TD 39

Using HEC-RAS for Dam
Break Studies
August 2014
Approved for Public Release. Distribution Unlimited.
TD-39
Form Approved OMB No. 0704-0188
REPORT DOCUMENTATION PAGE
The public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions,
searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments
regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to the Department of
Defense, Executive Services and Communications Directorate (0704-0188). Respondents should be aware that notwithstanding any other provision of
law, no person shall be subject to any penalty for failing to comply with a collection of information if it does not display a currently valid OMB control
number.
PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ORGANIZATION.
1. REPORT DATE (DD-MM-YYYY)
2. REPORT TYPE
3. DATES COVERED (From - To)
August 2014
Training Document
4. TITLE AND SUBTITLE
5a. CONTRACT NUMBER
Using HEC-RAS for Dam Break Studies
5b. GRANT NUMBER

5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S)
5d. PROJECT NUMBER
Gary Brunner
5e. TASK NUMBER
5F. WORK UNIT NUMBER
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)
8. PERFORMING ORGANIZATION REPORT NUMBER
U.S. Army Corps of Engineers

Institute for Water Resources
Hydrologic Engineering Center (CEIWR-HEC)
609 Second Street
Davis, CA 95616-4687
TD-39
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)
10. SPONSOR/ MONITOR'S ACRONYM(S)

11. SPONSOR/ MONITOR'S REPORT NUMBER(S)
12. DISTRIBUTION / AVAILABILITY STATEMENT
Approved for Public Release. Distribution Unlimited.

13. SUPPLEMENTARY NOTES
14. ABSTRACT
This document provides information on how to use the HEC-RAS (River Analysis System) software when performing
a dam break analysis. The document presents the unique hydraulic modeling aspects that are required, plus routing the
inflow flood through a reservoir; estimating dam breach characteristics; and, downstream routing/modeling issues.
15. SUBJECT TERMS
HEC-RAS; hydraulic; model; terrain data; hydrologic; boundary conditions; terrain roughness; hydraulic structures;
calibration; unsteady flow; dam break; routing; inflow flood; reservoir; dam breach; downstream; modeling; flood
hydrograph; one-dimensional; Saint Venant equations; two-dimensional; level pool routing; scenarios; water surface
slope; pool; breach; dam; dynamic wave routing
16. SECURITY CLASSIFICATION OF:
a. REPORT
b. ABSTRACT
c. THIS PAGE
17. LIMITATION
OF
ABSTRACT
UU
18. NUMBER
OF
PAGES
74
19a. NAME OF RESPONSIBLE PERSON
19b. TELEPHONE NUMBER

Standard Form 298 (Rev. 8/98)
Prescribed by ANSI Std. Z39-18
Using HEC-RAS for Dam

Break Studies
August 2014

Institute for Water Resources
Hydrologic Engineering Center
609 Second Street
Davis, CA 95616
(530) 756-1104
(530) 756-8250 FAX
www.hec.usace.army.mil
TD-39
Table of Contents
Table of Contents
List of Figures ...................................................................................................................................... iii
List of Tables ........................................................................................................................................ v
Abbreviations ...................................................................................................................................... vii
Overview .............................................................................................................................................. 1
Routing the Inflow Flood through a Reservoir ................................................................................ 1
Full Dynamic Wave Routing ............................................................................................................. 3
Level Pool Routing ............................................................................................................................ 4
Estimating the Dam Breach Characteristics .................................................................................... 6
Causes and Types of Dam Failures ................................................................................................... 7
Estimating Breach Parameters ........................................................................................................... 8
User Entered Data Method ............................................................................................................. 8
Simplified Physical Breaching Method........................................................................................ 23
Physically-Based Breach Computer Models ................................................................................ 24
Peak Flow Equations and Envelope Curves .................................................................................... 26
Site Specific Data and Engineering Analysis .................................................................................. 28
Recommended Approach ................................................................................................................ 29
Example Application ....................................................................................................................... 31
Downstream Flood routing/Modeling Issues .................................................................................. 35
Cross Section Spacing and Hydraulic Properties ............................................................................ 36
Computational Time Step ................................................................................................................ 38
Manning's Roughness Coefficients ................................................................................................. 41
Downstream Storage, Tributaries, and Levees ................................................................................ 44
Modeling Bridge and Culvert Crossings ......................................................................................... 47
Modeling Steep Streams .................................................................................................................. 48
Drops in the Bed Profile .................................................................................................................. 50
Initial Conditions and Low Flow ..................................................................................................... 51
Downstream Boundary Conditions ................................................................................................. 53
Using Two-Dimensional Flow Areas for Dam Break Analysis ...................................................... 54
References.......................................................................................................................................... 57
Table of Contents
ii
List of Figures
List of Figures
Figure
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Page
Error in Level Pool Routing Compared to Full Dynamic Wave Routing. ................................ 3
Cross Section Layout for One-Dimensional Full Dynamic Routing Through a Reservoir. ...... 4
Reservoir Pool and Downstream Area Modeled with Two-Dimensional Flow Areas. ............ 5
Storage Area and Cross Section Layout for Level Pool Routing .............................................. 6
Example Breach Process for an Overtopping Failure ............................................................. 11
Example Breach Process for a Piping Failure ......................................................................... 12
Description of the Breach Parameters ..................................................................................... 14
Summary of Regression Equations for Breach Size and Failure Time (Wahl, 1998) ............. 16
HEC-RAS Simplified Physical Breach Option. ...................................................................... 24
Envelope of Experienced Outflow Rates from Breached Dams ............................................. 28
Dam Break Flood Wave Progression Downstream ................................................................. 30
Example Cross Section Layout (Ackerman, 2014) ................................................................. 36
Numerical Error Due to Cross Section Spacing ...................................................................... 37
Example Model Instability Due to Very Short Cross Section Spacing. .................................. 39
Example of Varying Computational Time Step ...................................................................... 40
Significant Turbulence and Sediment Load During the Teton Dam Failure (Olsen, 1976) .... 43
Cross Section Layout for a Tributary Coming Into a Main Stem River (Ackerman, 2014) ... 45
Example of Using Storage Areas and Lateral Weirs to Account for Flow Reversals up
Tributaries (Ackerman, 2014) ................................................................................................. 46
Tributary Storage Modeled as Cross Section Ineffective Flow Areas (Ackerman, 2014) ...... 46
Example of Using Lateral Structures and a Storage Area to model a Protected Area ............ 47
High Ground (Road or Levee) Represented as Part of the Cross Section ............................... 48
Example Bridge with Pre-Processed Bridge Curves ............................................................... 49
Model Instability Due to a Drop in the Bed Profile ................................................................ 51
Example of Initial Conditions for a Reservoir and Lateral Structures Connected to
Storage Areas .......................................................................................................................... 52
Example Model Due to Bad Downstream Boundary Condition ............................................. 54
Example of a Storage Area Connected to a Two-Dimensional Flow Area ............................. 55
iii
List of Figures
iv
List of Tables
List of Tables
Table
Number
1
2
3
4
5
6
Page
Possible Failure Modes for Various Dam Types ....................................................................... 8
Dam Breach Weir and Piping Coefficients ............................................................................. 13
Ranges of Possible Values for Breach Characteristics ............................................................ 15
Physically-Based Embankment Dam Breach Computer Software.......................................... 25
Summary of Erosion Process Models Currently Under Development .................................... 26
Summary of Breach Parameter Estimates ............................................................................... 35
List of Tables
vi
Abbreviations
Abbreviations
ARS
Agricultural Research Service
weir coefficient
FERC
Federal Energy Regulatory Commission
FT
Froude Number Threshold
H:V
horizontal/vertical
HEC
Hydrologic Engineering Center
HEC-HMS
Hydrologic Modeling System software
HEC-RAS
River Analysis System software
LPI
Local Partial Inertia Technique
exponent
NOAA
National Oceanic and Atmospheric Administration
NWS
National Weather Service
PMF
probable maximum flood
SCS
Soil Conservation Service
USACE
USBR
U.S. Bureau of Reclamation
vii
Abbreviations
viii
Training Document No. 39

Overview
The development of an HEC-RAS (Hydrologic Engineering Center's (HEC), River Analysis
System) hydraulic model requires an accurate representation of the terrain data and the
hydrologic inputs used as boundary conditions. Additionally, appropriate model parameters for
terrain roughness and hydraulic structures must be estimated and then calibrated in order to have
confidence in the model results. The guidelines in this document are focused on the
development and use of unsteady flow models for dam break studies. Discussions of basic data
requirements, hydraulic parameter estimates, and model calibration/validation are not covered in
this document. The HEC-RAS User's Manual (HEC, 2014) contains information describing
model input, data requirements, parameter estimation, and model calibration.
This document presents hydraulic modeling aspects that are unique to performing a dam break
analysis. Topics include: routing the inflow flood through a reservoir; estimating dam breach
characteristics; and downstream routing/modeling issues.
Routing the Inflow Flood through a Reservoir

HEC-RAS can be used to route an inflowing flood hydrograph through a reservoir with any of
the following three methods:
one-dimensional unsteady flow routing (full Saint Venant equations);
two-dimensional unsteady flow routing (Full Saint Venant equations or Diffusion wave
equations); or
with level pool routing
In general, full unsteady flow routing (one- or two-dimensional) will be more accurate for both
the with and without breach scenarios. The unsteady flow routing method can capture the water
surface slope through the pool as the inflowing hydrograph arrives, as well as the change in
water surface slope that occurs during a breach of the dam. Reservoirs with long narrow pools
will exhibit greater water surface slope upstream of the dam than reservoirs that are wide and
short. Therefore, the most accurate modeling technique to capture pool elevations and outflows
of long narrow reservoirs is full dynamic wave (unsteady flow) routing. For wide and short
reservoirs, level pool routing may be appropriate.
Several items must be taken into account before choosing the appropriate flood routing technique
for a given study:
In situations where the population is at risk and any damage centers are far enough
downstream, differences in peak outflow and the shape of the breach hydrograph may not be
significant by the time the flood wave reaches the downstream locations. Two hydrographs
that have the same volume, but different peak flows and shape, will tend to converge as they
1
are routed downstream through the river and floodplain. In this situation, the reservoir can
be modeled with either full unsteady flow routing or level pool routing.
The ability to acquire accurate cross section data (or terrain data for two-dimensional
routing) through the pool can be problematic. Detailed bathymetric surveys may be
required to accurately describe the elevation-volume relationship of the reservoir pool. If
detailed bathymetric data are not available, and full unsteady-flow routing is still desired,
cross section data can be modified to match the published elevation-volume curve of the
reservoir pool. This can be accomplished by running a series of steady flow profiles from
the dam to the upstream end of the pool, using a small flow and varying the downstream
starting condition for different pool elevations. HEC-RAS will compute the volume under
each profile. The elevation-volume curve computed by HEC-RAS can then be compared to
the published curve. Start with the lowest elevations. If the computed volume does not
match the published volume, the cross sections should be modified to increase or decrease
the volume required. The Channel Design/Modification Editor in HEC-RAS may prove
very useful for this task.
Capturing the full reservoir volume upstream of the dam will require the modeler to extend
cross sections far enough upstream, such that the invert elevation of the most upstream cross
section is higher than the highest elevation that will be modeled in the dam during the
largest event. Rough guidance would be to add a few feet to the top of the dam, and then
extend the model upstream far enough so that the most upstream cross section's invert is
higher than the highest elevation of the dam.
If there are significant numbers of tributaries, or some large tributaries upstream of the dam
that enter the pool directly, then storage volume due to backwater up the tributaries must be
accounted for as well as their inflows. For one-dimensional unsteady flow routing,
tributaries can be modeled in several manners. One option is to model all of the significant
tributaries as separate river reaches, using cross sections. A second option is to model the
tributaries as storage areas, and connect those storage areas to the main pool with a lateral
structure (weir). This will allow water to back up into the tributary as a level pool of water,
thus accounting for its volume. A third option is to extend the reservoir cross sections up the
tributaries and define that portion of the reservoir cross section as an ineffective flow area.
The differences between level pool routing and full unsteady flow routing through a reservoir
can be very difficult to quantify. In order to decide if level pool routing is adequate, it is helpful
to estimate the potential error in the peak flow of the routed outflow hydrograph, due to the use
of level pool routing. Dr. Danny Fread (National Weather Service, NWS) performed several
numerical experiments in which Dr. Fread compared both full dynamic wave routing to level
pool routing (Fread, 2006). From these experiments, Dr. Fread developed a set of equations and
a graph that can be used to estimate the error in using level pool routing for a given reservoir and
flood event. The graph and equations are shown in Figure 1 (Fread, 2006).
where:
Dr = the average depth of water in the reservoir (feet); approximated as Dmax/2.
Lr = the length of the reservoir pool in feet
Tr = the time of rise if the inflowing hydrograph in hours
2
Figure 1. Error in Level Pool Routing Compared to Full Dynamic Wave Routing
In order to compute the error in level pool routing (Figure 1), the user must calculate l, v, and
t. Once these three parameters are calculated, a percent error in the rising limb/peak flow of the
outflow hydrograph can be estimated. This error represents the difference in the answers
between using level pool routing and full dynamic wave routing.
Full Dynamic Wave Routing

As discussed previously, full dynamic wave (unsteady flow) routing through the reservoir pool is
the most accurate methodology and therefore should be performed for dam break analyses of
USACE (U.S. Army Corps of Engineers) dams whenever practical. To model the reservoir
using full dynamic wave routing with HEC-RAS, the user can either model the pool with onedimensional cross sections throughout the entire reservoir, as would be done for a normal river
reach, or they can model the reservoir pool as a single two-dimensional flow area. The dam is
modeled with the Inline Structure option in HEC-RAS. An example plot of modeling the pool
with one-dimensional cross sections is shown in Figure 2.
The inflow hydrographs (computed with HEC-HMS (Hydrologic Modeling System) can be
entered as boundary conditions at the upper end of the pool (flow hydrograph), and at any of the
locations within the reservoir pool (lateral inflow hydrographs).
When modeling the pool with cross sections, the engineer should be aware that after a dam
breach occurs, the upper reach will no longer be fully inundated from the reservoir pool, thus
acting more like a normal river reach. If the inflowing hydrograph recedes to a very low flow at
the tail of the event, there could be some potential model instabilities resulting from the
3
Figure 2. Cross Section Layout for One-Dimensional Full Dynamic Routing through a Reservoir
combination of a low flow and irregular channel geometry. One way around this is to increase
the base flow on the recession of the upstream hydrographs. Another approach is to smooth out
any major irregularities in the channel invert for the cross sections upstream of the dam.
Sometimes, the combination of these two suggestions may be necessary to keep a stable solution
above the dam for the tail end of the hydrograph.
If the reservoir pool is modeled with a two-dimensional flow area, then it can go completely dry
without any model stability issues when the two-dimensional cells dry out. An example of
modeling a reservoir with a two-dimensional flow area is shown in Figure 3.
Level Pool Routing

