Beruflich Dokumente
Kultur Dokumente
Prepared for:
Federal Emergency Management Agency
June 2011
Prepared by:
FLO-2D Software, Inc.
Nutrioso, Arizona
Endorsements
The following agencies have reviewed this document, endorse the findings presented herein and
support the recommendation that the FLO-2D flood routing model (Versions 2009 and up) to be
included on FEMA's list of approved hydraulic and hydrologic models for Flood Insurance Studies.
Flood Control District of Maricopa County, Phoenix, Arizona (FCDMC)
U.S. Army Corps of Engineers, Sacramento District (Corps)
California Department of Water Resources, Sacramento, California (DWR)
Truckee River Flood Project, Reno, Nevada (TRFP)
U.S. International Boundary and Water Commission, El Paso, Texas (IBWC)
City of Camarillo, California
Mohave County Flood Control District, Kingman, Arizona (MCFCD)
Cochise County Flood Control District, Bisbee, Arizona (CCFCD)
Some of these agencies have a very long history and current projects using the FLO-2D model (Corps,
DWR, FCDMC, IBWC, TRFP). The other three agencies have FIS current projects using Version 2009.
Some of the agencies provided formal letters of endorsement sent to FEMA (FCDMC, MCFCD, CCFCD),
other agencies sent a formal endorsement by email to FLO-2D Software, Inc. (Corps, DWR, FRFP). These
letters and email endorsement are presented in the Appendix. The City of Camarillo and IBWC indicated
their endorsement with a simple email message that was not included in the Appendix.
Executive Summary
Since 1999 and through Version 2007 FLO-2D has been listed as a FEMA approved hydraulics model. In
June 2010 FLO-2D Software, Inc. petitioned the Federal Emergency Management Agency (FEMA) to
include Version 2009 of the FLO-2D model on FEMA's list of approved hydraulics models. This petition
procedure to list a new model version was accomplished with previous versions according to a
Memorandum of Understanding agreement between FEMA and FLO-2D Software, Inc. In February
2011, FEMA informed FLO-2D Software, Inc. that supporting validation documentation was required for
listing both new models and to update model versions. This document presents information that the
FLO-2D model for Versions 2009 and subsequent releases accurately predicts flood hydrology and
hydraulics and distributes flood volume. Specifically, this document was prepared to support a FEMA
listing of the FLO-2D Model (Versions 2009 and up) as an approved model for both hydraulic and
hydrologic Flood Insurance Studies (FIS).
Following International Association for Hydro-Environment Engineering and Research (IAHR) guidelines
for documenting model validation, this document discusses the FLO-2D model representation of the
flood physical system, the validation approach, the computational core validation, range of model
applicability, and the accuracy of the model results. The model background information is presented in
Chapter I along with validation goals and objectives. An overview of the model is presented in Chapter II
including model formulation, algorithm solution, numerical stability criteria, model organization and
assumption. Model validation for hydrology and hydraulics is discussed in Chapter III focusing on the
validity of the computation core for overland flow, river channel flow and channel-floodplain exchange.
Chapter III is organized according to the IAHR format for documenting the validity of computational
modeling software which presents a validation claim followed by a section on the substantiation of the
claim. The substantiations for the claims are based on a series of case study projects that validate
specific model components. Results from each case study project are detailed to verify model accuracy
and consistency. Further details of hydraulic validation case studies are presented in Chapter IV. Model
certification letters are presented in the Appendix.
This document demonstrates that the FLO-2D model provides accurate hydrology and hydraulic results
for a broad range complex flood studies from river flooding to urbanized alluvial fans. This document
also asserts that the computation core has provided consistent hydraulic results throughout the last
eleven years of new releases. Benchmark testing ensures that the model engine will provide future
accurate and reliable results.
ii
Table of Contents
Endorsements .............................................................................................................................................................i
Executive Summary ................................................................................................................................................. ii
FLO-2D Model Validation .......................................................................................................................................1
I.
Introduction .......................................................................................................................................................1
1.1
1.2
1.3
1.4
II.
