A variety of mathematical techniques and processes dealing
Methods with the collection, methods, analysis, interpretation, and
presentation of large quantities of geoscience and engineering
data to (mathematically) describe the variability and
uncertainties within any reservoir unit or pool, specifically
related here to resources estimates.
Gross Rock Volume 3.3.1, 5.4 The total volume of rock within an evaluation interval, typically
(GRV) a specific formation, from top to base of said interval, including
both reservoir and non-reservoir rock prior to the application of
net pay cutoff criteria.
High Estimate 2.1 With respect to resources categorization, this is considered to
(Resources) be an optimistic estimate of the quantity that will actually be
recovered from an accumulation by a project. If probabilistic
methods are used, there should be at least a 10% probability
(P10) that the quantities actually recovered will equal or
exceed the high estimate.
Highest Known 4.1 The shallowest occurrence of a producible hydrocarbon
Hydrocarbons accumulation as interpreted from some combination of well
log, flow test, pressure measurement, and core data.
Hydrocarbons may or may not extend above this depth.
Modifiers are often added to specify the type of hydrocarbons
(for instance, “highest known gas”).
History Match 6.4.3 The process of calibrating the underlying rock and fluid
properties, geologic model, and resultant predicted
performance of a numerical model to reasonably match the
actual performance history (pressure depletion, physics of fluid
flow, etc.) of a reservoir in terms of the fluid flow
characteristics, geological framework, etc. Due to the number
of variables involved, the history match of a reservoir is not
usually a unique realization.
Improved Recovery General The extraction of additional petroleum, beyond primary
(IR) recovery, from naturally occurring reservoirs by supplementing
the natural forces in the reservoir. It includes waterflooding
and gas injection for pressure maintenance, secondary
processes, tertiary processes, and any other means of
supplementing natural reservoir recovery processes. Improved
recovery also includes thermal and chemical processes to
improve the in-situ mobility of viscous forms of petroleum.
(Also called enhanced recovery.)
Justified for 2.7, 2.8 A development project that has reasonable forecast
Development commercial conditions at the time of reporting and there are
reasonable expectation that all necessary approvals/ contracts
will be obtained. A project maturity sub-class of Reserves.
363
THESE
GUIDELINES
Kerogen 10.3.2, 10.6.1, The naturally occurring, solid, insoluble organic material that
10.6.2.3 occurs in source rocks and can yield oil upon heating. Kerogen
is also defined as the fraction of large chemical aggregates in
sedimentary organic matter that is insoluble in solvents (in
contrast, the fraction that is soluble in organic solvents is
called bitumen). (See also Oil Shales.)
Known General An accumulation that has been discovered.
Accumulation
Lease Fuel 11.4 Oil and/or gas used for field and processing plant operations.
For consistency, quantities consumed as lease fuel should be
treated as part of shrinkage. However, regulatory guidelines
may allow lease fuel to be included in Reserves estimates.
Where claimed as Reserves, such fuel quantities should be
reported separately from sales and their value must be
included as an operating expense. [See also Consumed in
Operations (CiO).]
Liquefied Natural 11.2 Liquefied Natural Gas projects use specialized cryogenic
Gas (LNG) Project processing to convert natural gas into liquid form for tanker
transport. LNG is about 1/614 the volume of natural gas at
standard temperature and pressure.
Low/Best/High 2.1, 2.4, 2.5 Reflects the range of uncertainty as a reasonable range of
Estimates estimated potentially recoverable quantities.
Low Estimate 2.1 With respect to resources categorization, this is a conservative
estimate of the quantity that will actually be recovered from the
accumulation by a project. If probabilistic methods are used,
there should be at least a 90% probability (P90) that the
quantities actually recovered will equal or exceed the low
estimate.
Mean 8.1.1 The sum of a set of numerical values divided by the number of
values in the set.
Median 8.1.1 That point in the distribution of an uncertainty parameter at
which 50% of the possible outcomes are below and 50% of the
outcomes are above the value at that point.
Migration 3.5.3 Process of moving dipping reflections to their true subsurface
positions and collapses diffractions, thus increasing spatial
resolution and yielding a seismic image of the subsurface.
(See also Post-Stack Migration and Pre-Stack Migration.)
Mineral Interest 12.4.2, 12.4.3 See Economic Interest.
Mineral Lease 12.5.1 An agreement in which a mineral owner (lessor) grants an
entity (lessee) rights. Such rights can include (1) a fee
ownership or lease, concession, or other interest representing
the right to extract oil or gas subject to such terms as may be
imposed by the conveyance of that interest; (2) royalty
interests, production payments payable in oil or gas, and other
non-operating interests in properties operated by others; and
(3) those agreements with foreign governments or authorities
under which a reporting entity participates in the operation of
the related properties or otherwise serves as producer of the
underlying reserves (as opposed to being an independent
purchaser, broker, dealer, or importer).
364
THESE
GUIDELINES
Monte Carlo General A type of stochastic mathematical simulation that randomly