If it is not possible, necessary, or reasonable to perform full dynamic wave routing though the
reservoir, or if the presumed difference between level pool routing and dynamic routing is small,
then level pool routing can be performed with HEC-RAS. To model a reservoir using level pool
routing in HEC-RAS, the pool area is modeled with a storage area (HEC-RAS option for
modeling an area with level pool routing). That storage area is connected to a downstream river
reach, and that river reach must have a cross section that is inside the reservoir pool. The first
cross section in the reach is tied to the storage area by the fact that it will always have the same
water surface elevation during the computations. The dam is modeled as an inline structure,
which requires one cross section upstream of the inline structure. However, the cross section
upstream of the inline structure is tied to the inline structure boundary condition, and it cannot be
the first cross section of the reach. Because of this limitation in HEC-RAS, the result is that the
model must have two cross sections upstream of the inline structure: one cross section for the
connection to the storage area, and the second cross section for the inline structure boundary
condition. Both of the upstream cross sections should be representative of the reservoir area
4
Figure 3. Reservoir Pool and Downstream Area Modeled with Two-Dimensional Flow Areas
immediately upstream of the dam. The distance between these two cross sections should be
short (ten to twenty feet), so that the storage volume between the two cross sections is small. An
example diagram of modeling the reservoir with a storage area in HEC-RAS is shown in
Figure 4.
The engineer must enter an elevation-volume curve as part of the storage-area data describing the
reservoir. The minimum elevation of the two upstream cross sections should be roughly equal to
the minimum elevation specified for the storage area in order to prevent any instability once the
storage area is emptied.
When a dam break is modeled, the breach discharge will be computed by using the same
equations as the full dynamic wave method. The only difference is that the water supplied to the
dam will come from the storage area, and the storage area elevation will drop as a level pool as
water flows out of the breach. As noted above, when a rapidly forming breach occurs, the water
surface upstream of the dam will often have a significant slope to it. With the level pool routing
method, the water surface in the reservoir is always horizontal. This may or may not produce
5
Figure 4. Storage Area and Cross Section Layout for Level Pool Routing
significant differences in the outflow hydrograph, depending on many factors as outlined in this
Section.
Estimating the Dam Breach Characteristics

The estimation of a dam breach location, dimensions, and development time are crucial in any
assessment of a dam's potential risk. This is especially true in a risk assessment where dams will
be ranked based on the potential for loss of life and property damage. The breach parameters
will directly affect the estimate of the peak flow coming out of the dam, as well as any possible
warning time available to downstream locations. Unfortunately, the breach location, size, and
formation time, are often the most uncertain pieces of information in a dam failure analysis.
When performing a dam breach analysis, one must first estimate the characteristics of the breach.
Once the breaching characteristics are estimated, then HEC-RAS can be used to compute the
outflow hydrograph from the breach and perform the downstream routing.
The breach dimensions and development time must be estimated for every failure scenario that
will be evaluated. This requirement includes different failure modes as well as different
hydrologic events. The breach parameters associated with a PMF (probable maximum flood)
hydrologic event will be greatly different than the breach parameters for a sunny day failure at a
normal pool elevation. Therefore, for each combination of pool elevation (hydrologic event) and
failure scenario, a corresponding set of breach parameters must be developed.
A dam's potential breach characteristics can be estimated in several ways, including: comparative
analysis (comparing your dam to historical failures of dams of similar size, materials, and water
volume); regression equations (equations developed from historical dam failures in order to
estimate peak outflow or breach size and development time); utilization of velocity (or shear
stress) vs. erosion rates; and physically based computer models (software that attempts to model
the physical breaching process by using sediment transport/erosion equations, soil mechanics,
and principles of hydraulics). All of these methods are viable techniques for estimating breach
6
characteristics. However, each of these methods has strengths and weaknesses and should be
considered as a way of "estimating" the parameters and not utilized as absolute values.
In addition to the methods described above, site specific information, structural, and geotechnical
analyses should be used to refine and support the estimates of the breach parameters for each
failure scenario/hydrologic event. Historic breach information, regression equations, and
physically based computer models all have limitations that must be well understood when they
are applied. In any dam safety study it is important to consider a range of parameter estimates
for the breach size and development time for each failure scenario/event, and then perform a
sensitivity analysis of the breach parameters to identify their affect on the outflow hydrograph,
downstream stages and flows, and warning time to any population at risk.
The following section will cover causes and types of dam failures; estimating breach parameters;
recommended approach; and an example application.
As with many aspects of dam failure modeling in risk assessment studies, the level of effort in
estimating breach parameters should be consistent with the type of risk assessment. In general,
the level of effort and detail will increase from dams that are classified as "Low Hazard", to
dams that are classified as "High Hazard".
Causes and Types of Dam Failures

Historically, all types of dams have experienced failures due to one or more type of
event/loading. However, by far the majority of dam failures that have occurred have been
earthen dams, caused by some level of flood. The types of dams that are commonly built and
found in the field are:
Earthen embankment/rockfill
Concrete arch and multi arch
Concrete gravity
Buttress (combination of concrete gravity and arch dam)
Steel, timber, and composite materials
There are many mechanisms that can be the driving force of a dam failure. The following is a
list of mechanisms that can cause dam failures:
Flood event
Piping/seepage (internal and underneath the dam)
Landslide
Earthquake
Foundation failure
Equipment failure/malfunction (gates, etc.)
Structural failure
Upstream dam failure
Rapid drawdown of pool
Sabotage
Planned removal
7
Given the different mechanisms that cause dam failures, there can be several possible ways a
dam may fail for a given driving force/mechanism. Table 1 shows a list of dam types versus
possible modes of failure (Costa, 1985; Atallah, 2002).
Costa (1985) reports that of all dam failures as of 1985, 34 percent were caused by overtopping,
thirty percent due to foundation defects, 28 percent from piping and seepage, and eight percent
from other modes of failure. Costa (1985) also reports that for earth/embankment dams only, 35
percent have failed due to overtopping, 38 percent from piping and seepage, 21 percent from
foundation defects; and six percent from other failure modes.
Table 1. Possible Failure Modes for Various Dam Types
Earthen/
Concrete Concrete Concrete
Failure Mode
Embankment Gravity
Arch
Buttress
Overtopping
X
X
X
X
Piping/Seepage
X
X
X
X
Foundation Defects
X
X
X
X
Sliding
X
X
X
Overturning
X
X
Cracking
X
X
X
X
Equipment failure
X
X
X
X
Concrete
Multi-Arch
X
X
X
X
X
Estimating Breach Parameters

The estimation of the breach location, size, and development time are crucial in order to make an
accurate estimate of the outflow hydrographs and downstream inundation. However, these
parameters are some of the most uncertain in the entire analysis. Currently within HEC-RAS,
the user has two breaching methodologies to choose from, either "User Entered Data" or
"Simplified Physical". The User Entered Data method requires the user to enter all of the
breach information (i.e., breach size, breach development time, breach progression, etc.). The
Simplified Physical breaching method allows the user to enter velocity versus breach downcutting and breach widening relationships, which are then used dynamically to figure out the
breach progression versus the actual velocity being computed through the breach, on a time step
by time step basis.
User Entered Data Method
When using the User Entered Data option in HEC-RAS, the software requires the user to enter
the following information to describe a breach:
Location: centerline stationing of the breach in the dam
Failure Mode: overtopping or piping
Shape: bottom elevation, bottom width, left and right side slopes H:V
Time: critical breach development time
Trigger Mechanism: pool elevation; pool elevation plus duration; or clock time
8
Weir and Piping Coefficients: weir coefficients are used to compute overtopping/weir flow,
and an orifice coefficient is used to compute piping/pressure flow.
Failure Location. The breach failure location is based on many factors (type and shape of
dam, failure type, mode, and driving force of the failure). In general, one should consider all
factors about the dam, including any historical knowledge of seepage and foundation
problems, and place the breach location in the most probable location for each failure type.
The geotechnical engineer should be involved in determining the appropriate placement of
the breach.
Failure Mode. While HEC-RAS hydraulic computations are limited to overtopping and
piping failure modes, all other failure modes can be simulated with one of these two
methods. Failure mode is the mechanism for starting and growing the breach. Overtopping
failures start at the top of the dam and grow to maximum extents, while a piping failure mode
can start at any elevation/location and grow to the maximum extents. The ultimate breach
size and breach formation time are much more critical in the estimation of the outflow
hydrograph, than the actual failure initiation mode.
Critical Breach Development Time. HEC-RAS requires the user to enter what is called the
"critical breach development time". The critical breach development time for HEC-RAS can
be described as follows:
Overtopping Failure: The HEC-RAS breach start time is considered to be when the
erosion process has migrated to the upstream face of the dam (this is the start of a breach
for HEC-RAS). This is the point at which the outflow from the dam will start to increase
due to the breach. This condition is depicted in Figure 5C-D. The end of the breach
development time for HEC-RAS is when the breach is fully formed and significant
erosion has stopped. The breach development ending time should not include the time to
completely drain the reservoir pool.
Piping Failure: The HEC-RAS breach starting time for a piping failure is considered to
be when a significant amount of flow and material are coming out of the piping failure
hole. The breach ending time is considered to be when the breach is, for the most part,
fully formed (significant erosion has stopped, not the time until the reservoir pool is
emptied).
The estimation of the critical breach development time must be done outside of the HECRAS software and entered as input data. Descriptions on how to estimate this time are
provided.
Breach Weir and Piping Flow Coefficients. Weir and piping coefficients must be
entered by the user in HEC-RAS. These coefficients directly affect the magnitude of the
peak outflow hydrograph for any given breach. Unfortunately, exact knowledge of the
magnitude of these coefficients for a dam failure (overtopping or piping failure) is not
known.
In order to estimate the weir and piping flow coefficients, it is necessary to understand
the basic failure process. The following is a generalized description of the breach process
for an overtopping failure of an earthen dam. This description may not be true for all
earthen dams, as the breach process is a function of many parameters, such as: height of
the dam; volume of water behind the dam (including the inflowing hydrograph);
materials that the dam is constructed of; depth and duration of overtopping; outer
protective cover on the downstream and upstream side of the embankment; and other
parameters.
Overtopping Failure. In general, during an overtopping failure (Figure 5) of an earthen
dam, a headcut erosion process will first start on the downstream side of the dam
embankment (Figure 5A). While water is going over the dam crest, the dam crest acts
like a broad-crested weir. The headcut will erode back towards the center of the dam and
widen over time (Figure 5B). As the headcut begins to cut into the dam crest, the weir
crest length will become shorter, and the appropriate weir coefficient will trend towards a
sharp-crested weir value (Figure 5C). The time for breach initiation used in HEC-RAS is
shortly after what is depicted in Figure 5C. When the headcut reaches the upstream side
of the dam crest, a mass failure of the upstream crest may occur, and the hydraulic
control section will act very much like a sharp-crested weir (Figure 5D). The headcut
will continue to erode upstream through the dam embankment, as well as erode down
through the dam and widen at the same time (Figure 5E). During this process, the
appropriate weir coefficient will begin to trend back towards a broad-crested weir
coefficient. As the downward cut reaches the natural river bed elevation, and the breach
is more in a widening phase, the appropriate weir coefficient is more in the range of a
broad-crested weir value.
Piping Failure. A general description of a piping failure (Figure 6) is as follows. Water
is seeping through the dam at a significant enough rate, such that it is internally eroding
material and transporting it out of the dam. As the material is eroded, a larger hole is
formed, thus able to carry more water and erode more material (Figure 6A). The
movement of water through the dam during this process is modeled as a pressurized
orifice type of flow. During the piping flow process, erosion and headcutting will begin
to occur on the downstream side of the dam (Figure 6B) as a result of flow exiting the
pipe. As the piping hole grows larger, material above the hole will begin to slough off
and fall into the moving water (Figure 6C). The headcutting and material sloughing
processes will continue to move back towards the upstream side of the dam, while the
piping hole continues to grow simultaneously (Figure 6D). If the piping hole is large
enough, the weight of the material above the hole may be too great to be maintained, and
a mass caving of material will occur. This will result in a large rise in the outflow
through the breach and will accelerate the breaching process. Also at this point, the
hydraulics of the flow transitions from a pressure/orifice type flow to an open air weir
type flow. The headcutting and erosion process then continues back through the dam, as
well as downward (Figure 6E). Additionally, the breach will be widening. Depending on
the volume of water behind the dam, the breach may continue to cut down and widen
until the natural channel bed is reached. Then the breach will go into a widening phase.
10
Figure 5. Example Breach Process for an Overtopping Failure

11
Figure 6. Example Breach Process for a Piping Failure

12
As you can imagine from the description of the breach processes, as well as other factors and
complications that may occur in the real world, estimating these parameters can be difficult.
Currently in software such as HEC-RAS, the user is only allowed to enter a single value for the
breach weir coefficient and for the piping coefficient. Because the estimate of the peak flow is
so important in this process, one should try to estimate these coefficients based on the phase of
the breach process in which they think the largest flows will most likely occur. For example,
earthen dams with medium to very large storage volumes upstream, will most likely have failed
all the way down to the natural stream bed elevation, and be in the breach widening phase when
the peak outflow occurs. This would suggest using a weir coefficient (C) that is typical of a
broad-crested weir with a long crest length (i.e., C = 2.6). However, for dams with a relatively
low volume of water in comparison to the height of the dam, the peak flow may occur during the
phase of the breach in which the breach is still cutting down through the dam. For this case, a
weir coefficient typical of a sharp-crested weir would be more appropriate (i.e., C = 3.2). Other
factors to consider are the material types of the dam. Dams that have a clay core, and are
generally constructed of clay material, will tend to have a much more pronounced headcut
process. While dams that are more in the sand and gravel range will have a less pronounced
headcut process. This may lead to using higher weir coefficients for a clay dam (i.e., C = 3.2,
sharp-crested weir) versus a gravel/sand dam (i.e., C = 2.6, broad-crested weir).
During a piping failure breach, the rate of water flowing through the dam is modeled with an
orifice pressure flow equation. This equation also requires a discharge coefficient, which is a
measure of how efficiently the flow can get into the pipe orifice. Because a piping failure is not
a hydraulically designed opening, it is assumed that the entrance is not very efficient.
Recommended values for the piping/pressure flow coefficients are in the range of 0.5 to 0.6.
Guidelines for selecting breach weir and piping flow coefficients are provided in Table 2.
Table 2. Dam Breach Weir and Piping Coefficients
Overflow/Weir
Dam Type
Coefficients
Earthen Clay or Clay Core
2.6 3.3
Earthen Sand and gravel
2.6 3.0
Concrete Arch
3.1 3.3
Concrete Gravity
2.6 3.0
Piping/Pressure Flow
Coefficients
0.5 0.6
0.5 0.6
0.5 0.6
0.5 0.6
Breach Shape Definitions. For the purposes of these guidelines, the physical description of the
breach will consist of the height of the breach, breach width, and side slopes in H:V (side slopes
are expressed in units of distance horizontal to every one unit in the vertical). These values
represent the maximum breach size. A diagram describing the breach is shown in Figure 7.
The breach width is described as the average breach width (Bave) in many equations, while HECRAS requires the breach bottom width (Wb) for input. The breach height (hb) is the vertical
extent from the top of the dam to the average invert elevation of the breach. Many publications
and equations also use the height of the water (hw), which is the vertical extent from the
maximum water surface to the invert elevation of the breach. The side slopes are expressed in
H:V.
The breach dimensions, as well as the breach formation time must be estimated outside of the
HEC-RAS software, and entered into the program. Many case studies have been performed on
13
Figure 7. Description of the Breach Parameters
data from historic dam failures, leading to guidelines, regression equations, and computer
modeling methodologies for prediction of the dam breach size and time. One of the most
comprehensive summaries of the literature on historic dam failures is a U.S. Bureau of
Reclamation (USBR) report written by Mr. Tony Wahl titled "Prediction of Embankment Dam
Breach Parameters - A Literature Review and Needs Assessment" (Wahl, 1998). This report
discusses all types of dams, however the report focuses on earthen/embankment dams for the
discussion of estimating breach parameters. Much of what is presented in this section of the
guidelines was extracted from that report. Guidelines for breach parameters for concrete (arch,
gravity, buttress, etc.), steel, timber, and other types of structures, is very sparse, and is limited to
simple ranges.
Federal Agency Guidelines. Many federal agencies have published guidelines in the form of
possible ranges of values for breach width, side slopes, and development time. Table 3
summarizes some of these guidelines.
The guidelines shown in Table 3 should be used as minimum and maximum bounds for
estimating breach parameters. More specific ways to estimate breach characteristics are
addressed below.
Regression Equations. Several researchers have developed regression equations for the
dimensions of the breach (width, side slopes, volume eroded, etc.), as well as the failure time.
These equations were derived from data for earthen dams, earthen dams with impervious cores
(i.e., clay, concrete, etc.), and rockfill dams. Therefore, these equations do not directly apply to
concrete dams or earthen dams with concrete cores. The report by Wahl (1998) describes several
equations that can be used for estimating breach parameters. Summarized in Figure 8 are the
regression equations developed to predict breach dimensions and failure time from the USBR
report (Wahl, 1998).
Since the report by Wahl (1998), additional regression equations have been developed to
estimate breach width and breach development time. In general, several of the regression
equations should be used to make estimates of the breach dimensions and failure time. These
estimates should then be used to perform a sensitivity analysis, as discussed later in this
14
Table 3. Ranges of Possible Values for Breach Characteristics