Background .................................................................................................................................... 1
Some Comments on Modeling Free Surface Flows ....................................................................... 2
Validation Goal and Objectives ...................................................................................................... 2
Model Speed, Accuracy, Stability and Sensitivity .......................................................................... 3
iii
4.1
4.2
4.3
4.4
4.5
4.6
V. References ....................................................................................................................................................... 78
Appendix
U.S. Army Corps of Engineers Original Two Letters of Certification
Letters of Endorsement
Email Endorsements
List of FLO-2D Model Revisions
List of Figures
Figure 1. Unconfined Discharge Computations .......................................................................................................6
Figure 2. FLO-2D Numerical Stability Flow Chart .................................................................................................9
Figure 3a. SUMMARY.OUT File Report. ............................................................................................................ 18
Figure 3b. SUMMARY.OUT File for the Camarillo Project ................................................................................ 18
Figure 4. FLO-2D Version 2009 Results for Maximum Velocity and Depth ........................................................ 21
Figure 5. Version 2007 Flume Case Results Showing Identical Results as Version 2009 in Figure 4. ................. 21
Figure 6. Truckee River Observed (Black Outline) and FLO-2D Predicted Area of Flood Inundation ............... 22
Figure 7. Predicted Truckee River Water Surface Profiles for HEC-2 and FLO-2D............................................. 23
Figure 8. FLO-2D Predicted Stage for the 1997 Flood Event at the Reno Gage .................................................. 24
Figure 9. FLO-2D Predicted Stage for the 1997 Flood at the New Vista Gage .................................................... 25
Figure 10. FLO-2D Version 2009 Predicted Stage for the 1997 Flood at the New Vista Gage ............................ 26
Figure 11. FEMA 2005 FIRM MAP for Ventura County ..................................................................................... 27
Figure 12. FLO-2D Predicted Maximum Flow Depths for the City of Camarillo ................................................. 28
Figure 13. FLO-2D 2011 Model FIRM for the City of Camarillo ......................................................................... 28
Figure 14. USGS San Felipe Gage 1998 (FLO-2D Predicted vs. Measured Discharge) ....................................... 30
Figure 15. FLO-2D Predicted vs. USGS Gage Measured Discharge at Chama River Confluence ....................... 31
Figure 16. FLO-2D Predicted Area of Inundation Compared with the Corps' Digitized Shape File for the 2005
Area of Inundation Monitoring Project ................................................................................................................... 32
Figure 17. FLO-2D Predicted vs. USGS Gage Discharge at San Felipe May 24-June 1, 2005............................. 33
Figure 18. FLO-2D Predicted vs. USGS Gage Discharge at Albuquerque May 24-June 1, 2005 ......................... 33
Figure 19. FLO-2D Predicted vs. USGS Gage Discharge at Isleta Lakes May 24-June 1, 2005 .......................... 34
iv
Figure 20. FLO-2D Predicted vs. USGS Gage Discharge at San Acacia May 24-June 1, 2005............................ 34
Figure 21. 1997 High Flow Season Hydrograph at Jensen, Utah (FLO-2D vs. Measured Discharge) .................. 42
Figure 22. November 1998 Power Plant Discharge Release from Flaming Gorge Dam ....................................... 43
Figure 23. FLO-2D Predicted vs. Measured Discharge, Jensen USGS Gage for Power Plant Releases at Flaming
Gorge Dam ............................................................................................................................................................. 44
Figure 24. Version 2009 FLO-2D Predicted vs. Measured Discharge, Jensen USGS Gage for Power Plant
Releases at Flaming Gorge Dam............................................................................................................................. 45
Figure 25. Channel-Floodplain Flow Exchange ................................................................................................... 46
Figure 26. NEXRAD Cell Total Precipitation July 31, 2006 on the Catalina Mountain Watersheds .................... 49
Figure 27. Lower Soldier Canyon Alluvial Fan Aerial Photo. Agua Caliente Wash crosses ............................... 50
Soldier Trail (center photograph). Photograph by Terry Hendricks (PCRFCD; 7-31-06). .................................... 50
Figure 28. Soldier Canyon Alluvial Fan FLO-2D Predicted Area of Inundation for the July 31, 2006 Flood ...... 51
Figure 29. FLO-2D Predicted Maximum Depth for Rainbow Wash (FCDMC, 2010).......................................... 52
Figure 30. FLO-2D Predicted Hydrograph vs. Gage Data and HEC-2D Results (FCDMC, 2010) ....................... 52
Figure 31. FLO-2D Predicted Inundation Area vs. the Observed Flooding from 2005 Storm (FCDMC, 2010) ... 53
Figure 32. FLO-2D Model Predicted Depth and Velocities for White Tanks Fan 36 (JE Fuller, 2010) ............... 56
Figure 33. FLO-2D Mode Predicted Depth and Velocities for White Tanks Fan 36 ............................................ 57