Simulation and repeatedly samples input distributions (e.g., reservoir
properties) to generate a resulting distribution (e.g.,
recoverable petroleum quantities).
Multi-Scenario 7.1, 7.3.2, 7.4.2 An extension of the deterministic scenario method. In this
Method case, a significant number of discrete deterministic scenarios
are developed by the evaluator, with each scenario leading to
a single deterministic outcome. Probabilities may be assigned
to each discrete input assumption from which the probability of
the scenario can be obtained; alternatively, each outcome may
be assumed to be equally likely.
Natural Bitumen 10.6.1, 10.6.2.1, The portion of petroleum that exists in the semi-solid or solid
10.6.2.2 phase in natural deposits. In its natural state, it usually
contains sulfur, metals, and other non-hydrocarbons. Natural
bitumen has a viscosity greater than 10 000 mPa·s (or 10,000
cp) measured at original temperature in the deposit and
atmospheric pressure, on a gas free basis. In its natural
viscous state, it is not normally recoverable at commercial
rates through a well and requires the implementation of
improved recovery methods such as steam injection. Natural
bitumen generally requires upgrading before normal refining.
Natural Gas Liquids 11.2, 11.3, 11.9, A mixture of light hydrocarbons that exist in the gaseous
(NGL) 11.10, 11.10.2 phase in the reservoir and are recovered as liquids in gas
processing plants. NGLs differ from condensate in two
principal respects: (1) NGLs are extracted and recovered in
gas plants rather than lease separators or other lease
facilities, and (2) NGLs include very light hydrocarbons
(ethane, propane, or butanes) as well as the pentanes-plus
that are the main constituents of condensates.
Net Cash Flow Ch.9–General A cash flow is the calculation, at discrete intervals (typically
(NCF) reported on an annual basis but perhaps calculated on a
monthly basis), of the cash inflow (e.g., revenue from product
sales) and the cash outflow (e.g., operating costs, capital
expenditures, taxes, etc.) attributable to a specific project. The
NCF is the difference between the inflow and the outflow to the
entity’s interest. The NCF may be undiscounted or a
discounted cash flow (DCF).
Net to Gross (NTG) 3.3.1, 5.4 The ratio of net reservoir rock to the gross rock volume, i.e.,
net pay divided by gross vertical thickness of the
rock/formation under evaluation.
Net Pay 5.4 The portion (after applying cutoffs) of the thickness of a
reservoir from which petroleum can be produced or extracted.
Value is referenced to a true vertical thickness measured.
Net Profits Interest 12.5.12 An interest that receives a portion of the net proceeds from a
well, typically after all costs have been paid.
Net Revenue 9.5, 12.4.2 An entity’s revenue share of petroleum sales after deduction of
Interest royalties or share of production owing to others under
applicable lease and fiscal terms.
Non-Hydrocarbon 11.2, 11.5, 11.9 Associated gases such as nitrogen, carbon dioxide, hydrogen
Gas sulfide, and helium that are present in naturally occurring
petroleum accumulations.
365
THESE
GUIDELINES
Non-Sales 11.3 That portion of estimated recoverable or produced quantities
that will not be included in sales as contractually defined at the
reference point. Non-sales include quantities CiO, flare, and
surface losses, and may include non-hydrocarbons.
Oil Sands 10.6.2.2 Sand deposits highly saturated with natural bitumen. Also
called “tar sands.” Note that in deposits such as the western
Canada oil sands, significant quantities of natural bitumen may
be hosted in a range of lithologies, including siltstones and
carbonates.
Oil Shales 10.1, 10.6.1, Shale, siltstone, and marl deposits highly saturated with
10.6.3 kerogen. Whether extracted by mining or in-situ processes, the
material must be extensively processed to yield a marketable
product (synthetic crude oil). (Often called kerogen shale.)
On Production 2.7, 2.8 A project maturity sub-class of Reserves that reflects the
operational execution phase of one or multiple development
projects with the Reserves currently producing or capable of
producing. Includes Developed Producing and Developed
Non-Producing Reserves.
Overlift/Underlift 11.8.1 Production entitlements received that vary from contractual
terms resulting in overlift (producing over the entitled contract
quantity) or underlift (producing below the entitled contract
quantity) positions. This can occur in annual records because
of the necessity for companies to lift their entitlement in parcel
sizes to suit the available shipping schedules as agreed upon
by the parties.
Overriding Royalty 9.3.6, 12.6.2 A revenue interest free of any cost obligation, created by the
Interest (ORRI) operating entity and/or working interest owner and paid by the
operating entity and/or working interest owner out of revenue
from the property. This differs from the Royalty Interest as
ownership of the minerals is not the basis for this interest but
rather another form of economic interest in the property.
P1 2.1 Denotes Proved Reserves. P1 is equal to 1P
P2 2.1 Denotes Probable Reserves.
P3 2.1 Denotes Possible Reserves.
P10 2.1, 2.4 Probabilistic resources estimate designation equivalent to
“high estimate.”
P50 2.1, 2.4 Probabilistic resources estimate designation equivalent to
“best estimate.”
P90 2.1, 2.4 Probabilistic resources estimate designation equivalent to “low
estimate.”
Petroleum General Defined as a naturally occurring mixture consisting of
hydrocarbons in the gaseous, liquid, or solid phase. Petroleum