Horizontal
Component of
Average
Breach Side
Breach Width
Slope (H)
(Bave)
(H:V)
Dam Type
(0.5 to 3.0) x HD
0 to 1.0
(1.0 to 5.0) x HD
0 to 1.0
Earthen/Rockfill
(2.0 to 5.0) x HD 0 to 1.0 (slightly larger)
(0.5 to 5.0) x HD*
0 to 1.0
Multiple Monoliths
Vertical
Usually 0.5 L
Vertical
Concrete Gravity
Usually 0.5 L
Vertical
Multiple Monoliths
Vertical
Concrete Arch
Slag/Refuse
Entire Dam
Entire Dam
(0.8 x L) to L
(0.8 x L) to L
(0.8 x L) to L
(0.8 x L) to L
Valley wall slope

0 to valley walls
0 to valley walls
0 to valley walls
1.0 to 2.0
Failure
Time, tf
(hours)
0.5 to 4.0
0.1 to 1.0
0.1 to 1.0
0.1 to 4.0*
0.1 to 0.5
0.1 to 0.3
0.1 to 0.2
0.1 to 0.5
Agency
USACE 1980
FERC
NWS
USACE 2007
USACE 1980
FERC
NWS
USACE 2007
0.1 USACE 1980

0.1
FERC
0.1
NWS
0.1 USACE 2007
0.1 to 0.3
FERC
0.1
NWS
*Note: Dams that have very large volumes of water, and have long dam crest lengths, will continue to erode for long durations
(i.e., as long as a significant amount of water is flowing through the breach), and may therefore have longer breach widths and
times than what is shown in Table 3. HD = height of the dam; L = length of the dam crest; FERC - Federal Energy Regulatory
Commission; NWS - National Weather Service
document. The user should try to pick regression equations that were developed with data that is
representative of the study dam. In many cases this may not be possible, due to the fact that
most of the historic dam failures for earthen dams have occurred on smaller structures. In fact,
out of the 108 historic dam breaches listed in the USBR report (Wahl, 1998), only thirteen of the
dams are over 100 feet (30.5 meters) high and only five of the dams had a storage volume greater
than 100,000 acre-feet (123.4x106 cubic meters) at the time of failure. Additionally, most of the
regression equations were developed from a smaller subset of this data (20 to 50 dams), and the
dams included in the analysis are a mixture of homogenous earthen dams and zoned earthen
dams (dams with clay cores, or varying materials). Therefore, the use of any of the regression
equations should be done with caution, especially when applying them to larger dams that are
outside the range of data for which the equations were developed. The use of regression
equations for situations outside of the range of the data they for which were developed for may
lead to unrealistic breach dimensions and development times.
The following regression equations have been used for several dam safety studies found in the
literature (except the Xu and Zhang equations, which are presented because of their wide range
of historical data values), and are presented in greater detail in this document:
Froehlich (1995a)
Froehlich (2008)
MacDonald and Langridge-Monopolis (1984)
Von Thun and Gillette (1990)
Xu and Zhang (2009)
15
Table 2 - Breach Parameter relations based on dam-failure case studies.

For explanations of symbols see the Notation section at the end of this report.
Reference
Johnson and Illes (1976)
Singh and Snorrason
(1982, 1984)
MacDonald and
Langridge-Monopolis
(1984)
Number of
Case Studies
20
42
FERC (1987)
Froehlich (1987)
43
Relations Proposed
(S.I. units, meters, m3/s, hours)
0.5hd B 3hd for earthfill dams
2hd B hd
0.15 meters dovtop 0.61 meters
0.25 hours tf 1.0 hours
Earthfill dams:
Ver = 0.0261(Vout*hw)0.769
[best-fit]
0.564
[upper envelope]
tf = 0.0179(Ver)
Non-earthfill dams:
Ver = 0.00348(Vout*hw)0.852
[best-fit]
B is normally 2-4 times hd
B can range from 1-5 times hd
Z = 0.25 to 1.0
[engineered, compacted dams]
Z = 1 to 2
[non-engineered, slag or refuse dams]
[engineered, compacted earth dams]
tf = 0.1-1 hours
tf = 0.1-0.5 hours
[non-engineered, poorly compacted]
= 0.47Ko(S*)0.25
Ko = 1.4 overtopping; 1.0 otherwise
0.73
Reclamation (1988)
Singh and Scarlatos
(1988)
Von Thun and Gillette
(1990)
Dewey and Gillette (1993)
Froehlich (1995b)
52
)1.57

= 0.75 (
Kc = 0.6 with corewall; 1.0 without a corewall
= 79(S*)0.47
B = (3)hw
tf = (0.011)B
Breach geometry and time of failure tendencies
Btop/Bbottom averages 1.29
57
B, Z, tf guidance (see discussion)
57
63
Breach initiation model; B, Z, tf guidance

= 0.1803 0.32 0.19
(0.90)
= 0.002540.55
Ko = 1.4 for overtopping; 1.0 otherwise
Figure 8. Summary of Regression equations for Breach size and Failure Time (Wahl 1998)
These regression equations have been used on several dam break studies and have been found to
give a reasonable range of values for earthen, zoned earthen, earthen with a core wall (i.e., clay),
and rockfill dams. The following is a brief discussion of each equation set.
Froehlich (1995a): Froehlich utilized 63 earthen, zoned earthen, earthen with a core wall
(i.e., clay), and rockfill data sets to develop as set of equations to predict average breach
width, side slopes, and failure time. The data that Froehlich used for his regression analysis
had the following ranges:
Height of the dams: 3.66 92.96 meters (12 305 feet)
(with 90% < 30 meters, and 76% < 15 meters)
Volume of water at breach time: 0.0130 660.0 m3 x 106 (11 - 535,000 acre-feet)
(with 87% < 25.0 m3 x 106, and 76% < 15.0 m3 x 106)
16
Froehlich's regression equations for average breach width and failure time are:
Bave = 0.1803 K o Vw0.32 hb0.19
tf = 0.00254 Vw0.53 hb0.90

where:
Bave
Ko
Vw
hb
tf
=
=
=
=
=
average breach width (meters)

constant (1.4 for overtopping failures, 1.0 for piping)
reservoir volume at time of failure (cubic meters)
height of the final breach (meters)
breach formation time (hours)
Froehlich states that the average side slopes should be:

1.4H:1V overtopping failures
0.9H:1V otherwise (i.e., piping/seepage)
While not clearly stated in Froehlich's paper, the height of the breach is normally calculated
by assuming the breach goes from the top of the dam all the way down to the natural ground
elevation at the breach location.
Froehlich (2008): In 2008, Dr. Froehlich updated his breach equations based on the addition
of new data. Dr. Froehlich utilized 74 earthen, zoned earthen, earthen with a core wall (i.e.,
clay), and rockfill data sets to develop as set of equations to predict average breach width, side
slopes, and failure time. The data that Froehlich used for his regression analysis had the
following ranges:
Volume of water at breach time: 0.0139 660.0 m3 x 106 ( 11.3 - 535,000 acre-feet)
(with 86% < 25.0 m3 x 106, and 82% < 15.0 m3 x 106)
Froehlich's regression equations for average breach width and failure time are:
Bave = 0.27 K o Vw0.32 hb0.04
tf = 63.2
where:
Bave
Ko
Vw
hb
g
tf
=
=
=
=
=
=
Vw
ghb2

constant (1.3 for overtopping failures, 1.0 for piping)
gravitational acceleration (9.80665 meters per second squared)
breach formation time (seconds)
17
Froehlich's 2008 paper states that the average side slopes should be:
1.0 H:1V overtopping failures
0.7 H:1V otherwise (i.e., piping/seepage)
While not clearly stated in Froehlich's paper, the height of the breach is normally calculated
by assuming the breach goes from the top of the dam all the way down to the natural ground
elevation at the breach location.
MacDonald and LangridgeMonopolis (1984): MacDonald and Langridge-Monopolis
utilized 42 data sets (predominantly earthfill dams, earthfill dams with a clay core, rockfill
dams) to develop a relationship for what they call the "Breach Formation Factor". The
Breach Formation Factor is a product of the volume of water coming out of the dam and the
height of water above the dam. MacDonald and Langridge-Monopolis then related the
breach formation factor to the volume of material eroded from the dam's embankment. The
data that MacDonald and Langridge-Monopolis used for their regression analysis had the
following ranges:
Breach Outflow Volume: 0.0037 660.0 m3 x 106 (3 - 535,000 acre-feet)
(with 79% < 25.0 m3 x 106, and 69% < 15.0 m3 x 106)
The following is the MacDonald and Langridge-Monopolis equation for volume of material
eroded and breach formation time, as reported by Wahl (1998):
For earthfill dams:
Veroded = 0.0261 (Vout * hw )
0.769
tf = 0.0179 (Veroded )
0.364
For earthfill with clay core or rockfill dams:
Veroded = 0.00348 (Vout * hw )
0.852
where:
Veroded = volume of material eroded from the dam embankment (cubic meters)
Vout
= volume of water that passes through the breach (cubic meters); for example,
storage volume at time of breach plus volume of inflow after breach begins,
minus any spillway and gate flow after breach begins.
hw
= depth of water above the bottom of the breach (meters).
tf
= breach formation time (hours).
The value of the Vout parameter is not exactly known before performing the breach analysis,
as it is the volume of water that passes through the breach (not including flow from gates,
spillways, and overtopping of the dam away from the breach area). A good first estimate is
the volume of water in the reservoir at the time the breach initiates. Once a set of parameters
18
are estimated, and a breach analysis is performed, the user should go back and try to make a
better estimate of the actual volume of water that passes through the breach. Then
recalculate the parameters with that volume. The recalculation of the volume makes the
method iterative. The actual breach dimensions are a function of the volume eroded.
MacDonald and Langridge-Monopolis stated that the breach should be trapezoidal with side
slopes of 0.5H:1V. The breach size is computed by assuming the breach erodes vertically to
the bottom of the dam and it erodes horizontally until the maximum amount of material has
been eroded or the abutments of the dam have been reached. The base width of the breach
can be computed from the dam geometry with the following equation (State of Washington,
1992):
Wb =
where:
Wb =
hb =
C
=
Z3 =
Z1 =
Z2 =
Zb =
Veroded hb2 (CZ b + hb Z b Z 3 / 3)

hb (C + hb Z 3 / 2)
bottom width of the breach (meters)

height from the top of the dam to bottom of breach (meters)
crest width of the top of dam (meters)
Z1 + Z2
average slope (Z1:1) of the upstream face of dam
average slope (Z2:1) of the downstream face of dam
side slopes of the breach (Zb:1), 0.5 for the MacDonald method
Note: MacDonald and Langridge-Monopolis stated that the equation for the breach
formation time is an envelope of the data from the earthfill dams. An envelope equation
implies that the equation will tend to give high estimates (too long) of the actual breach time
(for homogenous earthfill dams). Wahl's study states this method will over predict times in
some cases, while many equations will under predict.
Von Thun and Gillette (1990): Von Thun and Gillette used 57 dams from both the
Froehlich (1987) paper and the MacDonald and Langridge-Monopolis (1984) paper to
develop their methodology. The method proposes to use breach side slopes of 1.0H:1.0V,
except for dams with cohesive soils, where side slopes should be on the order of 0.5H:1V to
0.33H:1V. The data that Von Thun and Gillette used for their regression analysis had the
following ranges:
Volume of water at breach time: 0.027 660.0 m3 x 106 ( 22 - 535,000 acre-ft)
(with 89% < 25.0 m3 x 106, and 84% < 15.0 m3 x 106)
The Von Thun and Gillette equation for average breach width is:
Bave = 2.5 hw + Cb
where:
Bave = average breach width (meters)
hw = depth of water above the bottom of the breach (meters)
Cb = coefficient, which is a function of reservoir size, see the following table.
19
Reservoir Size
(cubic meters)
< 1.23 106
6
1.23 10 6.17 106
6.17 106 1.23 107
> 1.23 107
Cb
(meters)
6.1
18.3
42.7
54.9
Reservoir Size
(acre-feet)
< 1,000
1,000 - 5,000
5,000 - 10,000
> 10,000
Cb
(feet)
20
60
140
180
Von Thun and Gillette developed two different sets of equations for the breach development
time. The first set of equations shows breach development time as a function of water depth
above the breach bottom:
t f = 0.02 hw + 0.25
t f = 0.015 hw
(erosion resistant)
(easily erodible)
where:
= breach formation time (hours)
tf
hw = depth of water above the bottom of the breach (meters)
The second set of equations shows breach development time as a function of water depth
above the bottom of the breach and average breach width:
tf =
tf =
Bave
4 hw
Bave
4 h w + 61.0
(erosion resistant)
(easily erodible)
where:
Bave = average breach width (meters)
Note: Von Thun and Gillette's breach formation time equations are presented for both
"erosion resistant" and "easily erodible" dams. Von Thun and Gillette's paper states: "It is
suggested that these limits be viewed as upper and lower bounds corresponding respectively
to well-constructed dams of erosion resistant materials and poorly-constructed dams of easily
eroded materials".
Xu and Zhang (2009): In 2009 a paper was published by Dr.'sY. Xu and L.M. Zhang in the
Journal of Geotechnical and Geo-Environmental Engineering. The database gathered by
Dr.'s Xu and Zang contained 182 earth and rockfill dams from the United States and China,
with nearly 50 percent of the dams greater than 15 meters in height. However, their final
equations are based on a much smaller subset of these dams due to missing data. Their paper
shows details for 75 dams that were comprised of homogeneous earth fill, zoned-filled, dams
with corewalls, and concrete faced dams. Their final equation for the average breach width
is based on 45 dam failures, and their equation for the time of failure is based on only 28 dam
20
failures. The data that Xu and Zhang used for their regression analysis had the following
ranges:
Volume of water at breach time: 0.105 660.0 m3 x 106 (11.3 - 535,000 acre-feet)
(with 80% < 25.0 m3 x 106, and 67% < 15.0 m3 x 106)
Xu and Zhangs regression equation for average breach width is:
h
Bave
= 0.787 d
hb
hr
where:
Bave =
Vw =
hb =
hd =
hr
=
hw
B3
b3
=
=
=
b4
b5
=
=
0.133
Vw1 / 3
h
w
0.652
e B3

height of the Dam (meters)
fifteen meters, is considered to be a reference height for distinguishing large dams
from small dams
height of the water above the breach bottom elevation at time of breach (meters)
b3+b4+b5 coefficient that is a function of dam properties
-0.041, 0.026, and -0.226 for dams with corewalls, concrete faced dams, and
homogeneous/zoned-fill dams, respectively
0.149 and -0.389 for overtopping and seepage/piping, respectively.
0.291, -0.14, and -0.391 for high, medium, and low dam erodibility, respectively
Height of the Breach (hb). While Xu and Zhang present an equation for the height of the
breach (hb), the coefficient of determination, R2 was only 0.35 for their best equation. This is
a very poor correlation, and therefore it is suggested to assume the breach height goes from
the top of the dam all the way down to the natural ground elevation at the breach location
(i.e., set hb = hd). Additionally, Xu and Zhang's equation for breach height can produce
breach heights greater than the height of the dam, which implies a scour hole forming. While
this can happen, it is not appropriate to use this as the breach height in a model like
HEC-RAS, as it is applying the weir equation to the full breach shape. If the scour hole is
included in the breach height, you would over predict the outflow out of the dam, as the
middle of the scour hole is not the hydraulic control for water leaving the dam, and thus too
large of a flow area would be used in the computations.
The Xu and Zhang paper does not provide estimates for side slopes directly. Instead, they
provide an equation to estimate the top width of the breach, which can then be used with the
average breach width, to compute the corresponding side slopes. Here is their equation for
the breach top width:
21
h
Bt
= 1.062 d
hb
hr
0.092
Vw1 / 3
h
w
0.508
e B2
where:
Bt = breach top width (meters)
B2 = b3+b4+b5 coefficient that is a function of dam properties
b3 = 0.061, 0.088, and -0.089 for dams with corewalls, concrete faced dams, and
homogeneous/zoned-fill dams, respectively.
b4 = 0.299 and -0.239 for overtopping and seepage/piping, respectively.
b5 = 0.411, -0.062, and -0.289 for high, medium, and low dam erodibility, respectively.
Breach side slopes can be computed with the following equation:
Z=
Bt Bave
hb
Important Note: Xu and Zhang data used in the development of the equation for breach
development time include s more of the initial erosion period and post erosion period than
what is generally used in HEC-RAS for the critical breach development time. In general, this
equation will produce breach development times that are greater than the other four equations
described above. Because of this fact, the Xu and Zhang equation for breach development
time should not be used in HEC-RAS. However, it is shown here for completeness of their
method:
h
= 0.304 d
Tr
hr
Tf
where:
Tf =
Tr =
Vw =
hd =
hr
=
hw =
B5 =
b3 =
b4
b5
=
=
0.707
1.228
Vw1 / 3
h
w
e B5
breach formation time (hours)