Figure 34. Rio Grande Overbank Flooding Confined by Levees (red outline in Maxplot) ................................... 59
Figure 35. Simple Levee Example Project with 1,000 cfs Steady Discharge ........................................................ 60
Figure 36. Corps' Monroe Alluvial Fan Project with a Defined Channel Using Version 2007 .............................. 62
Figure 37. Corps' Monroe Alluvial Fan Project Difference in Maximum Depth .................................................... 62
(all the differences in maximum depth are 0.2 ft or less) ........................................................................................ 62
Figure 38. Sparks Truckee Meadows Area General Peak Flow Path (Google Earth, 2011) .................................. 68
Figure 39. City of Camarillo 10 ft FLO-2D Grid System with the Building Shape File ....................................... 69
Figure 40. Detail of the City of Camarillo 10 ft FLO-2D Grid System with LiDAR Data .................................... 69
Figure 41. Spatially Variable n-value Shapefile .................................................................................................... 69
Figure 42. Flooding Differences Between the Models (100 ft model left and 10 ft model right) .......................... 70
Figure 43. Flooding along Ponderosa Drive (100 ft model left and 10 ft model right) ......................................... 70
Figure 44. 100 ft Model - Flooding across Highway 101. ...................................................................................... 71
Figure 45. 10 ft Model - Flooding is confined by Highway 101. ........................................................................... 71
Figure 46. FLO-2D Predicted vs. USGS Gage Discharge at San Marcial April 3-11, 2004 ................................. 73
Figure 47. FLO-2D Predicted vs. USGS Gage Discharge at San Marcial May 5-22, 2004 .................................. 73
Figure 48. Aerial Image of the FLO-2D Reach of the California Aqueduct Near Huron ...................................... 74
Figure 49. California Aqueduct Near Huron (from Kewaneh, Google Earth, 2011) ............................................. 74
Figure 50. California Aqueduct Bedslope Near Huron .......................................................................................... 75
Figure 51. Green River Bed Slope from Flaming Gorge Dam to the Colorado River Confluence ........................ 76
Figure 24. FLO-2D Predicted vs. Measured Discharge, Jensen USGS Gage for the Flaming Gorge Releases ..... 76
Figure 52. Flaming Gorge Power Plant Spike Release Hydrograph at the Ouray Gage ........................................ 77
Figure 53. Flaming Gorge Power Plant Spike Release Hydrograph at the end of Uinta Valley ............................ 77
Figure 54. Flaming Gorge Power Plant Spike Release Hydrograph at the Colorado River Confluence ............... 77
List of Tables
Table 1. FLO-2D Models for the Middle Rio Grande ........................................................................................... 29
Table 2. Floodplain Area of Inundation Calibration .............................................................................................. 32
vi
solutions began in 1999 in support of litigation in Federal Court with the California Department of Water
Resources and federal agencies over the release of flood waters from the Aqueduct onto farmland.
Benchmark testing started with Version 2002 and became more extensive with Version 2004. This
testing program expedited the development and modification of model components.
The goal of model validation is to have consistent results with the development of new features that
may also improve model simulation accuracy and stability. The validation objectives are:
There are several ways to detect and control numerical instability that will be discussed in Chapter II.
h h V
+
=i
t
x
S f = So -
h V V V V 1 V
x g x g x g t
where h is the flow depth and V is the depth-averaged velocity in one of the eight flow directions x. The
excess rainfall intensity (i) may be nonzero on the flow surface. The friction slope term Sf is based on
Mannings equation. The other terms include the bed slope So, pressure gradient and convective and
local acceleration terms. If the pressure term and acceleration terms are negligible (Sf = So), the flow is
considered steady and uniform (referred to as the kinematic wave equation). The addition of the
pressure gradient term (h/ x) to the kinematic wave equation results in the diffusive wave equation,
which is required for floodwave attenuation and change in storage on the floodplain. The remaining
local and convective acceleration terms are important to the flood routing for flat or adverse slopes,
very steep slopes, or unsteady flow conditions. Including these additional terms results in the full
dynamic wave momentum equation. Only the full dynamic wave representation of the momentum
equation is used in the FLO-2D model.
For unconfined overland flow, the equations of motion in FLO-2D are applied by computing the average
flow velocity across a grid element boundary one direction at time. There are eight potential flow
directions, the four compass directions (north, east, south and west) and the four diagonal directions
(northeast, southeast, southwest and northwest). Each velocity computation is essentially onedimensional in nature and is solved independently of the other seven directions. The stability of this
explicit scheme is based on strict numerical criteria to control the magnitude of the variable
computational timestep. For a discussion of the supplemental equation necessary for
hyperconcentrated sediment flow refer to the FLO-2D Reference Manual.
