See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/301694041

Magnetotelluric (MT) Method in Hydrocarbon Exploration: A New Perspective

Article · December 2004

CITATIONS READS

2 1,967

3 authors, including:

Hendra Grandis Djedi S. Widarto

Bandung Institute of Technology Pertamina University
55 PUBLICATIONS   82 CITATIONS    40 PUBLICATIONS   127 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Yunus Daud View project

Neotectonic of Papua, Indonesia View project

All content following this page was uploaded by Djedi S. Widarto on 29 April 2016.

The user has requested enhancement of the downloaded file.

 Magnetotelluric (MT) Method in Hydrocarbon Exploration: A New Perspective

Hendra Grandis 1), Djedi S. Widarto 2) , Agus Hendro 3)

1) Geophysics Program, Dept. of Geophysics and Meteorology – ITB

Jl. Ganesha 10, Bandung-40132 e-mail: grandis@geoph.itb.ac.id
2) Research Center for Geotechnology – LIPI
3) PT. Geoservices (Ltd.) Bandung

ABSTRACT

Magnetotelluric (MT) method has been used for petroleum exploration MT worldwide, especially in
reconnaissance stage of an exploration program and also in difficult areas due to topographic and structural
complexity. However, application of MT in Indonesia is very limited due many restraining circumstances.
The paper describes the basic principles of MT method and its potential application for hydrocarbon
exploration. Examples using both synthetic and field data from typical overthrust structures are discussed.

1. INTRODUCTION properties, i.e. conductivity or resistivity (see

Figure 1).
The seismic reflection method has become a
standard tool in oil and gas exploration for decades, Electric currents caused to flow in the earth by
and almost all of resources were found by earth’s changing magnetic field are called “telluric
extensive use of the seismic method. The trend currents” (from “Tellus”, one of the Latin names
towards increasingly difficult exploration targets for Earth). Hence the name “magnetotellurics”.
where seismic is less effective has led The EM waves which constitute the signal for MT
explorationists to consider other geophysical arise from (a) distant lightning (above ~1 Hz) and
techniques in a multi-disciplinary exploration (b) electric currents flowing in the ionosphere
program. (below ~1 Hz). The EM field caused by these
changing currents radiates around the earth (as do
Among supplementary geophysical techniques
radio waves), reflected repeatedly between the
in petroleum exploration, the magnetotelluric (MT)
conductive ionosphere and the relatively
method has been known for its capability in
conductive earth (Vozzof, 1991).
resolving geological structures less favorable for
seismic method (e.g. volcanic or basaltic high- The MT method has a wide range of
velocity cover, complex overthrust structures etc.). applications, from shallow investigations
However, the MT method has not been employed (geotechnics, groundwater and environment) to
for oil and gas exploration in Indonesia. Therefore, moderate and deep target in exploration of natural
the objective of this paper is to introduce the MT resources (mineral, geothermal and petroleum)
method and its application for petroleum depending on the frequency band used. However,
exploration. the MT method had been long known as the
preferred method in geothermal exploration due to
its great investigation depth and its effectiveness in
2. BASIC PRINCIPLES OF MT METHOD delineating low conductivity zones associated with
high temperature effects, e.g. alteration (Zhdanov
The magnetotelluric (MT) method is a & Keller, 1994).
frequency-domain electromagnetic (EM) sounding
technique used to investigate the electrical The Maxwell’s equations define the
structure of the earth’s subsurface. The method fundamental relationship between EM field vectors,
exploits naturally existing EM fields (typically in i.e. E (electric) and H (magnetic). Diffusion
frequency bands that span from 0.001 Hz to 10 equation of EM waves in a homogeneous medium
kHz) as signal source. These primary fields induce leads to solution with exponential decay of EM
secondary electric and magnetic fields in the waves amplitude with depth. In this case, skin
conductive earth. The transient variation of the depth is defined as the depth at which the
EM fields recorded at the surface of the earth is amplitude of EM waves become e-1 of that at the
therefore diagnostic to the subsurface electrical surface.