1 hour (unit duration)
height of the dam (meters)
fifteen meters, which is considered to be a reference height for distinguishing large
dams from small dams
height of the water above the breach bottom elevation at time of breach (meters)
b3+b4+b5 coefficient that is a function of dam properties
-0.327, -0.674, and -0.189 for dams with corewalls, concrete faced dams, and
-0.579 and -0.611 for overtopping and seepage/piping, respectively
-1.205, -0.564, and 0.579 for high, medium, and low dam erodibility, respectively
22
Simplified Physical Breaching Method

The Simplified Physical breaching method in HEC-RAS allows the user to enter velocity versus
breach down-cutting and breach widening relationships, which are then used dynamically to
figure out the breach progression versus the actual velocity being computed through the breach,
on a time step by time step basis. The main data requirement differences between this method
and the "User Entered Data" breach method are the following:
Max Possible Bottom Width - This field is now used to enter a maximum possible breach
bottom width. This does not mean this will be the final breach bottom width; it is really
being used to limit the breach bottom width growth to this amount. The actual bottom width
will be dependent on the velocity verses erosion rate data entered, and the hydraulics of flow
through the breach. This field is used to prevent breaches from growing larger than this user
set upper limit during the run.
Min Possible Bottom Elev - This field is used to put a limit on how far down the breach can
erode during the breaching process. This is not necessarily the final breach bottom elevation;
it is a user entered limiter (i.e., the breach cannot go below this elevation). The final breach
elevation will be dependent on the velocity verses erosion rate data entered, and the
hydraulics of flow through the breach.
Starting Notch Width or Initial Piping Diameter - If the overtopping failure mode is
selected, the user will be asked to enter a starting notch width. The software will use this
width at the top of the dam to compute a velocity, from the velocity it will get a down cutting
erosion rate (based on user entered data), which will be used to start the erosion process. If a
piping failure model is selected, the user must enter an initial piping diameter. Once the
breach is triggered to start, the initial breach hole will show up immediately. A velocity will
be computed through it, and then the down cutting and widening process will begin based on
user entered erosion rate data.
Mass Wasting Feature - This option allows the user to put a hole in the dam or the levee at
the beginning of the breach, in a very short amount of time. This option would probably
most often be used in a levee evaluation, in which a section of the levee may give way (Mass
Wasting), then that initial hole would continue to erode and widen based on the erosion
process. The required data for this option is a width for the mass wasting hole; duration in
hours that the mass wasting occurs over (this would normally be a short amount of time);
and, finally the bottom elevation of the initial mass wasting hole (it is assumed that the hole
is open all the way to the top of the levee or dam if this option is used).
Velocity versus. Downcutting and Widening Erosion Rates - When using the Simplified
Physical breaching option, the user is required to enter "Overtopping Downcutting" (velocity
vs. downcutting erosion rates), as well as "Widening Relationship" (velocity vs. erosion
widening rates). An example of the required data input for this method is shown in Figure 9.
As shown in Figure 9, the user is required to enter "Overtopping Downcutting " and "
Widening Relationship". This data is often very difficult to come by. Users will need to
consult with geo-technical engineers to come up with reasonable estimates of this data for a
23
Figure 9. HEC-RAS Simplified Physical Breach Option
simulating a historic levee or dam breach, and adjusting the velocity versus erosion rate data
until the model simulates the correct breach width and time. This is obviously an iterative
process, and may require the user to perform this at multiple locations to see if there is a
consistent set or erosion rates that will provide a reasonable model for simulating levee
breaches (or dams) in your geographical area. We realize that this data is not readily
available for any specific levee or dam. The hope is that over time we will be able to develop
guidelines for these erosion rates based on analyzing historical levee and dam breaches.
Physically-Based Breach Computer Models
Several computer models have been developed that attempt to model the breach process using
sediment transport theories, soil slope stability, and hydraulics. Mr. Wahl summarized some of
these models in his report (Wahl, 1998). A table from Wahl's (1998) report, which summarizes
the physically based computer models he reviewed, is shown in Table 4.
In general, all of the models listed in Table 4 rely on the use of bed-load sediment transport
equations, which were developed for riverine sediment transport processes. The use of these
models should be viewed as an additional way of "estimating" the breach dimensions and breach
development time.
Of all the models listed in Table 4, the BREACH model developed by Dr. Danny Fread (1988)
has been used the most for estimating dam breach parameters. Dr. Fread's model can be used for
constructed earthen dams as well as landslide formed dams. The model can handle forming
breaches from either overtopping or piping/seepage failure modes. The software uses weir and
orifice equations for the hydraulic computation of flow rates. The Meyer-Peter and Muller
sediment transport equation is used to compute transport capacity of the breach flow. Breach
24
Table 4. Physically-Based Embankment Dam Breach Computer Models

Sediment
Breach
Model and Year
Transport
Morphology
Parameters
Constant breach
Angle of repose,
Cristofano (1965)
Empirical formula
width
others
Harris and Wagner
Breach
(1967)
Schoklitsch
Parabolic breach
dimensions,
formula
shape
BRDAM (Brown
sediments
and Rogers, 1977)
Other
Features
Lou (1981);
Ponce and
Tsivoglou (1981)
Meyer-Peter and
Mller formula
Regime type
relation
Critical shear
stress, sediment
Tailwater effects
BREACH (Fread,
1988)
Meyer-Peter and
Mller modified by
Smart
Rectangular,
triangular, or
trapezoidal
Critical shear,
sediment
Tailwater effects,
dry slope stability
BEED (Singh and

Scarlatos, 1985)
Einstein Brown
formula
Rectangular or
trapezoidal
Sediments, others
Tailwater effects,
saturated slope
stability
FLOW SIM 1 and

FLOW SIM 2
(Bodine, undated)
Linear
predetermined
erosion;
Schoklitsch
formula option
Rectangular,
triangular, or
trapezoidal
Breach
dimensions,
sediments
enlargement is governed by the rate of erosion, as well as the collapse of material from slope
failures. Dr. Fread's model can handle up to three material layers (inner core, outer portion of the
dam, and a thin layer along the downstream face). The material properties that must be
described are: internal friction angle; cohesive strength, grain size of the material (D50), unit
weight, porosity, ratio of D90 to D30, and Manning's n. This software has been tested on a
limited number of data sets, but has produced reasonable results.
Additional research on the erosion process of earthen embankments that are overtopped is being
conducted in the United States as well as Europe. The Agricultural Research Service (ARS) has
been testing earthen embankment failures at sizes ranging from small scale laboratory models to
near prototype scale dams (up to seven feet high) for several years (Hanson, et al., 2003; Hassan,
et al., 2004). Similar tests have been performed in Norway for earthen dams, five to six meters
high, constructed of rock, clay, and glacial moraine (Vaskinn, et al., 2004). The hope is that this
research work will lead to the development of improved computer models of the breach process.
A dam safety interest group made up of U.S. Government agencies (USBR, ARS, USACE),
private industry, and Canadian and European research partners is currently evaluating new
technologies for simulating the breach process. The goal of this effort is to develop computer
simulation software that can model the dam breach process by progressive erosion for earthen
dams initiated by either overtopping flow or seepage. Computer models that are currently being
evaluated are: WinDAM (Temple, et al., 2006); HR-BREACH (Mohammed, 2002); and
FIREBIRD (Wang and Kahawita, 2006). Table 5 provides a summary of these models
capabilities (Wahl, 2009):
25
Table 5. Summary of Erosion Process Models Currently Under Development

Model
and
Embankment
Failure
Year
Types
Modes
Erosion Processes
Homogeneous with
Headcut formation on downstream face,
WinDAM
varying levels of
Overtopping deepening, and upstream advancement;
cohesiveness
lateral widening
Homogeneous cohesive,
Variety of sediment transport/erosion
or simple composite
equations and multiple methods of
embankments with
Overtopping
application. Discrete breach growth using
HR-BREACH noncohesive zones,
Piping
bending, shear, sliding and overturning
surface protection
failure of soil masses.
(grass or rock), and
cohesive cores
Homogeneous cohesive
Coupled equations for hydraulics and
FIREBIRD
Overtopping
or noncohesive
sediment transport.
Peak Flow Equations and Envelope Curves

Several researchers have developed peak flow regression equations from historic dam failure
data. The peak flow equations were derived from data for earthen, zoned earthen, earthen with
impervious core (i.e., clay, concrete, etc.) and rockfill dams only, and do not apply to concrete
dams. In general, the peak flow equations should be used for comparison purposes.
Once a breach hydrograph is computed in HEC-RAS, the computed peak flow from the model
can be compared to these regression equations as a test for reasonableness. However, one should
use great caution when comparing results from these equations to model predictions. First, the
user should go back to the original paper for each equation and evaluate the data sets and
assumptions that were used to develop that equation. Many of the equations were developed
from limited data sets, and most were for smaller dams. Also, when using these equations to
compare against model results, the event being studied can have a significant impact on the
model results peak flow. For example, studies being performed with PMF inflows may have
larger computed peak outflows than what will be predicted by some of the peak flow equations.
This is due to the fact that none of the historic data sets were experiencing a PMF level flood
when they failed.
Shown below is a summary of some of the peak flow equations (all equations are in metric form)
that have been developed from historic dam failures:
USBR (1982): Q = 19.1(hw)1.85
(envelope equation)
MacDonald and Langridge-Monopolis (1984):

Q = 1.154(Vwhw)0.412
Q = 3.85(Vwhw)0.411
(envelope equation)
Froehlich (1995b): Q = 0.607Vw0.295hw1.24
26
Xu and Zhang (2009):

Q
5/3
w
gV
h
= 0.175 d
hr
0.199
Vw1 / 3
hw
1.274
e B4
Kirkpatrick (1977): Q = 1.268(hw+0.3)1.24

Soil Conservation Service (SCS,1981): Q=16.6hw1.85
Hagen (1982): Q=0.54(S hd)0.5
Singh & Snorrason (1984):
Q = 13.4(hd)1.89
Q = 1.776(S)0.47
Costa (1985):
Q = 1.122(S)0.57
Q = 0.981(S hd)0.42
Q = 2.634(S hd)0.44 (envelope equation)
Evans (1986): Q = 0.72Vw0.53
Walder and O'Connor (1997): Q estimated by computational and graphical method
using relative erodibility of dam and volume of reservoir.
where:
Q =
hw =
Vw =
S =
hd =
hr =
B4 =
b3 =
b4
b5
=
=
peak breach outflow (cubic meters per second)

depth of water above the breach invert at time of breach (meters)
volume of water above breach invert at time of failure (cubic meters)
reservoir storage for water surface elevation at breach time (cubic meters)
height of dam (meters)
fifteen meters, which is considered to be a reference height for distinguishing large
dams from small dams.
b3+b4+b5 coefficients that are a function of dam properties
-0.503, -0.591, and -0.649 for dams with corewalls, concrete faced dams, and
-0.705 and -1.039 for overtopping and seepage/piping, respectively
-0.007, -0.375, and -1.362 for high, medium, and low dam erodibility, respectively
In addition to the peak flow equations, one can also compare computed model peak outflows to
envelope curves of historic failures. One such curve is shown in Figure 10 (HEC, 1980).
When comparing computed results to the envelope curve shown in Figure 9, keep in mind that
this envelope curve was developed from only fourteen data sets, and may not be a true upper
bound of peak flow versus hydraulic depth.
27
Figure 10. Envelope of Experienced Outflow Rates from Breached Dams
Site Specific Data and Engineering Analysis