The full dynamic wave equation is expressed as a second order, non-linear, partial differential form in
the FLO-2D model. To solve the equation, initially the flow velocity is calculated at the grid element
boundary with the diffusive wave equation using the energy slope (bed slope plus pressure head
gradient). This velocity is then used as a first estimate (or a seed) in the second order Newton-Raphson
tangent method to determine the roots of the full dynamic wave equation (James, et. al., 1985).
Mannings equation is used to compute the friction slope. If the Newton-Raphson solution fails to
converge after 3 iterations, the algorithm defaults to the diffusive wave solution.
In the full dynamic wave momentum equation, the local acceleration term is the difference in the
velocity over the previous timestep for a given flow direction. The convective acceleration term is
evaluated as the difference in the flow velocity across the grid element from the previous timestep. For
example, in the notation below, the local acceleration term (1/g*V/t) for grid element 251 in the east
(2) direction converts to:
(Vt Vt-1)251 (g * t)
where Vt is the velocity in the east direction for grid element 251 at time t, Vt-1 is the velocity at the
previous timestep (t-1) in the east direction, t is the timestep in seconds, and g is the acceleration due
to gravity. A similar construct for the convective acceleration term (Vx/g*V/x) can be made where V2
is the velocity in the east direction and V4 is the velocity in the west direction for grid element 251:
V2 * (V2 V4)251 (g * x)
Subsequent to Version 2009, the FLO-2D model also includes diffusion terms in the full dynamic wave
solution which helps to dampen the velocity oscillations for certain flow conditions:
Di (V2 - 2. * Vnew + V4)251 Length2
where Di = hardwired coefficient (~0.25), Vnew = seed velocity, and Length is either the channel length or
flow path within the grid element.
The discharge across the grid element boundary is computed by multiplying the velocity times the cross
sectional flow area. After the discharge is computed for all eight directions, the net change in discharge
(sum of the discharge in the eight flow directions) in or out of the grid element is multiplied by the
timestep to determine the net change in the grid element water volume (see Figure 1). This net change
in volume is then divided by the available surface area (Asurf = storage area) on the grid element to
obtain the increase or decrease in flow depth h for the timestep. The channel routing integration is
performed essentially the same way except that the flow depth for next timestep (i+1) is a function of
the channel cross section geometry and there is usually only one upstream and one downstream
channel grid element for sharing discharge.
i+1
Qx = Qn + Qe + Qs + Qw + Qne + Qse + Qsw + Qnw Asurf h/t
1. The average flow geometry, roughness and slope between two grid elements are computed.
2. For computing the velocity across a grid boundary for the next timestep (i+1), the flow depth dx is a
linear estimate (average depth between two elements) from the previous timestep i.
i+1
i
i
d x = d x + d x1
3. The first estimate of the velocity is computed using the diffusive wave equation. The only unknown
variable in the diffusive wave equation is the velocity for overland, channel or street flow.
4. The predicted diffusive wave velocity for the current timestep is used as a seed in the NewtonRaphson solution to solve the full dynamic wave equation for the solution velocity. The NewtonRaphson convergence criteria for the root solution is 0.01.
5. The discharge Q across the boundary is computed by multiplying the velocity by the cross sectional
flow area (flow width times depth for overland flow). The flow width is adjusted by the width
reduction factors (WRFs).
6. The incremental discharge for the timestep across the eight boundaries (or upstream and
downstream channel elements) are summed:
Decrease Timesteps,
Reset Hydraulics, Restart
Flood Routing
Initialize Variables
Start Flood
Routing Loop
Channel Subroutine,
Hydraulic Structures,
Mudflow, Sediment,
Transport, Infiltration
Yes
Channel/Street and
Floodplain Interface
Channel
Flow
Channel Stability
Criteria Satisfied
No
No
Sediment Distribution on
Channel/Floodplain Bed
Yes
Rainfall Runoff
and Evaporation
Yes
No
Output Interval
Complete
Overland Flow
Sediment Transport
Infiltration, Gully Flow,
Hydraulic Structures,
Mudflow
Yes
Simulation
Time Complete
Numerical Stability
Criteria Satisfied
No
Yes
Yes
Street Flow
No
End
Yes
Numerical Stability
Criteria
No
No
Yes
No
10
computations for that timestep are discarded and the timestep is decreased. The dynamic wave
stability criteria was determined to be useful for maintaining numerical stability in channel confluences
and other unique channel flow conditions.