may also contain non-hydrocarbon compounds; common
examples of which are carbon dioxide, nitrogen, hydrogen
sulfide, and sulfur. In rare cases, non-hydrocarbon content of
petroleum can be greater than 50%.
Petroleum Initially- General The total quantity of petroleum that is estimated to exist
in-Place originally in naturally occurring reservoirs, as of a given date.
Crude oil in-place, natural gas in-place, and natural bitumen
in-place are defined in the same manner.
366
THESE
GUIDELINES
Portfolio Effect 8.3.3 The stabilizing effect of having a population of individual
outcomes. As the population of independent portfolio elements
increases, the diversification effect of the larger population
allows the aggregated result to approach the population’s
mean. Consequently, the aggregated portfolio’s expected
result has reduced variance (e.g., the P10:P90 ratio) as the
population increases.
Possible Reserves General Possible Reserves are those additional reserves that analysis
of geoscience and engineering data indicates are less likely to
be recoverable than Probable Reserves. The total quantities
ultimately recovered from the project have a low probability to
exceed the sum of Proved plus Probable plus Possible (3P),
which is equivalent to the high-estimate scenario. When
probabilistic methods are used, there should be at least a 10%
probability that the actual quantities recovered will equal or
exceed the 3P estimate.
Post-Stack 3.4.2, 3.5.1 Migration applied on the Common Mid-Point stacking data.
Migration
Pre-Stack Migration 3.4.2, 3.5.1, Migration before Common Mid-Point stacking is done to avoid
3.5.3 the reflection-point smearing of dipping reflections and to
accommodate strong lateral velocity gradients. It can also be
used when the hyperbolic moveout assumption breaks down.
Probability 2.4, 2.5, 2.6, The extent to which an event is likely to occur, measured by
Ch. 7–General the ratio of the favorable cases to the whole number of cases
possible. PRMS convention is to quote cumulative probability
of exceeding or equaling a quantity where P90 is the small
estimate and P10 is the large estimate. (See also Uncertainty.)
Probability Density 8.3.2, 8.3.4 A probability density function (PDF), or density of a continuous
Function (PDF) random variable, is a function whose value at any given
sample (or point) in the sample space (the set of possible
values taken by the random variable) can be interpreted as
providing a relative likelihood that the value of the random
variable would equal that sample.
Probabilistic Method 2.4, 2.5, Ch. 7– The method of estimation of Resources is called probabilistic
General when the known geoscience, engineering, and economic data
are used to generate a continuous range of estimates and
their associated probabilities.
Probable Reserves General An incremental category of estimated recoverable quantities
associated with a defined degree of uncertainty. Probable
Reserves are those additional Reserves that are less likely to
be recovered than Proved Reserves but more certain to be
recovered than Possible Reserves. It is equally likely that
actual remaining quantities recovered will be greater than or
less than the sum of the estimated Proved plus Probable
Reserves (2P). In this context, when probabilistic methods are
used, there should be at least a 50% probability that the actual
quantities recovered will equal or exceed the 2P estimate.
Production General The cumulative quantities of petroleum that have been
recovered at a given date. Production can be reported in terms
of the sales product specifications, but project evaluation
requires that all production quantities (sales and non-sales), as
367
THESE
GUIDELINES
measured to support engineering analyses requiring reservoir
voidage calculations, are recognized.
Production Forecast 2.10, 9.2, 9.4.1, A forecasted schedule of production over time. For Reserves,
9.5 the production forecast reflects a specific development
scenario under a specific recovery process, a certain number
and type of wells and particular facilities and infrastructure.
When forecasting Contingent or Prospective Resources, more
than one project scope (e.g., wells and facilities) is frequently
carried to determine the range of the potential project and its
uncertainty together with the associated resources defining the
low, best, and high production forecasts. The uncertainty in
resources estimates associated with a production forecast is
usually quantified by using at least three scenarios or cases of
low, best, and high, which lead to the resources classifications
of, respectively, 1P, 2P, 3P and 1C, 2C, 3C or 1U, 2U, and 3U.
Production-Sharing 9.3.6, 12.2, A contract between a contractor and a host government in
Contract (PSC) 12.5.2, 12.5.2.1, which the contractor typically bears the risk and costs for
12.7 exploration, development, and production. In return, if
exploration is successful, the contractor is given the
opportunity to recover the incurred investment from
production, subject to specific limits and terms. Ownership of
petroleum in the ground is retained by the host government;
however, the contractor normally receives title to the
prescribed share of the quantities as they are produced. [Also
termed production-sharing agreement (PSA).]
Profit Split 12.5.2, 12.5.2.1, Under a typical production-sharing agreement, the contractor
12.7.1.3 is responsible for the field development and all exploration and
development expenses. In return, the contractor is entitled to a
share of the remaining profit oil or gas. The contractor receives