Site specific information about the dam should always be collected and evaluated. Site specific
information that may be useful in this type of analysis includes: materials/soil properties used in
building the dam; if the dam includes an impervious core/filter or not; material used for
impervious core/filter; embankment protection materials (rock, concrete, grass, etc.);
embankment slopes of the dam; historic seepage information; known foundation or abutment
problems; known problems/issues with gates and spillways; etc.
Whenever possible a geo-technical analysis of the dam should be performed. Geo-technical
evaluations can be useful in the selection of dam breach parameters. Specifically, geotechnical
analyses can be used to estimate appropriate breach side slopes based on soil material properties.
Additionally, a geo-technical analysis can be used to make a qualitative assessment of the breach
parameters estimated by the various methods described above (historic comparisons, regression
equations, and physically based model results).
Consideration of structural features such as spillway gates should also be considered for
determination of the appropriate breach geometry for failure modes involving gate malfunction,
blockage, or loss of the structure.
28
Recommended Approach
In general, several methods should be used to predict a range of breach sizes and failure times for
each failure mode/hydrologic event. It is recommended that the modeler select several
regression equations to estimate breach parameter values. Care must be taken when selecting
regression equations, such that the equations are appropriate for the dam being investigated.
Regression equations that have been used for earthen, zoned earth, earth with a clay core, and
rockfill dams are: Froehlich (1995a), Froehlich (2008), MacDonald and Langridge-Monopolis
(1984), Von Thun and Gillette (1990), and Xu and Zhang (2009). If the dam under investigation
is outside the range of data used in the development of the regression equations, resulting breach
parameter estimates should be scrutinized closely for reasonableness. Note: Never mix and
match breach parameters from multiple regression equations. In other words, use the average
breach width and time of failure from the same equation set. Do not use a breach width from one
equation set and a time of failure from another. The breach widths and times are interrelated, as
they are derived from a specific data set.
In addition to the regression equations, physically based computer models should also be utilized
if appropriate for the level of study (NWS-BREACH, WinDAM, and HR-BREACH models are
currently recommended). Whenever possible, geotechnical analyses of the dam should be used
to assist in estimating the breach parameters (i.e., side slopes of the breach), or at least used as a
qualitative assessment of the estimates. Additionally, breach parameter estimates should be
compared to the government agency ranges provided in Table 3. If values are outside the
recommended ranges, those estimates may need to be adjusted, unless there is compelling
physical evidence that the values are appropriate. This will lead to a range of values for the
breach size and failure times. A sensitivity analysis of breach parameters and times should be
performed by running all of the parameter estimates within a HEC-RAS model.
Each set of breach parameters and failure times will produce a different outflow hydrograph.
However, once these hydrographs are routed downstream, they will tend to converge towards
each other. There are two main reasons for this convergence: (1) the total volume of water in
each of the different hydrographs is basically the same (being the stored water behind the dam at
the time of failure, plus whatever inflow occurs); (2) as the hydrographs move downstream, a
sharp hydrograph will attenuate much more quickly than a flat hydrograph. Hydrographs from
different assumed breach parameters can converge to produce similar peak flow and stage in a
surprisingly short distance. An example flow versus time plot from a study performed with
HEC-RAS is shown in Figure 11. However these differences could be huge for loss-of-life
calculations if a population at risk is immediately downstream of the dam.
In the example shown in Figure 10, three different sets of breach parameters were used for the
same model. The hydrographs coming out of the dam are very different in magnitude of peak
flow, but they have the same volume of water. In this example, as the hydrographs move
downstream they have substantially converged within four miles and are almost the same peak
flow by River Mile 10. The rate at which the hydrographs will converge is dependent on many
factors: steepness in the rise of the outflow hydrograph, volume of the outflow hydrograph, slope
of the downstream reach, roughness of the downstream reach, available storage in the
downstream floodplain, etc. The user will need to route all of the breach outflow hydrographs
29
Figure 11. Dam Break Flood Wave Progression Downstream
downstream through the entire study area in order to fully evaluate the affect of the breach
parameters on the resulting flood hydrographs and inundation levels.
For a risk assessment study, the user must select the set of breach parameters that are considered
to be most likely for each event/pool elevation. This will require engineering judgment. If all of
the breach estimates, for a given event/pool elevation, end up converging to the same flow and
stage before getting to any population at risk and potential damage areas, then the selection of a
final set of breach parameters should not affect the computations and a simple mean value should
be used. However, if the various sets of breach parameters produce significantly different flow
and stage values at downstream locations (population at risk locations and potential damage
zones), then engineering judgment will need to be used to pick a set of values that are considered
most likely. Conservatively high or low values should not be used, as this will bias the overall
results.
Once a final set of breach parameters is selected for a given event/failure mode, the computed
peak outflow from the breach can be compared to some of the peak flow equations as a check of
reasonableness. Keep in mind the limitations of the peak flow equations, as discussed in the
Peak Flow Regression Equations section (see page 26).
Another check for reasonableness should be done by evaluating the breach flow and velocities
through the breach, during the breach formation process. This can be accomplished by
reviewing the detailed output for the inline structure (dam) and reviewing the flow rate and
velocities going through the breach. This output is provided on the HEC-RAS detailed output
table for the inline structure. There are two things to check for here:
1) if the model reaches the full breach development time and size, and there are still very
high flow rates and velocities going through the breach, this is a sign that either the
30
breach is too small, or the development time is to short (unless there are some physical
constraints limiting the size of the breach);
2) if the flow rate and the velocities through the breach become very small before the
breach has reached its full size and development time, then this is an indicator that the
breach size may be too large, or the breach time may be too long.
Additional factors affecting this could be the breach progression curve and the hydraulic
coefficients (weir and piping) used. When you get into the situation described above in either
Scenario 1 or 2, the breach size and development time should be re-evaluated to improve the
estimates for that particular structure.
The level of effort in estimating breach parameters should be consistent with the type of risk
assessment. In general, the level of effort and detail will increase from Type 1 (Low Hazard
Potential) through Type 3 (High Hazard Potential). For Type 1 analyses a basic estimate of
breach parameters consistent with the range of values in Table 3 could be appropriate. Type 2
(Significant Hazard Potential) and Type 3 analyses will typically require a greater level of detail
and accuracy incorporating most if not all of the methods are provided in this section.
Example Application
In order to demonstrate how to estimate breach parameters, an example application for a
fictitious dam is provided below. The event being evaluated in the example is a PMF scale
event. This same process needs to be performed for each failure mode/event (fully modeled
hydrologic event or pool elevation for sunny day failures). The following is the necessary
information required about a dam in order to develop breach parameter estimates as outlined in
these guidelines.
Reservoir Data
Important Pool Elevations

Stream Bed
Multipurpose Pool
Top of Flood Control
Top of Dam
PMF Max Water Surface
Elevation
(meters)
1678.0
1692.1
1710.0
1720.9
1722.26
Dam Embankment Data

Crest Length: 4360 meters
Crest Width: 9.15 meters
Maximum Height above river bed: 42.9 meters
Average Upstream Embankment slope: 3.3H:1V
Average Downstream Embankment slope: 3.3H:1V
Embankment Material: Rolled earth, zoned
Embankment Core: Impervious core, clay
31
Volume
(m3)
0.0
15.81x106
151.64x106
327.01x106
357.98x106
Upstream slope Protection: 18" riprap

Downstream slope protection: Topsoil and grass
Regression Equations
For this example, the Froehlich (1995a), Froehlich (2008), MacDonald and LangridgeMonopolis (1984), Von Thun and Gillette (1990), and Xu and Zhang (2009) regression equations
for predicting breach size and development time were used. This dam is within the range of the
data used to develop these regression equations, therefore the equations are considered to be an
appropriate methodology for estimating the breach parameters. During the PMF event for this
dam it is overtopped by 1.36 meters. The mode of failure for this example will be assumed as an
overtopping failure. The failure location is assumed to be at the main channel centerline. The
breach bottom elevation is assumed to be at an elevation of 1,678 meters (invert of the main
channel). The water surface elevation at the initiation of the breach will be at an elevation of
1,722.26 meters (maximum pool for PMF event). The following are the calculations for each
method.
Froehlich (1995a):
Bave = 0.1803 Ko Vw0.32 hb0.19
Bave = 0.1803 (1.4) (357.98x106)0.32 (42.9)0.19
Bave = 281.5 meters
tf = 0.00254 Vw0.53 hb-0.90
tf = 0.00254 (357.98x106)0.53 (42.9)-0.90
tf = 2.95 hours
The Froehlich (1995a) method assumes a side slope of 1.4H:1V for an overtopping breach.
Given the breach height of 42.9 meters, this yields a bottom width for the breach of
Wb = 221.4 meters.
Froehlich (2008):
Bave = 0.27 Ko Vw0.32 hb0.04
Bave = 0.27 (1.3) (357.98x106)0.32 (42.9)0.04
Bave = 222.76 meters
tf = 63.2 ( Vw /(ghb2))0.5
tf = 63.2 (357.98x106/(9.80665 x (42.9)2))0.5
tf = 2.47 hours
The Froehlich (2008) method assumes a side slope of 1.0H:1V for an overtopping breach.
Given the breach height of 42.9 meters, this yields a bottom width for the breach of
Wb = 179.86 meters.
MacDonald and Langridge-Monopolis (1984): The MacDonald and Langridge-Monopolis
equation for an earthfill dam with a clay core is:
32
Veroded = 0.00348 (Vout * hw)0.852

Since the outflow volume through the breach is unknown before performing the analysis, a
good starting estimate is the volume of water in the dam at the peak stage of the event.
Veroded = 0.00348 (357.98x106 * 44.26 )0.852
Veroded = 1.70556x106 cubic meters of material
To compute the bottom width of the breach, the method says to use side slopes of 0.5H:1V.
The user must also estimate an average side slope for both the upstream and downstream
embankment of the dam. For this example average side slopes of 3.3H:1V were used for
both upstream and downstream. The bottom width equation (State of Washington, 1992) is:
Wb =
Veroded hb2 (CZ b + hb Z b Z 3 / 3)

hb (C + hb Z 3 / 2)
where:
Wb = (1.70556x106 42.92(9.15*0.5 + 42.9*0.5*6.6/3))/(42.9(9.15 + 42.9*6.6/2))
Wb = 249.0 meters
tf = 0.0179 (Veroded)0.364
tf = 0.0179 (1.70556x106)0.364
tf = 3.32 hours
Note: Once an actual breach hydrograph is computed with the MacDonald and LangridgeMonopolis parameters, the volume of water coming out of the breach should be calculated,
and the parameters should be re-estimated using that volume of water for Vout.
Von Thun and Gillette (1990): The Von Thun and Gillette equation for the breach average
width is:
Bave = 2.5 * hw + Cb
Bave = 2.5 * 44.26 + 54.9
Bave = 165.6 m
Von Thun and Gillette suggest using breach side slopes of 0.5H:1V for earthen dams with a
clay core. Given the dam height of 42.9 meter, the Breach bottom width will be Wb = 144.2
meters.
Von Thun and Gillette show two equations for predicting the breach failure time. One
equation is a function of the depth of water only, while the other is a function of depth of
water and the computed average breach width. Both equations are used below.
tf = 0.02 * hw + 0.25
tf = 0.02 * 44.26 + 0.25
tf = 1.14 hours
tf = Bave/(4*hw)
tf = 166/(4*44.26)
tf = 0.94 hours
33
Both of the Von Thun and Gillette equations yield similar answers for the breach time.
Reviewing the Von Thun and Gillette paper showed that the data they used in their
experiments were mostly earthen embankments with slightly cohesive materials. Given that
the example dam we are studying has an engineered clay core, the longer time estimate is
probably more appropriate. Therefore the selected failure time is tf = 1.14 hours.
Xu and Zhang (2009): The Xu and Zhang equation for the breach average width is:
h
Bave
= 0.787 d
hb
hr
0.133
Vw1 / 3
hw
0.652
e B3
Bave = (42.9)(0.787)(42.9/15)0.133((357.98x106)1/3/44.26)0.652 e-0.283

Bave = 178.67 meters
h
Bt
= 1.062 d
hb
hr
0.092
Vw1 / 3
h
w
0.508
e B2
Bt = (42.9)(1.062)(42.9/15)0.092((357.98x106)1/3/44.26)0.508 e0.071
Bt = 220.64 meters
Based on the computation of Bave and Bt abobe, the breach bottom width for this method is
Wb = 136.7 and the side slopes are Z = 0.98H:1V.
The breach development time from the Xu Zhang equation is as follows:
h
= 0.304 d
Tr
hr
Tf
0.707
1.228
Vw1 / 3
hw
e B5
Tf = (1.0)(0.304)(42.9/15)0.707((357.98x106)1/3/44.26)1.228 e-0.327
Tf = 13.92 hours*
*Note: Please see note about the Xu Zhang method over estimating the breach time under
the method description above.
Physically-Based Breach Computer Models
For this example, only Dr. Fread's NWS-BREACH model was run to make an estimate of breach
parameters from a physically based computer model. The physical dimensions of the dam, the
soil properties, and the hydrologic event data were entered into the BREACH model. The results
from the BREACH model for this example are:
Breach Bottom Width Wb
Breach side slopes
Breach Failure Time tf
238 meter
0.9H:1V
4.2 hours
34
Summary Results for Breach Parameters

Shown in Table 6 is a summary of the breach parameters computed from the regression
equations and the NWS-BREACH model.
Table 6. Summary of Breach Parameter Estimates
Breach
Bottom
Width
(meters)
Method
Froehlich (1995a)
Froehlich (2008)
MacDonald and Langridge-Monopolis
Von Thun and Gillette
Xu and Zhang (2009)
NWS-BREACH Computer Model
Breach
Side
Slopes
(H:1V)
221.4
179.9
253.0
144.2
136.7
238.0
Breach
Failure
Time
(hours)
1.40
1.00
0.50
0.50
0.98
0.90
2.95
2.47
3.32
1.14
13.92*
4.2
*Note: the data Xu and Zhang used in the development of their equation for breach development time includes more of the initial erosion period
and post erosion period than what is generally used in HEC-RAS for the critical breach development time. In general, this equation will produce
breach development times that are greater than the other four equations described above. Because of this fact, the Xu Zhang equation for breach
development time should not be used in HEC-RAS.
From here, all six sets of parameters should be entered into the HEC-RAS software and run as
separate breach plans. This will result in six different breach outflow hydrographs. However,
once the hydrographs are routed downstream, they will begin to converge towards each other.
The selection of a final set of breach parameters for this event should be based on guidance
provided in the Recommended Approach section (see page 29).
Downstream Flood Routing/Modeling Issues

The modeling of a dam break flood wave is one of the most difficult unsteady flow problems to
solve. Previous discussions in this document have focused on modeling the reservoir pool, the
dam itself, and estimating breach parameters to be used in computing the breach outflow
hydrograph. However, the most difficult part of performing a dam safety study is routing the
dam break flood wave downstream.
Within HEC-RAS, the user can model the downstream area in the following manner: as a
combination of one-dimensional streams and storage areas; a combination of one-dimensional
streams, storage areas, and two-dimensional flow areas; or as a single two-dimensional flow
area. There are many things that the hydraulic modeler must consider to get an accurate estimate
of the downstream flood stages and flows. The following is a list of things that should be
considered when developing an unsteady flow model for a dam break application. Most of these
issues are concerns for one-dimensional river reach modeling with cross sections.
Cross Section Spacing and Hydraulic Properties

Computational Time Step
Manning's Roughness Coefficients
Downstream Storage, Tributaries, and Levees
Modeling Bridge and Culvert Crossings
Modeling Steep Streams
35
Drops in the bed Profile

Initial Conditions (low flow)
Downstream Boundary Conditions
Cross Section Spacing and Hydraulic Properties

Cross-sectional cut lines should be created to capture the entire extent of flooding anticipated by
the dam break scenario. As in any hydraulic modeling study, cross sections must be laid out to
accurately describe the channel and floodplain geometry. Cross sections are laid out
perpendicular to the anticipated flow lines of both the channel and the floodplain, during high
flow conditions. There must be enough cross sections to describe: contractions and expansions
of the channel and/or floodplain; changes in bed slope; changes in roughness; and significant
changes in discharge. Cross sections also need to be added immediately upstream and
downstream of: tributary inflow locations; dams and other inline structures (weirs, drop
structures, or natural drops in the bed profile); bridge and culvert crossings; levees and other
types of lateral hydraulic structures. An example of a cross section layout is shown in Figure 12.
Figure 12. Example Cross Section Layout (Ackerman, 2014)
In addition to describing the physical changes and hydraulic structures within the channel and
floodplain, there are also numerical considerations for adding or removing cross sections.
Cross Sections Spaced To Far Apart. In general, cross sections spaced too far apart will cause
additional numerical diffusion of the floodwave, due to the derivatives with respect to distance
being averaged over to long of a distance. See an example of artificial numerical diffusion in
Figure 13. Figure 13 shows an upstream inflow hydrograph and two downstream hydrographs
after they have been routed through the river system. In this example, the channel is a
rectangular channel on a constant slope, with a constant Manning's roughness. The only change
in the example is the cross section spacing.
36
Figure 13. Numerical Error Due to Cross Section Spacing
Additionally, when cross sections are spaced far apart, and the changes in hydraulic properties
are great, the solution can become unstable. Instability can occur when the distance between
cross sections is so great, such that the Courant number becomes much greater than 1.0, and
numerical errors grow to the point of the model becoming unstable. Another way to say this is
that the cross section spacing is not commensurate with the hydrograph being routed and the
computational time step being used (i.e., the cross section spacing is much further than the flood
wave can travel within the computational time step being used).
Maximum Cross Section Spacing. A good starting point for estimating maximum cross section
spacing are two empirically derived equations by Dr. Danny Fread (Fread, 1993) and P.G.
Samuels (Samuels, 1989). These two equations represent very different methods for coming up
with spacing. Samuels' equation implies that smaller streams and steeper streams will require
tighter cross section spacing. In general, Samuels' equation was derived for typical flood studies,
in which the modeler is developing a steady state model for a typical floodplain study of the twoyear through 100 year events. For dam break flood studies, Samuels' equation may be too strict,
in that it requires much tighter cross section spacing than needed. Samuels' equation is as
follows:
0.15 D
S0
where:
x = the cross section spacing distance (feet)
D
= the average main channel bankfull depth (feet)
S0 = the bed slope (feet/feet)
Note: Samuels' equation was derived from data with slopes ranging from two to fifty feet/miles.
Dr. Fread's equation implies smaller streams and steeper hydrographs will require tighter cross
sections. Fread's equation is one set of three conditions he presented in his paper for determining
spacing. The equation is a theoretical derivation of spacing based on the inherent numerical
37
errors involved with linearizing the St. Venant Equations into a four-point implicit finitedifference scheme. The other two involve a check of the change in cross sectional area from one
cross section to the next, and the other accounts for changes in slope. Consequently, the spacing
determined by Freads equation may be too coarse, depending on the bed slope changes, the
contraction and expansion characteristics and other non-linear data. Dr. Fread's equation is as
follows:
cTr
20
where:
x = the cross section spacing distance (feet)
C
= the wave speed (feet per second)
Tr = time of rise (from low flow to peak) of the hydrograph (seconds)
Samuels' and Dr. Fread's equations are rough estimates of cross section spacing - a good place to
start. However, over time and practice, the modeler should be able to determine a good first
estimate based on experience.
Cross Sections Too Close Together. If the cross sections are too close together, then the
derivatives with respect to distance may be overestimated, especially on the rising side of the
flood wave. This can cause the leading edge of the flood wave to over steepen, to the point at
which the model may become unstable. An example of this is shown in Figure 14. In this
example, the only change made to the model was that cross sections were interpolated at very
short intervals (five feet). If it is necessary to have cross sections at such short intervals, then
much smaller time steps will need to be used in order for the numerical computations to solve the
equations over such short distances. In general, for most dam break flood studies, cross sections
should not be spaced at intervals closer than about 50 feet, unless you can use very small time
steps (i.e., a few seconds or less). However, cross sections can be placed at closer distances at
hydraulic structures, such as bridges/culverts, dams, and inline weirs, due to the fact that the
model does not solve the unsteady flow equations through these structures. Rather it uses
hydraulic equations specifically defined for those structures.
Computational Time Step