It has been determined that the Courant Number is more effective in controlling numerical surging
parameter for most FLO-2D flood simulations than the combined application of the percent change in
depth DEPTOL and WAVEMAX stability. While the Courant condition is a necessary condition for
solution convergence, it is not always sufficient to guarantee numerical stability. The WAVEMAX
stability parameter may still have some value in controlling numerical instability for some unique flow
conditions such as channel confluence flow or split flows. With Version 2009 for most flood simulations,
however, the DEPTOL and WAVEMAX stability criteria are not necessary if the CFL criteria is properly
applied.
The key to efficient computational flood routing is to have the largest timesteps possible and still avoid
numerical instability. Timesteps for most FLO-2D simulations typically range from 0.1 to 30.0 seconds.
Each flood simulation starts with a timestep set to 1.0 second and increases it until the numerical
stability criteria is exceeded, then the timestep is decreased until the minimum timestep is reached. If
the minimum timestep is not small enough to maintain numerical stability, then the minimum timestep
can be reduced, the numerical stability criteria can be adjusted or the input data can be modified. The
timesteps are a function of the discharge flux for a given grid element and its size. Small grid elements
with a steep rising hydrograph and large peak discharge require small timesteps. Accuracy is not
compromised for small timesteps, but the computational time can be long for a large grid system.
2.4 Courant Number Guidelines
The Courant Number is assigned by the user in the TOLER.DAT file line 2 as follows:
Line 1.
0.1
0.2
Line 2.
0.6
1.0
Line 2 is optional and if left unassigned, the default Courant Number is 0.6. A typical range of the
Courant Number is 0.3 (for a slower more stable model) to 1.0 (for a faster less stable model). The
default value of 0.6 is recommended as a starting value. For models that appear to be unstable,
reducing the Courant Number will help to control or eliminate the numerical surging.
Using the VELTIMEFP.OUT and VELTIMEC.OUT files that list the grid element or channel element
maximum velocity in descending order, unreasonable velocities can be identified. Based on the model
results, the following guidelines are recommended:
1. For the initial simulation:
Courant Number
DEPTOL
WAVEMAX
= 0.6
= 0.0
= 0.0
2. If the model has no numerical surging or unreasonable maximum velocities and it is desired to
have the model run faster increase the Courant Number to 0.7 or 0.8.
11
3. If the model has some numerical instability, decrease the Courant Number to 0.3 - 0.5.
4. If the model has numerical surging in the channel, set the WAVEMAX to 0.25 (this will reduce
the timestep and slow the model down).
5. After a flood simulation is complete, review the TIME.OUT file to determine which of the
stability criterion is slowing down the model.
6. It may be necessary to experiment with short duration simulations to determine which
combination of the stability criteria results in the fastest stable model.
2.5 The Importance of Volume Conservation
A review of a flood model results begins with volume conservation. Volume conservation is an
indication of model accuracy and numerical stability. Many numerical models either do not conserve
volume or do not report on volume conservation. In the FLO-2D model, the inflow volume, outflow
volume, change in storage and infiltration and evaporation losses from the grid system are summed at
the end of each time step. The difference between the total inflow volume and the outflow volume plus
the storage and losses is a measure of the volume conservation. FLO-2D volume conservation results
are written to the output files and to the screen at run time for the user specified output time intervals.
Any FLO-2D flood simulation not conserving volume must be revised. It should be noted that volume
conservation in any model flood simulation is not exact. Data errors, numerical instability, or poorly
integrated components may contribute to a loss of volume conservation. While some numerical error is
introduced by rounding numbers, approximations or interpolations (such as with rating tables), volume
should be conserved within a fraction of a percent of the inflow volume. The user must decide on an
acceptable level of error in the volume conservation. Most FLO-2D simulations are accurate for volume
conservation within a few millionths of a percent. Generally, volume conservation within 0.0001
percent or less can be considered as a successful flood simulation.
2.6 Organization of the FLO-2D Model
The FLO-2D software package includes a grid developer system (GDS) for generating data input files, a
Mapper program that creates plots of output data and flood maps, and the FREQPLOT program to
analyze flood frequency. The GDS will filter DTM points, interpolate the DTM data and assign elevations
to grid elements. It also graphically creates and edits component data such as channels, streets and
buildings. The Mapper program automates flood hazard delineation. Mapper will plot very detailed
flood inundation color contour maps and shape files. It will also replay flood animations and generate
flood damage and risk maps.
The FLO-2D manual is divided into a series of four documents. The Reference Manual is devoted to
model theory and component description. The Data Input Manual is subdivided into a series of data file
sections devoted to variable definitions, ranges, descriptions and data input instructional comments.