14
 JURNAL GEOFISIKA 2004/2

Skin depth (in meters) is expressed as is performed in the digital unit that consists of
(Kauffman & Keller, 1981): multi-channel analog to digital converters (ADCs)
and data storage. Field set-up of the equipment
δ ≈ 500 (ρ T)1/2 (1) using ZONGE GDP-32 is schematically presented
where ρ is resistivity (in Ohm.m) and T is period in Figure 2.
(in second). The equation (1) shows that slower Typically the frequency between 0.003 Hz to
attenuation of EM wave is obtained for longer 1 kHz is subdivided into three overlapping
period and higher resistivity of the medium. frequency bands, i.e. High Frequency (48 - 1024
Therefore, the skin depth is associated with the Hz), Medium Frequency (3 - 64 Hz) and LF (0.005
penetration depth of EM waves. It also describes - 4 Hz). High and medium frequency variations
the principle of MT sounding, i.e. information on are recorded first, continued by low and very low
variation of resistivity with depth is obtained by frequencies with longer recording time. At each
recording MT signals from a wide frequency band. frequency band, real-time quality control of data is
done by analyzing raw sounding curves obtained
from data processing performed in the receiver
unit. Based on coherency and outlier limit tests,
bad data are discarded and signal to noise ratio
(S/N) is increased by application of analogue
filters and amplifiers and also by accumulating
stacked data. Field processing capabilities allow
data quality control, immediate re-recording of
poor quality site and optimization of MT survey
grid.
Measurements are preferably conducted
during minimum anthropogenic noise since
sensitive sensors are used to measure very small
EM field variations. MT signals are usually of
good quality and strength after midnight and just
Figure 1. before dawn. Although distant lightning is the
Principles of electromagnetic induction (Telford et main EM signal source for MT, heavy rain and
al., 1990). thunderstorm occurred at or close to measurement
sites will saturate the data with noise. In such case,
the measurement is usually postponed.
Model response (forward modelling) of more
complicated resisitivity distribution is obtained by In general, one-day and up to two-day
resolving diffusion equation of EM waves in 1-D measurement is necessary to record complete MT
or 2-D media (e.g. Grandis, 1999; Uchida, 1993). signals (from high to low frequency) at each site.
The ratio of electric and magnetic fields, either This includes deployment and set-up of the
from forward modelling and from measurement, is equipment in the morning, measurement of MT
the impedance (Z) and usually expressed as signals from noon to morning of the next day,
apparent resistivity and phase: removing the equipment and moving to the next
site. In difficult areas, measurements are usually
1 ⎛ Im Z ⎞
Z ; φ = tan −1 ⎜
2
ρa = ⎟ (2) performed by flying camp.
ω µ0 ⎝ Re Z ⎠
MT sounding curves are log apparent resistivity
versus log period and phase versus log period.

3. MT DATA ACQUISITION AND

EQUIPMENT
EM fields variation with time is measured by
using coil magnetometers and non-polarizable
(porous-pot) electrodes. Two horizontal
components of electric field (Ex and Ey) and three
components of magnetic field (Hx, Hy and Hz) are Figure 2.
recorded. Signals are subject to filtering and Schematic diagram of the field set-up for MT data
amplification in an analog unit, while the recording acquisition (Zonge Engineering, 2001).