In the development of any unsteady flow model, stability and numerical accuracy can be
improved by selecting a time step that satisfies the Courant Condition. This is very important for
a dam break model. Too large a time step will cause numerical diffusion (attenuation of the
peak) and possibly model instability. Too small of a time step can lead to very long computation
times, as well as possible model instability.
Too large of a time step: When the solution scheme solves the unsteady flow equations,
derivatives are calculated with respect to distance and time. If the changes in hydraulic
properties at a given cross section are changing rapidly with respect to time, too large of a time
step may cause over estimation (too steep) of the time based derivatives, causing the program to
go unstable. The solution to this problem in general is to decrease the time step. Even if the
program does not go unstable, too large of time steps will cause numerical attenuation of the
hydrograph that is not physically related. An example of a model with varying time steps is
shown in Figure 15. In this example, all things in the model were exactly identical, except one
38
Figure 14. Example Model Instability due to Very Short Cross Section Spacing
run was done with a one minute time step (appropriate for this model), and the other was done
with a ten minute time step (too large for this model). As shown in Figure 15, the run with the
ten minute time step has a ten percent lower peak flow, and the flood wave is much more spread
out (diffused) than the run with the one minute time step.
Too Small of a Time Step. If a time step is selected that is much smaller than what the Courant
Condition would suggest for a given flood wave, then model runs times will be much longer than
necessary, and this can also cause model stability problems. In general to small of a time step
will cause the leading edge of the flood wave to steepen, possible to the point of oscillating and
going unstable. Extremely small time steps (less than one second) can cause round off errors
when storing numbers in the computer, which in turn can lead to numerical errors which can
grow over time.
Time Step Selection. As mentioned above, the best way to estimate a computational time step
for HEC-RAS is to use the Courant Condition. This is especially important for dam break flood
studies. The Courant Condition is the following:
C=
Vw T
1
X
and therefore:
x
Vw
39
River: Test
Reach: 1
RS: 0
120000
Le ge nd
10 min DT
1 min DT
100000
Flow (cfs)
80000
60000
40000
20000
0
0200
0300
0400
0600
0500
01Jan2007
Time
0700
0800
Figure 15. Example of Varying Computational Time Step
where:
C
T
x
Vw
=
=
=
=
courant number
time step (seconds)
distance step in feet (average cross section spacing or two-dimensional cell size)
wave speed (feet per second
The flood wave speed is based on capturing the speed of the rising side of the flood wave as it
propagates downstream. Flood wave speed is most accurately calculated in the area of the initial
rise of the flood wave, where there is the largest change in discharge with respect to the change
in cross sectional area (this is the leading edge of the dam break flood wave). The equation for
calculating flood wave speed is:
Vw =
dQ
dA
where:
Vw = flood wave speed (feet per second)
dQ = the change in discharge over a short time interval (Q2 Q1)
dA = the change in cross section area over a short time interval (A2 A1)
Note: dQ/dA can be approximated by calculating the change in discharge and flow area at a
single cross section over a single computational time step. This should be done while the flood
wave initial abrupt rise is occurring at that cross section.
40
For practical applications of the Courant Condition, the user can take maximum average velocity
from HEC-RAS and multiply it by 1.5, to get a rough estimate of flood wave speed in natural
cross sections.
For medium to large rivers the Courant Condition may yield time steps that are too restrictive
(i.e., a larger time step could be used and still maintain accuracy and stability). A practical time
step can be estimated as:
Tr
20
However, treat this estimate as an upper limit. Remember that for dam break models, typical
time steps are in the range of one to sixty seconds due to the short time of rise and very fast flood
wave velocities.
Manning's Roughness Coefficients

Roughness coefficients represent the resistance to flow in channels and floodplains. Roughness
is usually presented in the form of a Manning's n value in HEC-RAS. There is extensive
research and literature on methods to determine n values; however most of this work is
representative of only main channels and not floodplains. Additionally, the literature on
Manning's n values is for historically experienced floods, which are much lower than the flood
resulting from a dam break. The actual selection of n values to be used for each dam assessment
will require judgment by the engineer responsible for hydraulic model development.
A proper perspective is required before establishing a range of n values to be used in USACE
risk assessment studies. The following general guidelines of factors that affect n values should
be considered in developing representative values.
Base Surface roughness: Often represented by the size and shape of surface or channel
and floodplain material that produces a friction effect on flow.
Stage and Discharge: The n value in most streams decreases with increase in stage and
discharge. However, this is not always the case. If the channel bed is of lesser roughness
than the channel banks, then the composite channel n values will increase with channel
stage. Once the stage gets higher than the main channel banks, the roughness coefficient
could begin to decrease. The main point here is that the variation of Manning's n with
stage is site specific.
Obstructions: Objects constructed in the channel or in overbanks such as bridge piers or
buildings can potentially cause increases in n value. It is especially difficult to estimate
Manning's roughness coefficients to represent buildings in the floodplain, as there are
many factors to consider: the area obstructed and the density of the buildings, direction of
the flow in relation to the layout of the structures, roughness of all of the other boundaries,
slope of the terrain, velocities of the flow, etc.
Irregularities: Variations in cross-section size and shape along the floodplain.

Irregularities are often caused by natural constrictions and expansions, sand deposition
41
and scour holes, ridges, projecting points and depressions, and holes and humps on the
channel bed. Gradual and uniform changes will generally not appreciably affect n value.
Whereas, areas that have lots of sharp channel irregularities will tend to have higher
Manning's roughness coefficients.
Channel alignment: Smooth curvature with large radius will generally not increase
roughness values, whereas sharp curvature with severe meandering will increase the
roughness.
Vegetation: Dependent on height, density, distribution, and type of vegetation. Heavily

treed areas can have a significant affect for dam failures. In general a lower average
depth results in a higher n value. High velocities can potentially flatten the vegetation
and lowering n values.
Silting, Scouring, and Debris: Silting may change a very irregular channel into a
comparatively uniform one and decrease n and scouring may do the reverse. During a
dam break flood wave, there will be a tremendous amount of scouring occurring, as well
as lots of debris in the flow. The increase sediment load and debris will cause the flow to
bulk up (increase in stage). One way to account for this increased sediment load and
debris is to increase the Manning's n values.
The resulting maximum water surface profile associated with the failure of a dam will often be
much higher than any historically observed flood profile. In such cases, there is no historical
based model data to calibrate to floods of this magnitude. It is therefore incumbent upon the
engineer to determine reasonable roughness coefficients for flows and stages that will be higher
than ever experienced. To gain a perspective on how each modeling parameter affects results, a
bounding type sensitivity analysis can be performed regardless of the methods used to establish n
values.
Historical regional knowledge of channels and floodplains should be used along with published
guidelines in establishing a base level set of n values. Guidelines for establishing base level
Manning's n values can be found in Chapter 3 of the HEC-RAS Hydraulic Reference Manual
(HEC, 2014a). The base level n values should be adjusted up or down based on factors addressed
previously. Calibration to the largest historical events of record should be done whenever
possible. Once adjusted roughness coefficients are established, uncertainty analyses should be
performed by varying all values (two additional computational runs) by plus or minus twenty
percent. In general, channel n values for risk assessment may be in the range of 0.025 to 0.075.
The overbank n values may range between 0.04 and 0.25. Note that higher n values can be used
in areas to allow for storage embayments with little to no conveyance.
Manning's n Values Immediately below Dam. Significant turbulence, sediment load and
debris should be expected for the immediate reach downstream of a failed dam. This is obvious
when viewing the photo of the Teton Dam failure shown in Figure 16. Because HEC-RAS does
not directly account for high volumes of sediment in the flow, and the extreme turbulence in the
water surface caused by the breach, it is often a good idea to increase the Manning's n values just
downstream of the dam. The increased sediment and turbulence will cause higher water surfaces
to occur. The only way to mimic this is by increasing the roughness coefficients. Proper
42
Figure 16. Significant Turbulence and Sediment Load during the Teton Dam Failure (Olsen, 1976)
modification and variation of n values is one of the many uncertainties in dam failure modeling.
An accurate assessment can be confidently attained only after previous knowledge of a particular
dam failure event. A reasonable modeling approach may be to assume double the normal n
value directly downstream of the dam and transition to normal roughness coefficients where
failure induced turbulence, sediment load, and debris transport are expected to recede.
Roughness Coefficients for Steep Streams. Many of our dams are located in mountainous
regions, where the slopes of the stream are significantly steep. It is very common to
underestimate Manning's n values for steep terrain. Underestimation of the roughness
coefficients can cause water surface elevations to be too low, increased velocities, and possibly
even supercritical flow. In addition to this, abrupt changes in n values or underestimation of n
values can cause the model to go unstable. Dr. Robert Jarrett (Jarrett, 1984) collected some
extensive field data on steep streams (slopes greater than 0.002 feet/feet) in the Rocky
Mountains. Dr. Jarrett measured cross sectional shape, flow rates, and water surface elevations
at 21 locations for a total of 75 events. From this data Dr. Jarrett performed a regression analysis
and developed an equation to estimate the Manning's roughness coefficient of the main channel.
n = 0.39 S 0.38 R 0.16

43
where:
n = Manning's roughness coefficient of the main channel
S = energy slope (slope of the energy grade line, feet/feet)
R = hydraulic radius of the main channel (feet).
While Dr. Jarrett's equation is not necessarily applicable to all locations, it is often a useful check
for reasonableness of the Manning's n values in steep terrain.
Downstream Storage, Tributaries, and Levees

Accounting for downstream storage in the floodplain below the dam is crucial in order to get a
reasonable estimate of the flood wave propagation and attenuation as it moves downstream.
General floodplain storage (areas that get wet but have little to no velocity) can often be modeled
as part of the normal cross section by using ineffective flow areas. If a portion of a cross section
is wet, but it will have a very low velocity, high Manning's n values are another approach to
modeling that area of the cross section.
Modeling Tributaries. Tributaries that come into the main river downstream may have flow
reversals during the passing of the flood wave. Significant size tributaries need to be accounted
for, since they may represent a large amount of storage volume taken out of the flood wave.
Further, the resultant inundation maps will need to include the flooding extent up the tributaries.
These factors require scrutiny when developing geometric data for HEC-RAS and can be
addressed in three different ways when laying out data for tributaries. A tributary may be
modeled using: (1) a separate one-dimensional river reach, (2) a two-dimensional flow area, (3) a
storage area, or (4) an extension of the main river cross sections.
The most comprehensive way to model the effects of a tributary to the main river is to model the
tributary with one-dimensional cross sections or a two-dimensional flow area. If the computed
water surface along the tributary results in a sloped water surface, then modeling the tributary as
a separate river reach is the preferred modeling method. Tributaries that have significant inflows
to the overall flood hydrograph are strong candidates to be modeled as separate one-dimensional
reaches, or part of a two-dimensional flow area that is being used to model the downstream area.
When adding a tributary to the main stem of a river, it is important to differentiate between the
contributing area of the main stem cross sections versus the contributing area of the tributary
cross sections. At the stream junction, if flow from the two reaches will mix, a decision will
need to be made as to the line that represents the separation point of the tributary and the main
stem flows. Cross sections from one reach should end just where the cross sections of the other
reach begin, to insure complete inundation mapping. Cross sections should not overlap. Figure
17 depicts a tributary included in the model as a separate reach. As shown in Figure 17, the user
must identify the point at which to end the main stem cross sections and begin the tributary cross
sections.
The next best option for accounting for tributary storage is to model the tributary as a storage
area, and connect the storage area to the main river with a lateral structure. The lateral structure
can be a weir, in which the weir geometry is represented with a cross section from the tributary.
44
Figure 17. Cross Section Layout for a Tributary Coming into a Main Stem River (Ackerman, 2014)
This will allow water from the flood wave to back up and fill the storage area as a level pool of
water. An example of modeling tributaries with storage areas and lateral structures is shown in
Figure 18.
The third option is to extend the normal cross sections up into the tributary, and use ineffective
flow areas for that portion of the cross sections. This option is depicted in Figure 19.
Modeling Levees and Major Roads. Downstream levees and major roads, that normally
prevent water from getting into protected areas, must also be considered. In general it is best to
model the area behind the levees separately as a two-dimensional flow area, a storage area, a
series of interconnected storage areas, or another routing reach. The details of modeling an area
behind a levee will depend on the terrain and details of the interior area. A lateral structure
(weir) should be used to model the top of the levees and major roads. When using a Lateral
structure to model a levee in HEC-RAS, this allows the model to evaluate levee overtopping,
breaching, and the filling of the interior area separate from the main river and floodplain. An
example of modeling a levee and protected area with a single storage area is shown in Figure 20.
If a levee or road is only a small obstruction to the flow, such that it will be completely
overwhelmed during the routing of the dam break flood wave, then it may be better to model that
45
Figure 18. Example of Using Storage Areas and Lateral Weirs to Account for Flow Reversals up
Tributaries (Ackerman, 2014)
Figure 19. Tributary Storage Modeled as Cross Section Ineffective Flow Areas (Ackerman, 2014)
46
Figure 20. Example of Using Lateral Structures and a Storage Area to Model a Protected Area
levee/road as part of the general cross sections. This means using cross sections to model both
the interior and exterior area around the levee, and using the HEC-RAS cross section levee
option to keep flow in the river side of the levee until the levee is overtopped. This should only
be done for small levees/roads, in which the area behind these levees is not a significant
area/storage volume. An example of this type of modeling is shown in Figure 21.
Modeling Bridge and Culvert Crossings