Separate manuals are devoted to the application of the GDS and Mapper.
12
13
plotted as shaded color contours. GIS shape files (*.shp) are automatically created with any plotted
results. The GIS shape files can be then be imported into ArcView or other GIS programs. The Mapper
features and functions are described in its own manual. Some of the Mapper features include:
Importation of multiple geo-referenced aerial photos in various graphic formats such as TIFF,
BMP, JPG, etc.;
Flood damage assessment component to compute the flood damage as a function of the FLO-2D
predicted maximum depths, building shape files and building value tables (dbf file);
For creating FEMA Flood Insurance Rate Maps, the Mapper processor program has a Digital Flood
Insurance Rate Map (DFIRM) optional tool that was created by Anderson Consulting Engineers of Fort
Collins, Colorado. This tool will convert the FLO-2D water surface line contours to those used on a
DFIRM base map. All the required DFIRM symbols and map annotation tools are available. This is an
excellent and precise tool to create the FEMA DFIRMs.
Graphical displays of the output hydrographs and water surface profiles are provided in the HYDROG
and PROFILES post-processor programs. HYDROG will plot the hydrograph for every channel element or
selected floodplain cross section. HYDROG can also be used to evaluate the average channel hydraulics
in a given reach. The user can select the reach upstream and downstream channel elements and the
program will compute the average of all the channel hydraulics in the reach including: velocity, depth,
discharge, flow area, hydraulic radius, wetted perimeter, top width, width to depth ratio, energy slope,
and bed shear stress. The PROFILES program plots channel water surface and bed slope profiles.
2.7 FLO-2D Model Assumptions and Approximations
The basic assumptions and approximations associated the FLO-2D model are:
Hydraulic structure (bridge, culvert, weir) discharge that is approximated by a rating curve or
table;
Pressure pipe flow is not simulated except by a hydraulic structure rating curve or table.
These assumptions and approximations are self-explanatory and do not limit the model from accurately
predicting flood inundation and water surface elevations.
14
Each FLO-2D application involves a unique combination of physical components involving hydrology
(rainfall and inflow hydrographs), flow hydraulics (roughness or water surface control) and topographic
variability (elevations, slope and cross sections). The validation of the primary physical process
15
components will be discussed. The main issues in validating and verifying the FLO-2D model for the
various combinations of the physical components relate to:
The crucial questions regarding model validation and verification using project data bases are:
1.
2.
3.
4.
5.
Accurate model replication of a historical flood event depends on the rainfall or flood inflow
hydrograph(s) data base. The area of inundation is more dependent on the volume of the inflow
hydrograph (overall shape) and rainfall than the peak discharge that may define only a small portion of
the inflow hydrograph. For example, in some alluvial fan applications, the fan area may be comparable
to the watershed area and the fan rainfall may be critical to estimating the total area of inundation.
Infiltration and transmission losses can have a similar impact on the area of inundation and could affect
both model calibration and verification.
Model numerical instability surging occurs when the timestep is too large for the relationship between
the discharge flux and the potential surface area of a grid element. A high discharge flux on a small
surface area can lead to growth of numerical oscillation. The velocity and therefore the discharge may
become unreasonable for one or more timesteps. This instantaneous surge may not affect the overall
distribution of the flood, but it does require model adjustments to avoid reporting invalid maximum
flow depths or velocities. The model automatically reduces the computational timestep based on the
selected stability criteria that can be adjusted to preserve numerical stability. Very small timesteps,
however, can lead to long flood simulation computer runtimes. Long simulations can be avoided by
selecting an appropriate size grid system. There is a tradeoff between model speed (large grid
elements) and mapping resolution (small grid elements). When selecting grid element size, the FLO-2D
manual suggests that:
QpeakAsurf < 1.0 cfs/ft2
where Qpeak = peak discharge and Asurf = surface area of one grid element.
To verify the model for a given set of known or measured flow conditions, it may be necessary to adjust
some of the components that represent water surface control, available floodplain storage, flow path
obstruction, flow distribution, volume gains and losses, and flow confinement. Data bases with
highwater marks, areas of inundation and gage recording may be influenced by:
Complex urban or 2-D unconfined flood simulations typically cannot be calibrated and it is necessary to
perform either validation or verification with simpler models that isolate a specific component.
Furthermore, it is rare that the combination of discharge measurements and corresponding high water
marks are available for an infrequent flood event. A data base where there is corresponding flood cross
section surveys with measured discharge data is exceedingly rare. Validation is often limited to known
analog solutions or flume studies.