payment in oil or gas production and is exposed to both
technical and market risks.
Project 2.1.1, 2.2; A defined activity or set of activities that provides the link
General between the petroleum accumulation’s resources sub-class
and the decision-making process, including budget allocation.
A project may, for example, constitute the development of a
single reservoir or field, an incremental development in a
larger producing field, or the integrated development of a
group of several fields and associated facilities (e.g.
compression) with a common ownership. In general, an
individual project will represent a specific maturity level (sub-
class) at which a decision is made on whether or not to
proceed (i.e., spend money), suspend, or remove. There
should be an associated range of estimated recoverable
resources for that project. (See also Development Plan.)
Property 2.1.1 A defined portion of the Earth’s crust wherein an entity has
contractual rights to extract, process, and market specified in-
place minerals (including petroleum). In general, defined as an
area but may have depth and/or stratigraphic constraints. May
also be termed a lease, concession, or license.
Prospective 2.1, 2.2, 2.4, 2.6 Those quantities of petroleum that are estimated, as of a given
Resources date, to be potentially recoverable from undiscovered
accumulations by application of future development projects.
368
THESE
GUIDELINES
Proved Reserves General An incremental category of estimated recoverable quantities
associated with a defined degree of uncertainty. Proved
Reserves are those quantities of petroleum that, by analysis of
geoscience and engineering data, can be estimated with
reasonable certainty to be commercially recoverable, from a
given date forward, from known reservoirs and under defined
economic conditions, operating methods, and government
regulations. If deterministic methods are used, the term
“reasonable certainty” is intended to express a high degree of
confidence that the quantities will be recovered. If probabilistic
methods are used, there should be at least a 90% probability
that the quantities actually recovered will equal or exceed the
estimate.
Purchase Contract 12.5.10 A contract to purchase oil and gas provides the right to
purchase a specified volume of production at an agreed price
for a defined term.
Pure-Service 12.5.4, 12.5.5 Agreement between a contractor and a host government that
Contract typically covers a defined technical service to be provided or
completed during a specific time period. The service company
investment is typically limited to the value of equipment, tools,
and expenses for personnel used to perform the service. In
most cases, the service contractor’s reimbursement is fixed by
the contract’s terms with little exposure to either project
performance or market factors. No Reserves or Resources can
be attributed to these activities.
Qualified Reserves 5.5.1, 5.5.2, A reserves evaluator who (1) has a minimum of five years of
Evaluator (QRE) 5.5.5, 6.4.2 practical experience in petroleum engineering or petroleum
production geology, with at least three years of such
experience being in the estimation and evaluation of Reserves
information; and (2) either (a) has obtained from a college or
university of recognized stature a bachelor’s or advanced
degree in petroleum engineering, geology, or other discipline
of engineering or physical science or (b) has received, and is
maintaining in good standing, a registered or certified
professional engineer’s license or a registered or certified
professional geologist’s license, or the equivalent, from an
appropriate governmental authority or professional
organization. (Modified from SPE 2019 “Standards Pertaining
to the Estimating and Auditing of Oil and Gas Reserves
Information.”)
Rank 8.2.3 Rank is the probability position of an outcome across the full
uncertainty range. Rank correlation, therefore, is the similarity
between outcomes on the probability range for the
uncertainties, not the actual magnitudes.
Raw Production 4.5.1.6, 11.1, All components, whether hydrocarbon or other, produced from
11.3, 11.5 the well or extracted from the mine (hydrocarbons, water,
impurities such as non-hydrocarbon gases, etc.).
Reasonable General If deterministic methods for estimating recoverable resources
Certainty quantities are used, then reasonable certainty is intended to
express a high degree of confidence that the estimated
quantities will be recovered. Typically attributed to Proved
Reserves or 1C Resources quantities.
369
THESE
GUIDELINES
Reasonable 2.2, 2.11, 9.7 Indicates a high degree of confidence (low risk of failure) that
Expectation the project will proceed with commercial development or the
referenced event will occur. (Differs from Reasonable
Certainty, which applies to resources quantity technical
confidence, while Reasonable Expectation relates to
commercial confidence.)
Recoverable General Those quantities of hydrocarbons that are estimated to be
Resources producible by the project from either discovered or
undiscovered accumulations.
Recovery Efficiency 4.4.1.2, 4.4.1.3 A numeric expression of that portion (expressed as a
percentage) of in-place quantities of petroleum estimated to be
recoverable by specific processes or projects, most often
represented as a percentage. It is estimated using the
recoverable resources divided by the hydrocarbons initially in-
place. It is also referenced to timing. Current and ultimate (or
estimated ultimate) are descriptors applied to reference the
stage of the recovery. (Also called recovery factor.)