Bridges and culvert crossings can often be a source of model instability problems in a dam break
study. Many downstream bridges will be overtopped, and may even be washed away. If it is
almost certain that a downstream bridge/culvert will be washed away, then it probably does not
need to be included in the model. Additionally, if a structure is so high above the stream that the
water surface will not hit the low chord of the bridge deck (which may be the case for very large
highway bridges that are far downstream from the dam), then that bridge will also not need to be
modeled. However, if the road embankment, and the bridge/culvert will cause a backwater (i.e.,
a significant rise in the water surface), then it should be included in order to obtain the correct
stages upstream of the structure, and the increased storage behind the structure. If the impact of
the structure is unknown, then in general it should be modeled. Then once the model is up and
running, the structure could be evaluated for both its impact on the water surface and whether or
not it is expected to remain in place due to the forces placed on it during the event.
47
High Ground
Figure 21. High Ground (Road or Levee) Represented as Part of the Cross Sections
Bridge/culvert crossings are a common source of model stability problems when performing a
dam break analysis. Many bridges will be overtopped during such an event. Many of those
bridges may in fact be washed out during such an event. Common problems at bridges/culverts
are the extreme rapid rise in stages when flow hits the low chord of the bridge deck or the top of
the culvert. Modelers need to check the computed family of rating curves closely and make sure
they are reasonable. One solution to this problem is to use smaller time steps, such that the rate
of rise in the water surface is smaller for a given time step. Modelers may also need to change
hydraulic coefficients to get curves that have more reasonable transitions.
Just as with cross sections, HEC-RAS pre-processes bridges/culverts into a family of rating
curves. Users must ensure that these curves go high enough to capture all possible water surface
elevations and flows. An additional source of instability can arise when the curves do not go
high enough, and the program extrapolates from the last two points in the curve. This
extrapolation can cause problems when it is not consistent with the cross section geometry
upstream and downstream of the structure. The extrapolation is basically assuming that the
changes in conveyance, area, and other hydraulic parameters are linear with respect to increased
stage. However, these hydraulic properties are very non-linear. Therefore the extrapolation can
cause the unsteady flow equations to have difficulty in solving the equations. An example bridge
crossing and set of preprocessed curves is shown in Figure 22.
Modeling Steep Streams

Steep streams are very difficult to model with an unsteady flow model in general. Modeling a
dam break flood wave through a steep stream system is even more difficult. Steep streams tend
to have very high velocities and rapid changes in depth, area, and velocity, which makes it more
challenging to obtain a stable model solution through these areas.
48
Figure 22. Example Bridge with Pre-Processed Bridge Curves
The default solution methodology for the one-dimensional unsteady flow routing option within
HEC-RAS is generally for gradually varied flow. Areas of rapidly varied flow, such as flow
profiles transitioning from subcritical to supercritical flow, and hydraulic jumps, tend to cause
the one-dimensional solution scheme to have difficulties in remaining stable. Additionally, the
assumption of a hydrostatic flow distribution may not be valid. As Froude number approaches
1.0 (critical depth), the inertial terms of the St. Venant equations and their derivatives tend to
cause model instabilities (generally in rapid flow areas the derivatives are over estimated).
However, the HEC-RAS software does have an option to run the one-dimensional solution
scheme in a mixed flow regime mode, which allows it to solve through these types of flow
transitions.
Manning's n Values. If you are running the software in the default mode ( mixed flow option
not turned on), and if the program goes down to critical depth at a cross section, the changes in
area, depth, and velocity are very high. This sharp increase in the water surface slope will often
cause the program to overestimate the depth at the next cross section upstream, and possible
underestimate the depth at the next cross section downstream (or even the one that went to
critical depth the previous time step). One solution to this problem is to increase the Manning's n
value in the area where the program is first going to critical depth, and the steeper portions of the
reach. This will force the solution to a subcritical answer and allow it to continue with the run.
It is common for people to underestimate the magnitude of the Manning's roughness coefficient
for steep streams. Additionally, it is common to have pool and riffle sequences in steep streams.
In a pool and riffle sequence, Manning's n values will often be higher in the steeper riffle areas,
and lower in the flatter pool areas. This level of detail for modifying Manning's n values is often
not done, and can be a contributor to the instability of the model.
Mixed Flow Regime Option. If you feel that the true water surface should go to critical depth,
or even to an extended supercritical flow regime, then the mixed flow regime option should be
turned on when using one-dimensional river reaches to model steep areas. In order to solve the
49
stability problem for a mixed flow regime system, Dr. Danny Fread (Fread, 1986) developed a
methodology called the "Local Partial Inertia Technique" (LPI). The LPI method has been
adapted to HEC-RAS as an option for solving mixed flow regime problems when using the
unsteady flow analysis portion of HEC-RAS. This methodology applies a reduction factor to the
two inertia terms in the momentum equation as the Froude number goes towards a user defined
threshold.
The default values for the methodology are Froude Number Threshold (FT) = 0.8 and m
(exponent) = 4. When the Froude number is greater than the threshold value, the factor is set to
zero. The user can change both the Froude number threshold and the exponent. As you increase
the value of both the threshold and the exponent, you decrease stability but increase accuracy.
As you decrease the value of the threshold and/or the exponent, you increase stability but
decrease accuracy. To learn more about the Mixed Flow Regime option in HEC-RAS, please
see the HEC-RAS User's Manual (HEC, 2014).
Increased Baseflow. Another solution to the problem of flow going from subcritical to
supercritical flow and back again, is to increase the base flow in the hydrographs, as well as the
base flows used for computing the initial conditions. Increased base flow will often dampen out
any water surfaces going towards or through critical depth due to low flows that are in a pool
riffle sequence.
Modified Puls Routing. HEC-RAS has an option that will allow the user to define any portion
of a model to be solved with the Modified Puls routing method instead of the full unsteady flow
equations. This allows the user to define problem areas, such as very steep reaches, as Modified
Puls Routing reaches. A Modified Puls Routing reach can be defined at the upstream end of a
HEC-RAS river reach, at the downstream end, in the middle of a reach, or even defined for the
entire reach. The computations are performed in conjunction with the unsteady flow equations
on a time step by time step basis. Additionally, reaches that are defined as Modified Puls
reaches can contain bridges, culverts, and even lateral structures. The hydraulics of these
structure types are accounted for during the Modified Puls routing. To use this option, please
review the HEC-RAS User's Manual (HEC, 2014).
Two-Dimensional Flow Areas. The new two-dimensional flow area option in HEC-RAS
allows user to model areas with either the Full Saint Venant equations in two-dimensions, or the
diffusion wave form of the equations in two-dimensions. The new two-dimensional solver uses
a finite volume solution algorithm, which can handle subcritical, supercritical, and mixed flow
regime (including hydraulic jumps), much more robustly then the current one-dimensional finite
difference solution scheme. This makes it very easy to use two-dimensional flow areas to model
steep streams.
Drops in the Bed Profile

Significant drops in the bed profile can also be a source of model stability problems, especially at
low flows. Significant drops in the elevation of the channel bed can cause flow to pass through
critical depth and results in an unstable model solution. An example of this type of problem is
shown in Figure 23.
50
Figure 23. Model Instability due to a Drop in the Bed Profile.
If the drop is very small, then usually an increase in baseflow will drown out the drop, thus
preventing the model from passing through critical depth. If the drop is significant, then it
should be modeled with an inline structure using a weir profile at the top of the drop. This will
allow the model to use a weir equation for calculating the upstream water surface for a given
flow, rather than using the unsteady flow equations. This produces a much more stable model,
as the program does not have to model the flow passing through critical depth with the unsteady
flow equations. HEC-RAS automatically handles submergence on the weir, so this is not a
problem. An additional solution to this problem is to use the cross section rating curve option at
the top of the drop, which causes the program to interpolate the water surface from the rating
curve, rather than solving the unsteady flow equations through the drop in the bed profile.
Initial Conditions and Low Flow

Initial Conditions. In order for the unsteady flow model to run, the user must establish the
initial conditions in the entire system. This means that it must have a flow and a stage at every
cross section, as well as a stage in every storage area/two-dimensional flow area (storage areas
and two-dimensional flow areas can start dry). The most common way to establish the initial
conditions is for the user to enter a set of initial flows for all the reaches, and the software
performs a steady flow backwater profile to get the corresponding stages. The initial condition
flows entered by the user must be consistent with the all of the boundary condition flows at time
zero (the start of the unsteady flow run).
Initial reservoir elevations and gate settings must also be consistent with the initial condition
flows, such that the flow computed out of the reservoir at the first time step is consistent with
51
what the user entered to perform the initial conditions profile (Figure 24). If the user enters a
low flow for the initial conditions backwater profile, and then at the first unsteady flow time step
the program calculates a much larger flow coming out of the reservoir (due to gate settings an
initial reservoir stages), this can cause an instability in the area just below the dam.
Bald Eagle Creek Example Dam Break Study
Plan: PMF - Froehlich Breach Parameters
10/28/2008
Bald Eagle Cr. Lock Haven

Legend
680
EG 03JAN1999 0300
WS 03JAN1999 0300
Crit 03JAN1999 0300
Ground
660
Elevation (ft)
640
620
600
580
70000
75000
80000
85000
Main Channel Distance (ft)
Figure 24. Example of Initial Conditions for a Reservoir and Lateral Structures connected to Storage
Areas.
Another possible source of initial conditions causing the model to go unstable right away, are the
initial storage area elevations. It is up to the user to enter an initial storage area water surface
elevation for all storage areas; even if it is to start out dry (water surface is set to the lowest
elevation of the storage area). When a storage area is hydraulically connected to a river reach
(this is normally done with a lateral structure), and the initial water surface in the river reach is at
an elevation that will cause a flow interaction with a storage area (water surface is above the
lateral structure weir profile, or culverts, or gates), then that storage area needs to have an initial
water surface elevation set equal to the computed initial stage in the river. If the storage area is
set much higher or lower than the elevation of the river section it is connected to, then a large
discharge may be computed at the hydraulic structure that connects them. This large discharge
across the lateral structure will either take a lot of flow from the river (if the river stage is higher
than the storage area), or it will have a large inflow into the river (if the storage area stage is
much higher than the connected river stage). Either of these two cases can cause the model to go
unstable at the initial start of the unsteady flow computations. By setting the storage area
elevations to the same as the initial water surface of the cross section it is connected to, then the
computed flow across the lateral structure will be close to zero. Shown in Figure 24 are two
lateral structures, which are connected to storage areas. The initial condition water surface
elevation is higher than the downstream lateral structure. Therefore, the storage area connected
to this structure must be set to the initial condition water surface elevation in this area. Because
the initial water surface is lower than the most upstream lateral structure, the water surface
52
elevation for that connected storage area can be set to dry, or whatever elevation is appropriate
below the minimum elevation of the lateral structure.
Low Flow Conditions. Low flows can often be very difficult to model with an unsteady flow
model. Medium to steeper slope streams will often have a pool and riffle sequence at low flow,
and the water surface will generally pass through critical depth at the upper end of the riffle
(bottom of the pool). In addition to this, the depths of water are very shallow. Once the flood
wave begins the water surface will change quickly, and there will be a large change in depth with
respect to distance and time. The leading edge of a dam break flood wave will be very steep, and
can often be a source of model instability as it propagates down the river system. The finite
difference solution to the equations will generally have the most trouble balancing during the
initial dramatic rise at the beginning of the flood wave. The fact that the initial conditions may
be very low flows and depths can make it even more difficult to solve through those shallow and
steep riffle regions.
There are several things the modeler can do to allow the program to solve through this situation.
The easiest solution is to increase the base flow for the initial conditions. This will provide more
initial depth of water in general, and it may also drown out the pool and riffle sequence. A
general "rule of thumb" is to start out by trying a base flow around one percent of the peak flow
that will be routed. Increase the base flow if necessary, but never go above ten percent of the
peak flow. If you artificially use a base flow that is ten percent or more of the peak, the
computed peak flow and stage will be higher than it would have been otherwise.
If you have increased the base flow to a reasonable level, and are still having model stability
problems at the leading edge of the flood wave, then try adding a pilot for the reach in which the
model is having stability issues. A pilot channel is an option in which you can add some depth
without adding much flow area or conveyance. The pilot channel is an option in HEC-RAS, and
it is only used during low flow, once the cross sections get to some appreciable depth, the
program automatically removes it from the cross section. To learn more about the use of pilot
channels, please review the section on Pilot Channels in Chapter 6 of the HEC-RAS Users
Manual (HEC, 2014).
One other option that can help stabilize the model during the initial rise of the flood wave is
turning on the Mixed Flow Regime Option. This option drops the acceleration terms when the
Froude number gets greater than a user defined threshold, which is often the case on the leading
edge of the flood wave.
Downstream Boundary Conditions

Downstream boundary conditions are important for all hydraulic models, especially unsteady
flow models. Downstream boundary conditions can often be a source of model error, as well as
model instability. More often than not, the true stage for a given flow at the downstream end of
our models is not known. Because of this we often use either normal depth (Manning's
equation), or a rating curve computed from a steady flow model. The normal depth boundary
condition requires the user to enter a single energy slope, which is then in turn used in Manning's
equation to compute the downstream stage for any flow occurring. Occasionally this forced
slope or even a single valued rating curve can end up with stages that are not correct for the
given flow at a given point on the flood hydrograph. In general, the best solution is to make sure
53
that the downstream boundary condition is downstream from any of the locations in which stages
are being used to compute damages or loss of life, such that the error in the water surface
elevation at the boundary condition does not affect the area of interest.
Additionally, if a boundary condition is ill posed (rating curves with not enough points, or the
user entered stages are too low for a given flow rate; and normal depth boundaries where the user
has entered to steep of a slope for the energy gradeline), this can be a source of model instability.
In other words, the downstream boundary condition may be causing abrupt drops or rises in the
computed water surface near the location of the boundary condition. An example of what can
happen when using a normal depth boundary condition, and entering to steep of an energy slope
is shown in Figure 25. In this case, the steep energy slope caused the program to compute lower
stages than appropriate for a given flow, which in turn caused the model to over steepen the
flood wave at the downstream end of the model.
Figure 25. Example Model Error Due to Bad Downstream Boundary Condition
Using Two-Dimensional Flow Areas for Dam Break Analyses