3.3 Validity of Computational Core
The flood routing portion of the model essentially includes three core subroutines: channel routing,
floodplain routing and the channel floodplain exchange. The street routing component uses the channel
routing component to simulate a shallow rectangular channel and does not require separate validation.
An assessment of the validity of the computation core routines starts with a volume conservation
report. Every FLO-2D flood simulation must conserve volume or the simulation is inaccurate due to user
input data errors. FLO-2D reports on volume conservation to the computer screen during the flood
simulation and then writes the volume conservation results in a comprehensive output file
(SUMMARY.OUT) that can be reviewed after each computer run. The first FLO-2D validation claim and
substantiation proof is follows:
17
18
19
hydrograph is bulked with sediment, the mudflow is routed as a water and sediment continuum over
the hydrograph. The same water routing algorithm is used for mudflows but the momentum equation is
solved with the additional viscous and yield stress terms. The bulked sediment hydrograph is tracked
through system conserving volume for both water and sediment. Flow cessation and flow dilution are
possible outcomes of the mudflow routing.
The only distinction between unconfined floodplain and alluvial fan flooding is slope and as such it is not
necessary to have any special considerations for alluvial fan flooding. For design flood events, alluvial
fan flood inundation is dictated by topography and roughness and where appropriate by urban features
or hydraulic structures. Thus FLO-2D is an appropriate model for alluvial fan flooding as well as riverine
applications. The first FLO-2D 1999 certification for FEMA Flood Insurance Studies (FIS) flood studies by
the Corps of Engineers was for unconfined flooding on alluvial fans (see the Appendix for the Corps'
Certification Letter).
20
DEPTH = 6.82 ft
Figure 4. FLO-2D Version 2009 Results for Maximum Velocity and Depth
21
Figure 6. Truckee River Observed (Black Outline) and FLO-2D Predicted Area of Flood Inundation
22
The observed area of inundation in Figure 6 is displayed as a black outline. A comparison of fifty
surveyed high water marks and the FLO-2D predicted maximum water surface elevations resulted in
an average difference of only 0.2 ft (0.06 m). There are three important verification anchors to this
project:
The n-values were calibrated to match the original HEC-2 water surface elevation results (Figure
7) and the recorded stage at the USGS gages (Figure 8 and 9);
The predicted area of inundation matches the flooded developed from aerial photographs and
observations very well as shown in Figure 6;
The timing of the floodwave movement through the system matches the downstream Vista gage
recorded stage almost exactly (Figure 9).
Figure 7. illustrates that the FLO-2D predicted water surface elevation matches the HEC-2 water surface
profiles almost exactly through the urban areas. The City of Sparks, the City of Reno and the Truckee
Meadows are located at the downstream end of the profile in Figure 7 (the last 20,000 ft of the model
indicated by the flat slope). Figures 8 and 9 display the predicted versus measured stage reading for the
two USGS gages in operation during the 1997 flood. These stage recordings represent an accurate
assessment of the rise and fall of the hydrograph through the urban areas.
Figure 7. Predicted Truckee River Water Surface Profiles for HEC-2 and FLO-2D
23
Figure 8. FLO-2D Predicted Stage for the 1997 Flood Event at the Reno Gage
24
Figure 9. FLO-2D Predicted Stage for the 1997 Flood at the New Vista Gage
25
25
20
15
10
0
0
10
Time (days)
Figure 10. FLO-2D Version 2009 Predicted Stage for the 1997 Flood at the New Vista Gage
26
In February 2006 the City appealed and in April 2006 FEMA agreed with the appeal and requested more
information and data. The new FLO-2D model FIS study with a 10 ft grid element resulted in the
predicted maximum flow depths in Figure 12 and the FIRM mapping in Figure 13. This highly detailed
model included loss of storage for 5,000 buildings, spatially variable n-values (including the streets),
channel modeling of numerous drains (2,395 channel elements) and numerous hydraulic structures to
simulate the culverts and bridges. The model conserved volume and a review of predicted maximum
velocities and depths indicated that there was no surging in the channel or on the floodplain. Output
hydrographs didn't display any numerical instabilities.
The level of detail in the Camarillo project is unmatched by any comparable two-dimensional flood
project. The City closely tracked and scrutinized this project with their consultant (Kasraie Consulting)
and made recommendations for adjustments based on their flood experience. They reviewed and
concurred with the FLO-2D predicted area of flood inundation. The City of Camarillo approved this FIS
FIRM map submittal to FEMA. For a more detailed discussion of this project see Study Case 3 in Chapter
IV.