Reference Point 11.2, 11.3 A defined location within a petroleum extraction and
processing operation where quantities of produced product are
measured under defined conditions before custody transfer (or
consumption). Also called point of sale, terminal point, or
custody transfer point.
Reserves General Those quantities of petroleum anticipated to be commercially
recoverable by application of development projects to known
accumulations from a given date forward under defined
conditions. Reserves must satisfy four criteria: they must be
discovered, recoverable, commercial, and remaining (as of a
given date) based on the development project(s) applied.
Reservoir General A subsurface rock formation that contains an individual and
separate natural accumulation of petroleum that is confined by
impermeable barriers, pressure systems, or fluid regimes
(conventional reservoirs), or is confined by hydraulic fracture
barriers or fluid regimes (unconventional reservoirs).
Resources General Term used to encompass all quantities of petroleum
(recoverable and unrecoverable) naturally occurring in an
accumulation on or within the Earth’s crust, discovered and
undiscovered, plus those quantities already produced. Further,
it includes all types of petroleum whether currently considered
conventional or unconventional. (See Total Petroleum Initially-
in-Place.)
Resources 2.1.1, 2.4, 2.8, Subdivisions of estimates of resources to be recovered by a
Categories 2.10; General project(s) to indicate the associated degrees of uncertainty.
Categories reflect uncertainties in the total petroleum
remaining within the accumulation (in-place resources), that
portion of the in-place petroleum that can be recovered by
applying a defined development project or projects, and
variations in the conditions that may impact commercial
development (e.g., market availability and contractual
changes). The resource quantity uncertainty range within a
single resources class is reflected by either the 1P, 2P, 3P,
Proved, Probable, Possible, or 1C, 2C, 3C or 1U, 2U, 3U
resources categories.
370
THESE
GUIDELINES
Resources Classes 2.1.1, 2.4, 2.8, Subdivisions of resources that indicate the relative maturity of
2.10; General the development projects being applied to yield the
recoverable quantity estimates. Project maturity may be
indicated qualitatively by allocation to classes and sub-classes
and/or quantitatively by associating a project’s estimated
likelihood of commerciality.
Revenue-Sharing 12.5.3 Contracts that are very similar to the PSCs with the exception
Contract of contractor payment in these contracts, the contractor
usually receives a defined share of revenue rather than a
share of the production.
Reversionary 12.5.8, 12.5.9 The right of future possession of an interest in a project when
Interest a specified condition has been met.
Risk 8.2.4, 8.6 The probability of loss or failure. Risk is not synonymous with
uncertainty. Risk is generally associated with the negative
outcome, the term “chance” is preferred for general usage to
describe the probability of a discrete event occurring.
Risk and 12.4.1, 12.5.4 Risk and reward associated with oil and gas production
Reward activities are attributed primarily from the variation in revenues
cause by technical and economic risks. The exposure to risk in
conjunction with entitlement rights is required to support an
entity’s resources recognition. Technical risk affects an entity’s
ability to physically extract and recover hydrocarbons and is
usually dependent on a number of technical parameters.
Economic risk is a function of the success of a project and is
critically dependent on cost, price, and political or other
economic factors.
Risked-Service 12.5.4, 12.5.4.1 Agreements that are very similar to the production-sharing
Contract (RSC) agreements in that the risk is borne by the contractor but the
mechanism of contractor payment is different. With an RSC,
the contractor usually receives a defined share of revenue
rather than a share of the production.
Royalty, 9.3.6, 12.4.2, A type of entitlement interest in a resource that is free and
Royalty Interest 12.6.2, 12.6.2.1 clear of the costs and expenses of development and
production to the royalty interest owner. A royalty is commonly
retained by a resources owner (lessor/host) when granting
rights to a producer (lessee/contractor) to develop and
produce that resource. Depending on the specific terms
defining the royalty, the payment obligation may be expressed
in monetary terms as a portion of the proceeds of production
or as a right to take a portion of production in-kind. The royalty
terms may also provide the option to switch between forms of
payment at the discretion of the royalty owner. See also
Overriding Royalty Interest (ORRI).
Sales 11.2, 11.3 The quantity of petroleum and any non-hydrocarbon product
delivered at the custody transfer point (Reference Point) with
specifications and measurement conditions as defined in the
sales contract and/or by regulatory authorities.
Seismic Inversion 3.5.1 Seismic inversion combines seismic and well data to predict
rock properties (lithology, fluid content, porosity) across a
survey. These rock properties can be used to identify
hydrocarbon and reservoir. (SEG Wiki)
371
THESE
GUIDELINES
Seismic Resolution 3.2.1, 3.3.1 Seismic resolution is a measure of minimum spatial or
temporal separation between two reflection events so that they
can yet still be distinguished and resolved separately. Two
types of resolution are considered—vertical and lateral, both of
which are governed by the signal bandwidth. The vertical