15
JURNAL GEOFISIKA 2004/2

4. MT DATA PROCESSING where Zxx = Zyy = 0 and Zxy = – Zyx if the

subsurface is actually 1-D, and then ZDET is a
The recorded data are stored in the receiver’s
scalar impedance.
hard disk or flash memory capable for storage of
data for several sites or several days of If the structure is purely 2-D and the rotated
measurement. The time series data are then coordinate system is parallel or perpendicular to
downloaded to the personal computer at the the strike of the structure, then the impedance
basecamp for preliminary data processing. tensor will take a simpler form, i.e.
In frequency domain, horizontal components ⎛ 0 Z xy ⎞
Z 2− D = ⎜⎜ ⎟ (6)
of electric and magnetic fields are related by the
⎝ Z yx 0 ⎟⎠
following equation:
⎛ E x ⎞ ⎛ Z xx Z xy ⎞ ⎛ Hx ⎞ In equation (6) Zxy and Zyx are termed TE-mode
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜ E y ⎟ = ⎜ Z yx
or (transverse electric) and TM-mode (transverse
⎝ ⎠ ⎝ Z yy ⎟⎠ ⎜ Hy ⎟
⎝ ⎠ magnetic) impedances respectively if x-axes is
parallel to the strike. The impedance tensor can be
E=ZH (3) rotated incrementally to obtain impedance tensor
close to the one expressed in equation (6) and to
where each component is function of frequency or
estimate the strike direction.
period. The MT data processing is essentially
spectral analysis of electric and magnetic fields By using equation (2) each component of the
time series data from which the transfer function or impedance tensor (before rotation), principal
impedance tensor (Z) is estimated. The readers are components of Z2-D and ZDET can be expressed
refferred to standard texbooks and published as apparent resistivity and phase.
papers for detailed description of MT data
processing (e.g. Nabighian, 1991).
5. MODELLING AND INTERPRETATION
The impedance tensor obtained from data
OF MT DATA
processing is associated with the measurement
coordinate system, where x- and y- axes are Comprehensive 1-D and 2-D MT forward and
conventionally set to North and East respectively. inverse modelling are routinely performed to
The impedance tensor represents the subsurface obtain the resistivity structure of the earth, while
resistivity distribution at each site in a 3-D way, i.e. 3-D MT modelling is still in the research stage.
it varies as function of the coordinate system. A We have developed a smoothness-constrained
tensor can be rotated mathematically in order to inversion technique which results in smooth 1-D
obtain the tensor associated with any different and 2-D resistivity model of the subsurface
coordinate system. A clockwise rotation of an (Grandis et al., 1999; 2003). The inversion
angle α will results in an impedance tensor in a method is similar to Occam’s inversion developed
new coordinate system as follows: by Constable et al. (1987) for 1-D model and
deGroot-Hedlin & Constable (1990) for 2-D model.
+ ⎛ cos α sin α ⎞
Z = R Z RT , R = ⎜ ⎟⎟
⎜ − sin α cos α The 1-D MT inversion modelling is used to
⎝ ⎠ infer horizontally layered model from invariant (or
(4) one of TM- and TE-mode) sounding curve at each
site. By employing different inversion techniques
where R is a rotation matrix and (.)T is matrix and constraints, a model representing blocky
transpose notation. (discrete) or smooth variation of resistivity with
By assuming that in 1-D or layered medium depth can be obtained (see Figure 3). A 2-D
the impedance is scalar (it does not depend on the resistivity section is formed by joining 1-D models
coordinate system) then the so-called invariant from contiguous sites along a profile. This pseudo
parameters can be extracted from the impedance 2-D model can be used for preliminary
tensor. The invariant parameters that do not vary interpretation or as a starting model for a full 2-D
with the angle of rotation represent overall MT inversion modelling.
resistivity distribution at each sounding site. One In the model obtained from inversion, the
of the invariant paramaters is the well-known resistivity units are related to various rock units
determinant of the impedance tensor written as and structures in the survey area. Obviously, it is
(Ranganayaki, 1988): important that any other available geophysical or
geological data be integrated to reduce the
Z DET = Z xx Z yy − Z xy Z yx (5)
ambiguities in the final interpretation of the MT
data.

16
 JURNAL GEOFISIKA 2004/2

RESISTIVITY (Ohm.m)
1000
1 10 100 1000
APP. RESISTIVITY (Ohm.m)

100

100

10
obs. data
calc. data

DEPTH (m)
1
1000

90
PHASE (deg.)

45

0
10000
0.001 0.01 0.1 1 10 100 1000
PERIOD (sec.)

Figure 3.
Example of 1-D MT inversion modelling: apparent resistivity and phase sounding curves (left), smooth (red)
and blocky (dashed blue) 1-D models representing resistivity variation with depth (right).

6. MT CASE STUDIES contrasts between the resistive carbonate and its

conductive surroundings were tested to produce
Complicated overthrust structures are typical
TE-mode and TM-mode data. We found that a
of hidrocarbon prospects of eastern Indonesia. The
resistivity contrast of 1000 Ohm.m to 100 Ohm.m
MT method was used to infer resistive carbonate
and TE-mode data gave best results.
layers embedded in more conductive formations.
It was expected that MT method would better Figure 4 shows the 2-D smooth model from
define the extent and depth of the overthrusted inversion of TE-mode data along with resistivity
carbonate formation over areas where the boundaries used to create the synthetic model. In
information was not yet available. The MT trial general, the model from smoothness-constrained
measurements were conducted only at four sites. inversion is in agreement with the geological
model representing the subsurface. However,
Inversions of MT data were performed to
information below the resistive carbonate
obtain smooth 1-D resistivity variation with depth
formation is difficult to assess for two reasons: (i)
at each sounding site. We can only resolve the top
the depth is beyond the frequency range of the MT
of carbonate formation indicated by a resistive
data, (ii) the upper resistive formation acts as a
layer at depth supported by well data adjacent to
“screen” such that it masks the deeper formations
the site. The limited number of MT sounding sites
(which is a well-known phenomenon in EM
renders comprehensive interpreation difficult.
theory).
However, the pilot survey allowed confirmation
and newly gained information concerning the In order to verify our results on synthetic data
carbonate formation. The results were inversion, we performed audio-frequency MT
encouraging and it is likely that with more and (AMT) measurements over a relatively well-
closely spaced sounding sites, promising results known (from geology) overthrust structures
could be expected. located in West Java (i.e. Tanjungkerta,
Sumedang). The data were obtained from 27 MT
The possibility of imaging such exploration
stations across two parallel faults. These faults are
target by MT method was then studied by using
identified as surface manifestations of the
synthetic data. A geological model representing
oversthrust structures. Figure 5 presents a smooth
overthrust structures was digitized and resistivity
2-D model and its interpretation that confirms
values were attributed to the carbonate and the
folded and overthrust structures of the survey area.
surrounding formations. Several resistivity

17
JURNAL GEOFISIKA 2004/2

Figure 4.
2-D smooth model from inversion of TE-mode synthetic data corresponding to resistivity boundaries of
overthrust structures.