The latest version of HEC-RAS (5.0 or later) now has the ability to perform two-dimensional
flow routing. For a dam break study, the user can model the downstream area entirely with onedimensional elements (cross sections and storage areas); as a combination of one- and twodimensional elements (cross sections, storage areas, and two-dimensional flow areas); or the
entire downstream area can be modeled as a two-dimensional flow area.
Two-dimensional flow areas can be directly connected to storage areas by using a hydraulic
structure called a storage area or two-dimensional flow area hydraulic connector ("SA/2D Area
Conn"). See the example in Figure 26.
In the example shown in Figure 26, the storage area is upstream of the two-dimensional flow
area, so the positive flow direction is from the storage area to the two-dimensional flow area.
When defining the hydraulic structure that connects the two areas, the storage area will be
54
Figure 26. Example of a Storage Area connected to a Two-Dimensional Flow Area
considered the headwater side, and the two-dimensional flow area will be considered the
tailwater side. In the example shown in Figure 25, a storage area is being used to represent a
reservoir pool. The hydraulic connection between the storage area and the two-dimensional flow
area is used to model the dam. The two-dimensional flow area is being used to model the
hydraulics of the flow downstream of the dam. Additionally, the user could model the reservoir
pool with a one-dimensional river reach, or a two-dimensional flow area.
Using the approach shown in Figure 26 is a very quick way to get a dam break model up and
running. However, modeling the downstream area with a two-dimensional flow area does not
necessarily make this a detailed model. Downstream areas will often have bridges, culverts,
roads that are barriers to flow, levees protecting urban areas, etc. The types of areas require
detailed modeling to get accurate answers, whether you are modeling them as two-dimensional
flow areas or one-dimensional river reaches. Developing a detailed model for the downstream
55
area requires detailed terrain, hydraulic structure information, and the time to model those areas
correctly. If a two-dimensional flow area is used, it still requires lots of work to make the
computational mesh respect all of the barriers to flow (bridges, culverts, roads, levees, etc.).
Developing a detailed computational mesh that respects all of the flow barriers, and includes all
of the hydraulic structures is the most time consuming part of developing a model, but it is
necessary to get good answers downstream. If you do not take the time to do this, and you just
through in a two-dimensional flow area with a nominal grid size, do not assume you have
"accurate" answers just because you a doing two-dimensional modeling.
56
References
Ackerman, 2014.
Ackerman, Cameron T., 2014. Adapted from lecture material provided during HEC-RAS
training courses.
Atallah, 2002.
Atallah, Tony A., 2002. A Review on Dams and Breach Parameters Estimation. Master of
Science in Hydrosystems Engineering, Virginia Polytechnic Institute & State University,
Blacksburg, VA, January 2002.
Brown and Rogers, 1977.
Brown, R. J. and Rogers, D. C., 1977. A simulation of the hydraulic events during and following
the Teton Dam failure. Proceedings of Dam-Break Flood Routing Model Workshop, Bethesda,
MD, pages 131 - 163; 18 20 October,.
Costa, 1985.
Costa, John E., 1985. Floods from Dam Failures. United States Department of the Interior,
Geological Survey, Open-File Report 85-560, Denver, CO.
Cristofano, 1965.
Cristofano, Eugene A., 1965. Method of Computing Erosion rate for Failure of Earthfill Dams.
U.S. Bureau of Reclamation, Engineering and Research Center, Denver CO; 1965.
Dewey and Gillette, 1993.
Dewey, R. and Gillette, D., 1993. Prediction of Embankment Dam Breaching for Hazard
Assessment. ASCE Specialty Conference on Geotechnical Practice in Dam Rehabilitation,
Raleigh, North Carolina, 25-28 April 1993.
Evans, 1986.
Evans, Stephen. G., 1986. The maximum discharge of outburst floods caused by the breaching
of man-made and natural dams. Canadian Geotechnical Journal, 1986, 23(3), pages 385 - 387.
FERC, 1988.
United States Federal Energy Regulatory Commission (FERC) - Notice of Revised Emergency
Action Plan Guidelines, 22 February 1988.
Fread, 1986.
Fread, Dr. Danny L., 1986. Local Partial Inertia Technique (LPI).
Fread, 1988.
Fread, Dr. Danny L., 1988 (revised 1991). BREACH: An Erosion Model for Earth Dam
Failures. Hydrologic Research Laboratory, National Weather Service, National Oceanic and
Atmospheric Administration (NOAA), Silver Springs, MD, July 1988 (revised August 1991).
57
Fread, 1993.
Fread, Dr. Danny L., 1993. NWS FLDWAV model: The replacement of DAMBRK for dam-break
flood prediction. 10th Annual Conference of the Association of State Dam Safety Officials,
Kansas City, MO, 26 - 29 September 1993, pages 177184.
Fread, 2006.
Fread, Dr. Danny L., 2006. Course notes on "Dam Break Floods".
Froehlich, 1987.
Froehlich, David C., 1987. Embankment-Dam Breach Parameters. Hydraulic Engineering,
Proceedings of the 1987 National Conference, ASCE, Williams burg, VA, pages 570 - 575.
Froehlich, 1995a.
Froehlich, David C., 1995. Embankment Dam Breach Parameters Revisited. First International
Conference, Water Resources Engineering, Environmental and Water Resources Institute
(EWRI), 14 - 18 August 1995. ASCE, Water Resources Engineering Proceeding, pages 887 891.
Froehlich, 1995b.
Froehlich, David C., 1995. "Peak Outflow from Breached Embankment Dam". ASCE Journal of
Water Resources Planning and Management, Volume 121, Issue 1, January 1995, pages 90 - 97.
Froehlich, 2008.
Froehlich, David C., 2008. Embankment Dam Breach Parameters and Their Uncertainties.
ASCE, Journal of Hydraulic Engineering, Vol. 134, No. 12, December 2008, pages 1708-1721.
ISSN 0733-9429.
Hagen, 1982.
Hagen, V. K., 1982. Re-evaluation of design floods and dam safety. Proceedings of 14th
ICOLD Congress, Rio De Janeiro, Brazil. Vol. 1, pages 475 - 491, 1982.
Hanson, et al., 2003.
Hanson, Gregory J., Cook, K.R., and Britton, S.L., 2003. "Observed Erosion Processes During
Embankment Overtopping Tests. American Society of Agricultural Engineers (ASAE), Annual
Meeting, Proceedings, Paper No. 03-2066, July 2003.
Harris and Wagner, 1967.
Harris ,G.W. and Wagner, D.A., 1967. Outflow from Breached Earth Dams. University of
Utah, Salt Lake City, UT.
Hassan, et al., 2004.
Hassan, M., Morris, M., Hanson, G.J., and Lakhal, K., 2004. Breach formation: Laboratory and
numerical modeling of breach formation. Proceedings of the Annual Conference of the
Association of State Dam Safety Officials (ASDSO), Phoenix, AZ, 2004, pages 1321-1337.
58
HEC, 1980.
Hydrologic Engineering Center, 1980. Flood Emergency Plans - Guidelines for Corps Dams,
RD-13, June 1980. U.S. Army Corps of Engineers, Hydrologic Engineering Center, 609 Second
Street, Davis, CA.
HEC, 2014.
Hydrologic Engineering Center, 2014. HEC-RAS River Analysis System, User's Manual, CPD68. U.S. Army Corps of Engineers, Hydrologic Engineering Center, 609 Second Street, Davis,
CA.
HEC, 2014a.
Hydrologic Engineering Center, 2014. HEC-RAS River Analysis System, Hydraulic Reference
Manaual, CPD-69. U.S. Army Corps of Engineers, Hydrologic Engineering Center, 609 Second
Street, Davis, CA.
Jarrett, 1984.
Jarrett, Dr. Robert D, 1984. Hydraulics of High-Gradient Streams. ASCE, Journal of
Hydraulics, Volume 110, No. HY 11, pages 1519 - 1539; November 1984.
Johnson and Illes, 1976.
Johnson, F.A. and Illes, P., 1976. A Classification of Dam Failures. Water Power and Dam
Construction, pages 43 - 45; December 1976
Kirkpatrick, 1977.
Kirkpatrick, Gerald. W., 1977. Evaluation guidelines for spillway adequacy, The Evaluation of
Dam Safety. Engineering Foundation Conference, ASCE, New York, pages 395 - 414, 1977.
MacDonald and Langridge-Monopolis, 1984.
MacDonald, Thomas C., and Langridge-Monopolis, J, 1984. Breaching characteristics of dam
failures. ASCE Journal of Hydraulic Engineering, Vol. 110, No. 5, May 1984, pages 567 - 586.
Mohamed, 2002.
Mohamed, M.A.A., 2002. Embankment Breach Formation and Modeling Methods. The Open
University, England, UK.
Olsen, 1976.
Olsen, Eunice, 1976. Photos of the Teton Dam Failure, St. Anthony, ID, June 1976.
Ponce and Tsivoglou, 1981.
Ponce, Victor M., and Tsivolgou , Andrew J., 1981. Modeling Gradual Dam Breaches. ASCE
Journal of the Hydraulics, Proceedings, Volume107, No. 7, pages 829 - 838.
Samuels, 1989.
Samuels, P.G., 1989. Backwater lengths in rivers. Proceedings of the Institution of Civil
Engineers, Part 2, Volume 87, pages 571 - 582.
59
Singh and Scarlatos, 1985.

Singh, V. P. and Scarlatos, P. D., 1985b. Breach erosion of earth-fill dams and flood routing:
BEED Model. Military Hydrology Report, page 131; U.S. Army Engineer Waterways
Experiment Station, Environmental Laboratory, Vicksburg, MS.
Singh and Scarlatos, 1988.
Singh, V.P., Scarlatos, P.D., Collins, J.G. and Jourdan, M.R., 1988. Breach Erosion of Earthfill
Dams (BEED) Model. Natural Hazards, Volume 1, pages 161 - 180.
Singh and Snorrason, 1982.
Singh, Krishan P. and Snorrason A., 1982. Sensitivity of Outflow Peaks and Flood Stages to the
Selection of Dam Breach Parameters and Simulation Models. University of Illinois, State Water
Survey Division, Surface Water Section, Champaign, IL, 179 pages; June1982.
Singh and Snorrason, 1984.
Singh, Krishan P., and Snorrason, A., 1984. Sensitivity of outflow peaks and flood stages to the
selection of dam breach parameters and simulation models. Journal of Hydrology, Vol 68,
Issues 1-4, February 1984, pages 295 - 310. Global Water: Science and Engineering The Ven Te
Chow Memorial Volume
SCS, 1981.
SCS, 1981. Simplified Dam-Breach Routing Procedure. United States Department of
Agriculture, Soil Conservation Service (SCS), Technical Release No. 66 (Rev. 1), 39.
Temple, et al., 2006.
Temple, Darrel M., G.J. Hanson, and Neilsen, M.L., 2006. WINDAM Analysis of Overtopping
Earth Embankment Dams. ASABE Annual International Meeting Proceedings, Paper No. 062105; 9 - 12 July 2006, Portland, OR.
USBR, 1982.
USBR, 1982. Guidelines for defining inundated areas downstream from Bureau of Reclamation
dams. Reclamation Planning Instruction, No. 82-11; 15 June 1982, U.S. Bureau of Reclamation.
Vaskinn, et al., 2004.
Vaskinn, Kjetil A., Lovoll, M.A., Hoeg, K., Morris, M., Hanson, G., and Hassan, M., 2004.
Physical Modeling of Breach Formation; Large Scale Field Tests. State Dam Safety Officials
Association Proceedings. Phoenix, AZ, September 2004.
Von Thun and Gillette, 1990.
Von Thun, J. Lawrence, and Gillette, D. R., 1990. Guidance on breach parameters.
Unpublished internal document, U.S. Bureau of Reclamation, Denver, CO. March 1990, 17
pages.
Wahl, 1998.
Wahl, Tony L., 1998. Prediction of Embankment Dam Breach Parameters - A Literature Review
and Needs Assessment. Dam Safety Research Report, DSO-98-004. Water Resources Research
Laboratory, U.S. Dept. of the Interior, Bureau of Reclamation, Dam Safety Office (DSO), July
1998.
60
Wahl, 2009.
Wahl, Tony L., 2009. "Evaluation of New Models for Simulating Embankment Dam Breach.
Association of State Dam Safety Officials (ASDO) Conference, Hollywood, Florida, 27
September - 1 October 2009.
Walder and O'Connor, 1997.
Walder, Joseph S. and O'Connor, J. E., 1997. Methods for predicting peak discharge of floods
caused by failure of natural and constructed earth dams. Water Resources Research, Vol. 33
Issue 10, pages 2337 - 2348. October 1997.
Wang and Kahawita, 2006.
Wang, P., Kahawita, R., Mokhtari, A., Phat, T.M. and Quach, T.T., 2006. Modelling breach
formation in embankments due to overtopping. ICOLD Conference, Barcelona, Spain, June
2006.
Xu and Zhang, 2009.
Xu, Y. and Zhang , L. M., 2009. Breaching Parameters for Earth and Rockfill Dams. ASCE
Journal of Geotechnical and Geoenvironmental Engineering, Volume 135, No. 12, pages 1957 1970, December 2009.
Additional References
CEATI (Centre for Energy Advancement through Technological Innovation) International Inc.,
Dam Safety Interest Group (DSIG), http://www.ceati.com/.
Fread, Dr. Danny L., 1984. DAMBRK: The NWS dam-break flood forecasting model. U.S.
Army Corps of Engineers, Hydrologic Engineering Center, 609 Second Street, Davis, CA. CPD40
Hanson, G. J., Temple, D. M., Morris, M., Hassan, M., Cook, K., 2005). Simplified breach
analysis model for homogeneous embankments: Part II, Parameter inputs and variable scale
model comparisons U.S. Society on Dams Annual Meeting and Conference, Proceedings, Salt
Lake City, UT, June 2005, pages 163 - 174.
Hydrologic Engineering Center, 2010a. HEC-RAS River Analysis System, Hydraulic Reference
Manual, CPD-69. U.S. Army Corps of Engineers, Hydrologic Engineering Center, 609 Second
Street, Davis, CA.
Hydrologic Engineering Center, 2011. HEC-GeoRAS, User's Manual, CPD-83 U.S. Army
Corps of Engineers, Hydrologic Engineering Center, 609 Second Street, Davis, CA.
Hydrologic Engineering Center, 2010. HEC-HMS Hydrologic Modeling System, User's Manual,
CPD-74A. U.S. Army Corps of Engineers, Hydrologic Engineering Center, 609 Second Street,
Davis, CA.
Wahl, Tony L., 2004. Uncertainty of Predictions of Embankment Dam Breach Parameters",
ASCE, Journal of Hydraulic Engineering, Volume 130, No. 5, pages 389 - 397, May 2004.
61
Wahl, Tony L., Hanson, G.J., Courivaud J., Morris, M., Kahawita, R., McClenathan, J.T., and
Gee, D.M., 2008. Development of Next-Generation Embankment Dam Breach Models. United
States Society on Dams Proceedings, pages 767 - 779; Portland, OR; 28 April - 2 May 2008.
Washington State Dept. of Ecology, 1992. Dam Safety Guidelines, Technical Note 1, Dam
Break Inundation Analysis and Downstream Hazard Classification. Water Resources Program,
Dam Safety Office, P.O. Box 47600, Olympia, WA 98504-7600, July 1992 (revised October
2007), https://fortress.wa.gov/ecy/publications/publications/9255e.pdf
62

TD 39

Uploaded by

Copyright:

Available Formats

TD 39

Uploaded by

Document Information

Original Description:

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

TD 39

Uploaded by

Copyright:

Available Formats

Using HEC-RAS for Dam

Approved for Public Release. Distribution Unlimited.

Form Approved OMB No. 0704-0188

REPORT DOCUMENTATION PAGE

4. TITLE AND SUBTITLE

5a. CONTRACT NUMBER

Using HEC-RAS for Dam Break Studies

5b. GRANT NUMBER

5d. PROJECT NUMBER

8. PERFORMING ORGANIZATION REPORT NUMBER

U.S. Army Corps of Engineers

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)

10. SPONSOR/ MONITOR'S ACRONYM(S)

12. DISTRIBUTION / AVAILABILITY STATEMENT

Approved for Public Release. Distribution Unlimited.

15. SUBJECT TERMS

19a. NAME OF RESPONSIBLE PERSON

19b. TELEPHONE NUMBER

Using HEC-RAS for Dam

U.S. Army Corps of Engineers

Agricultural Research Service

Federal Energy Regulatory Commission

Froude Number Threshold

Hydrologic Engineering Center

Hydrologic Modeling System software

River Analysis System software

Local Partial Inertia Technique

National Oceanic and Atmospheric Administration

National Weather Service

probable maximum flood

Soil Conservation Service

U.S. Army Corps of Engineers

U.S. Bureau of Reclamation

Training Document No. 39

Using HEC-RAS for Dam Break Studies

Using HEC-RAS for Dam Break Studies

Routing the Inflow Flood through a Reservoir

Using HEC-RAS for Dam Break Studies

Training Document No. 39

Training Document No. 39

Using HEC-RAS for Dam Break Studies

Full Dynamic Wave Routing

Using HEC-RAS for Dam Break Studies

Training Document No. 39

Level Pool Routing

Training Document No. 39

Using HEC-RAS for Dam Break Studies

Using HEC-RAS for Dam Break Studies

Training Document No. 39

Estimating the Dam Breach Characteristics

Training Document No. 39

Using HEC-RAS for Dam Break Studies

Causes and Types of Dam Failures

Using HEC-RAS for Dam Break Studies

Training Document No. 39

Estimating Breach Parameters

Training Document No. 39

Using HEC-RAS for Dam Break Studies