resolution can be expressed as a fraction of the dominant
wavelength, which is a function of wave velocity and the
dominant frequency, for resolution criterion considered (criteria
of Rayleigh – λ/4, Ricker – λ/4, or Widess – λ/8, where λ is the
wavelength of dominant frequency). The lateral resolution is
expressed by the Fresnel zone, a function of depth of the
reflector, the velocity of the reflector and the dominant
frequency. Migration improves the lateral resolution by
decreasing the width of the Fresnel zone, thus separating
features.
Shale Gas 10.3.1, 10.4.2, Although the terms shale gas and tight gas are often used
10.4.5 interchangeably in public discourse, shale formations are only
a subset of all low-permeability tight formations, which include
sandstones and carbonates, as well as shales, as sources of
tight gas production.
Shale Oil 10.6.2.3 Although the terms shale oil and tight oil are often used
interchangeably in public discourse, shale formations are only
a subset of all low-permeability tight formations, which include
sandstones and carbonates, as well as shales, as sources of
tight oil production.
Split Classification 2.4, 9.3.4, A single project should be uniquely assigned to a sub-class
12.5.1 along with its uncertainty range, For example, a project cannot
have quantities categorized as 1C, 2P, and 3P. This is referred
to as Split Classification. If there are differing commercial
conditions, separate sub-classes should be defined.
Split Conditions 9.3.4 The uncertainty in recoverable quantities is assessed for each
project using resources categories. The assumed commercial
conditions are associated with resource classes or sub-
classes and not with the resources categories. For example,
the product price assumptions are those assumed when
classifying projects as Reserves, and a different price would
not be used for assessing Proved versus Probable reserves.
Stacking 3.4.2, 3.5.1, Usually means Common Midpoint (CMP) stacking. It is a major
3.5.3 process to enhance the signal-noise-ratio, reduce noise and
improve seismic data quality. In multichannel seismic
acquisition, the point on the surface halfway between the
source and receiver that is shared by numerous source-
receiver pairs. Traces from different shot records with a
common reflection point, such as CMP data, are stacked to
form a single trace during seismic processing. Such
redundancy among source-receiver pairs enhances the quality
of seismic data when the data are stacked.
372
THESE
GUIDELINES
Sub-Commercial 2.1, 2.3 A project subdivision that is applied to discovered resources
that occurs if either the technical or commercial maturity
conditions of project have not yet been achieved. A project is
sub-commercial if the degree of commitment is such that the
accumulation is not expected to be developed and placed on
production within a reasonable time-frame. Sub-commercial
projects are classified as Contingent Resources.
Sunk Cost 9.6.2 Money spent before the effective date and that cannot be
recovered by any future action. Sunk costs are not relevant to
future business decisions because the cost will be the same
regardless of the outcome of the decision. Sunk costs differ
from committed (obligated) costs, where there is a firm and
binding agreement to spend specified amounts of money at
specific times in the future (i.e., after the effective date).
Synthetic Crude Oil 10.6.1, 10.6.2.3 A mixture of hydrocarbons derived by upgrading (i.e.,
(SCO) chemically altering) natural bitumen from oil sands, kerogen
from oil shales, or processing of other substances such as
natural gas or coal. SCO may contain sulfur or other non-
hydrocarbon compounds and has many similarities to crude
oil.
Synthetic 3.4.1, 3.4.5, A 1D synthetic seismogram is formed by simply convolving an
Seismogram 3.5.2 embedded waveform with a reflectivity function. The
embedded waveform is sometimes a theoretical waveform
(such as a Ricker wavelet) and sometimes a waveform
resulting from analysis of actual seismic data (the embedded
wavelet, also called the equivalent wavelet). The reflectivity is
usually that calculated for normal incidence from velocity
(sonic) and density logs, but often only velocity changes are
considered because density changes are unknown (or else
some relationship between density and velocity is assumed). A
synthetic seismogram is used to correlate with seismic at the
well location.
Technical Forecast 2.10 The forecast of produced resources quantities that is defined
by applying only technical limitations (i.e., well-flow-loading
conditions, well life, production facility life, flow-limit
constraints, facility uptime, and the facility’s operating design
parameters). Technical limitations do not take into account the
application of either an economic or license cutoff. (See also
Technically Recoverable Resources.)
Technical 2.4 Indication of the varying degrees of uncertainty in estimates of
Uncertainty recoverable quantities influenced by the range of potential in-
place hydrocarbon resources within the reservoir and the
range of the recovery efficiency of the recovery project being
applied.
Technically 2.10, 2.11 Those quantities of petroleum producible using currently
Recoverable available technology and industry practices, regardless of
Resources (TRR) commercial or accessibility considerations.
Technology Under 2.3, 2.7, 2.10 Technology that is currently under active development and
Development that has not been demonstrated to be commercially viable.
There should be sufficient direct evidence (e.g., a test
project/pilot) to indicate that the technology may reasonably be