Figure 4.
2-D smooth model representing overthrust structures (interpreted geological boundaries) of Tanjungkerta,
Sumedang, West Java.

7. CONCLUDING REMARKS • to map frontier areas on a reconnaissance basis

This paper outlined MT method and its to isolate specific prospects for more expensive
potential application in hydrocarbon exploration in seismic detailing, i.e. before seismic,
geologically difficult areas. Results from synthetic • to complement seismic interpretation with
and field data presented in this paper – although additional structural or lithological information,
very limited – provide a promising perspective of i.e. in basins with volcanic covers or with
the application of MT method for petroleum complicated overthrust structures.
exploration in structurally difficult areas.
However, it is worth noting that the MT
From case studies presented above and others method has its disadvantages and advantages. The
found in the literature, it can be concluded that drawbacks of MT method include: irregularity of
expected contributions from MT data in a natural signal, problem due to both natural and
petroleum exploration are as follows: artificial noise and lower resolution than seismic.

18
 JURNAL GEOFISIKA 2004/2

Advantages of MT method are as follows : calculation, Computer and Geosciences, 25,

• great depth of penetration that is comparable to 119 - 125.
seismics depth of interest, Grandis, H., Menvielle, M., Roussignol, M., 1999,
• informative in non-seismic or poor seismic Bayesian inversion with Markov chains-I. The
areas, magnetotelluric one-dimensional case,
Geophysical Journal International, 138, 757 -
• light-weight equipment and no transmitter 768.
required, minimizing logistical problems and
ability to access almost any environment, Grandis, H., Mogi, T., Fomenko, 2003, Multi-
dimensional magnetotelluric inversion
• good production rate (approximately 0.5 to 1 modelling unsing Markov Chain Monte Carlo
point per day for complete tensor MT method, International Union of Geodesy and
measurements), Geophysics (IUGG) General Assembly,
• in general better resolution than gravity and Sapporo, Japan.
magnetics, Kauffman, A.A., Keller, G.V., 1981, The
• well-developed interpretation procedure that magnetotelluric sounding method, Elsevier.
facilitates fast interpretation,
Nabighian, M., (ed.); 1991, Electromagnetic
Methods in Applied Geophysics, Vol.2,
The proposed strategy for conducting MT
Application, Investigations in Geophysics No.3,
survey for hydrocarbon exploration is the
SEG publishing.
following. In a reconnaissance survey, several
profiles with audio-frequency MT (AMT) scalar Ranganayaki, R.P., 1984, An interpretive analysis
measurements are preferred. The inter-station of magnetotelluric data, Geophysics, 49, 1730 –
distance is approximately 250 meters, while MT 1748.
soundings are performed only every 1 – 2 km. For Telford, W.M., Geldart, L.P., Sheriff, R.E., 1990,
detailed study, complete MT tensor measurement Applied Geophysics (2nd edition), Cambridge
can be performed on a selected profile with University Press.
stations interval of approximately 200 – 300
Uchida, T., 1993, Smooth 2-D inversion of
meters.
magnetotelluric data based on statistical
criterion ABIC, Journal of Geomagnetism &
Geoelectricity, 45, 841 - 858.
REFERENCES
Vozoff; K., 1991, The magnetotelluric method, in
Constable, S.C., Parker, R.L., Constable, C.G., M. Nabighian (ed.) Electromagnetic Methods in
1987, Occam's inversion, A practical algorithm Applied Geophysics, Vol.2, Application,
for generating smooth models from Investigations in Geophysics No.3, SEG
electromagnetic sounding data, Geophysics, 52, publishing.
289 - 300.
Zhdanov, M., Keller, G.V., 1994, The geoelectrical
deGroot-Hedlin, C., Constable, S., 1990, Occam's methods in geophysical exploration, Elsevier.
inversion to generate smooth, two-dimensional
models from magnetotelluric data, Geophysics, Zonge Engineering, 2001, AMT/MT Program
55, 1613 - 1624. Manual.
Grandis, H., 1999, An alternative algorithm for
one-dimensional magnetotelluric response

19

View publication stats