expected to be available for commercial application.
373
THESE
GUIDELINES
Tight Gas 10.2.1, 10.2.2, Gas that is trapped in pore space and fractures in very low-
10.2.5, 10.4.2, permeability rocks and/ or by adsorption on kerogen, and
10.4.6, 10.5.4 possibly on clay particles, and is released when a pressure
differential develops. It usually requires extensive hydraulic
fracturing to facilitate commercial production. Shale gas is a
sub-type of tight gas.
Tight Oil 10.2.1, 10.2.2, Crude oil that is trapped in pore space in very low-permeability
10.2.5, 10.4.2, rocks and may be liquid under reservoir conditions or become
10.4.6 liquid at surface conditions. Extensive hydraulic fracturing is
invariably required to facilitate commercial maturity and
economic production. Shale oil is a sub-type of tight oil.
Time Migration 3.5.3 A migration technique for processing seismic data in areas
where lateral velocity changes are not too severe, but
structures are complex. Time migration has the effect of
moving dipping events on a surface seismic line from apparent
locations to their true locations in time. The resulting image is
shown in terms of travel time rather than depth and must then
be converted to depth with an accurate velocity model to be
compared to well logs.
Total General All estimated quantities of petroleum that are estimated to
Petroleum exist originally in naturally occurring accumulations,
Initially-in-Place discovered and undiscovered, before production.
Uncertainty 2.1, 2.2, 2.4, The range of possible outcomes in a series of estimates. For
Ch. 7–General recoverable resources assessments, the range of uncertainty
reflects a reasonable range of estimated potentially
recoverable quantities for an individual accumulation or a
project. (See also Probability.)
Unconventional Ch. 10–General Unconventional resources exist in petroleum accumulations
Resources that are pervasive throughout a large area and lack well-
defined OWC or GWC (also called “continuous-type deposits”).
Such resources cannot be recovered using traditional recovery
projects owing to fluid viscosity (e.g., oil sands) and/or
reservoir permeability (e.g., tight gas/oil/CBM) that impede
natural mobility. Moreover, the extracted petroleum may
require significant processing before sale (e.g., bitumen
upgraders).
Undeveloped 2.8, 2.11 Those quantities expected to be recovered through future
Reserves investments: (1) from new wells on undrilled acreage in known
accumulations, (2) from deepening existing wells to a different
(but known) reservoir, (3) from infill wells that will increase
recovery, or (4) where a relatively large expenditure (e.g., when
compared to the cost of drilling and completing a new well) is
required to recomplete an existing well.
Undiscovered 2.1 That quantity of petroleum estimated, as of a given date, to be
Petroleum Initially- contained within accumulations yet to be discovered.
in-Place
Unitization 12.6.3, 12.6.3.1 Process whereby owners group adjoining properties and
divide reserves, production, costs, and other factors according
to their respective entitlement to petroleum quantities to be
recovered from the shared reservoir(s).
374
THESE
GUIDELINES
Unrecoverable 2.2, 2.3, 2.7 Those quantities of discovered or undiscovered PIIP that are
Resources assessed, as of a given date, to be unrecoverable by the
currently defined project(s). A portion of these quantities may
become recoverable in the future as commercial
circumstances change, technology is developed, or additional
data are acquired. The remaining portion may never be
recovered owing to physical/chemical constraints represented
by subsurface interaction of fluids and reservoir rocks.
Upgrader 11.2 A general term applied to processing plants that convert extra-
heavy crude oil and natural bitumen into lighter crude and less
viscous synthetic crude oil. While the detailed process varies,
the underlying concept is to remove carbon through coking or
to increase hydrogen by hydrogenation processes using
catalysts.
Volumetric Analysis 4.4.1, 4.4.2 This procedure uses reservoir rock and fluid properties to
calculate PIIP and then estimate that portion that will be
recovered by a specific development project. The volumetric
estimate may be based on either probabilistic or deterministic
approaches.
Well Abandonment 9.2, 9.3.2 The permanent plugging of a dry hole, an injection well, an
exploration well, or a well that no longer produces petroleum
or is no longer capable of producing petroleum
profitably. Several steps are involved in the abandonment of a
well: permission for abandonment and procedural
requirements are secured from official agencies; the casing is
removed and salvaged if possible; and one or more cement
plugs and/or mud are placed in the borehole to prevent
migration of fluids between the different formations penetrated
by the borehole. In some cases, wells may be temporarily
abandoned where operations are suspended for extended
periods pending future conversions to other applications such
as reservoir monitoring, enhanced recovery, etc.
Wet Gas General Natural gas from which no liquids have been removed before
the reference point. The wet gas is accounted for in resources
assessments, and there is no separate accounting for
contained liquids. It should be recognized that this is a
resources assessment definition and not a phase behavior
definition.
Working Interest 12.4.2, 12.4.2.1 An entity’s equity interest in a project before reduction for
royalties or production share owed to others under the
applicable fiscal terms.
375

Prms Guidelines Combined-Final

Uploaded by

Copyright:

Available Formats

Prms Guidelines Combined-Final

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Prms Guidelines Combined-Final

Uploaded by

Copyright:

Available Formats

PRMS

Guidelines for Application of the

World Petroleum Council

Society of Petroleum Engineers (SPE)

© Copyright 2022 Society of Petroleum Engineers

version of the same-named document. It is a companion volume to the Petroleum Resources

a specific passage within the current chapter of the AG.

universally true for all situations and/or data.

conjunction with any other resources guidelines or regulatory reporting requirements.

2.0 PETROLEUM RESOURCES DEFINITIONS, CLASSIFICATION,

3.0 SEISMIC APPLICATIONS ...................................................................................... 26

4.0 ASSESSMENT OF PETROLEUM RESOURCES USING

4.1 Introduction ............................................................................................................................57

Appendix A: Special Considerations for Material Balance Analyses ..................... 95

5.0 PETROPHYSICS .................................................................................................. 104

7.0 PROBABILISTIC RESOURCES ESTIMATION .................................................... 146

8.0 AGGREGATION OF RESERVES AND RESOURCES ........................................ 182

9.0 EVALUATION OF PETROLEUM RESERVES AND RESOURCES .................... 216

10.0 UNCONVENTIONAL RESOURCES ESTIMATION ............................................ 240

11.0 PRODUCTION MEASUREMENT AND OPERATIONAL ISSUES ..................... 303

GLOSSARY ................................................................................................................ 355

1.1 Rationale for the Applications Guidelines Revision

1.2 History of Petroleum Reserves and Resources Definitions

Petroleum Resources Definitions,

Fig. 2.1—Resources classification framework.

2.2 Defining a Project

2.4 Range of Uncertainty Categorization

2.5 Methods for Estimating the Range of Uncertainty in Recoverable Quantities

2.6 Chance of Commerciality

2.7 Project Maturity Sub-Classes

Fig. 2.3—Sub-classes based on project maturity.

2.8 Reserves Status

Approved for Development

Table 2.1—Project Maturity sub-class relative to Reserves Status sub-class.

2.9 Economic Status

Project Maturity Economic

Economically Viable Contingent Resources

Development Unclarified Undetermined

Development Not Viable Economically Not Viable Contingent Resources

Table 2.2—Approximate Project Maturity sub-class relative to Economic Status sub-class.

* For this example, this is not always true (see text).

Table 2.3—Comparison of recoveries.

2.11 Example Application of Commercial Maturity Sub-Classes

Fig. 2.7—Sub-classes based on maturity.

3.1 Introduction and Overview

3.2 Shared Earth Modeling

3.2.1 Static Model—Geologic Framework Architecture. A static model is often called a

3.3 Volumetrics, Trap Geometry, and Gross Rock Volume

Fig. 3.4—Grading of amplitude conformance (after Roden et al. 2012).

Fig. 3.5—Grading of amplitude continuity (after Roden et al. 2012).

Fundamental analysis by an AVO/DHI expert should be part of the seismic interpretation

3.5 Seismic Inversion

Fig. 3.9—Simplified schematic description of seismic impedance inversion

2. Establish the internal continuity of the reservoirs with reasonable certainty.

Quantifying uncertainty is a fundamental step in seismic inversion and seismic classification.

3.6 Seismic Surveillance

3.6.1 Seismic Surveillance Example A: 4D Interpretation and 3D Inversion Lead to

3.6.3 Seismic Surveillance Example C: Evaluate Fault transmissibility and Reservoir

Assessment of Petroleum Resources