27
Figure 12. FLO-2D Predicted Maximum Flow Depths for the City of Camarillo
Figure 13. FLO-2D 2011 Model FIRM for the City of Camarillo
28
Year
Location
Purpose
1996
Floodplain Inundation
1998
BDANWR
Floodplain Inundation
2002
Flood Routing
2002
Model Calibration
2003
2003
Cochiti to Highway 44
URGWOM Corps
2003
URGWOM Support
2004
2004
Cochiti to Albuquerque
Levee Failure
2005
Restoration Plan
2006
2007
Restoration Plans
2008
Flood Routing
2009-10
Notes: SA = San Acacia, SM = San Marcial, BDANWR = Bosque del Apache National Wildlife Refuge, AMAFCA = Albuquerque
Metropolitan Arroyo Flood Control Authority, SOBTF = Save Our Bosque Task Force, Corps = Albuquerque District Corps of
Engineers, URGWOM = Corps Upper Rio Grande Water Operations Model, FWS = U.S. Fish and Wildlife Service, BOR = U.S.
Bureau of Reclamation, IBWC = U.S. International Boundary and Water Commission.
Since the completion of the Middle Rio Grande FLO-2D model from Cochiti Dam to the headwaters of
Elephant Butte Reservoir in 2003, the model has been used to support various Corps and IBWC flood
hazard and river operation studies. The FLO-2D model from Cochiti to Elephant Butte was used by the
Corps of Engineers to support their Upper Rio Grande Water Operations Model (URGWOM). Thirty
years of record were replicated with the calibrated FLO-2D model to predict water delivery to Elephant
Butte Reservoir. The Middle Rio Grande is about 173 miles in length from Cochiti Dam to Elephant Butte
Reservoir. Discharge hydrographs for seven gages throughout the reach for the 30 years of record were
developed. Figure 14 shows an example of the gage data replication. The model was also extended
upstream of Cochiti Dam to Abiquiu Reservoir on the Chama River. The release from the reservoir was
predicted at the confluence of the Chama River and the Rio Grande. The results are shown in Figure 15
29
which depicts the diurnal variations of the power plant release at Abiquiu Dam. FLO-2D predicted this
variation exactly in both magnitude and timing.
In 2005, the Corps implemented an intensive flow monitoring and data collection program for the spring
runoff Cochiti Dam releases that included aerial photography, video, cross section surveys and water
surface elevation measurements. This data base was used to develop a more detailed FLO-2D model.
The model was calibrated for future prediction of overbank flooding and floodwave attenuation. The
250 ft grid element model has 167,308 elements with 3,242 channel elements. This extensive data
collection effort during the controlled dam release enabled the FLO-2D model to be applied to the area
of floodplain inundation. The application used previous calibration infiltration and evaporation
parameters. For more discussion of this project, see Case Study 4 in Chapter IV.
4500
4000
3500
Discharge (cfs)
3000
2500
2000
1500
1000
500
0
04/05/98
04/15/98
04/25/98
05/05/98
05/15/98
05/25/98
06/04/98
06/14/98
06/24/98
07/04/98
07/14/98
07/24/98
Date, MM/DD/YY
Figure 14. USGS San Felipe Gage 1998 (FLO-2D Predicted vs. Measured Discharge)
30
1600
1400
1200
Discharge (cfs)
1000
800
600
400
Gage Discharge
FLO-2D Predicted
Discharge
200
5/9/2003
5/8/2003
5/7/2003
5/6/2003
5/5/2003
5/4/2003
5/3/2003
5/2/2003
5/1/2003
4/30/2003
4/29/2003
4/28/2003
4/27/2003
4/26/2003
4/25/2003
4/24/2003
4/23/2003
4/22/2003
Date
Figure 15. FLO-2D Predicted vs. USGS Gage Measured Discharge at Chama River Confluence
Rio Grande aerial photographs were flown between roughly 10 am and 12 noon on June 1, 2005 when
the Cochiti Dam release was 5,300 cfs with the mean daily flows ranging from 5,040 cfs to 5,650 cfs. A
typical area of inundation comparison with the shape file and aerial photograph in the background is
shown in Figure 16. Flooded overbank areas shown in magenta were digitized by the Corps. The Corps
shape file area estimates involved interpretation of wet floodplain area, estimates of flooded areas that
were obscured by dense vegetation, and the timing of the photography with relationship to the
floodwave movement. The FLO-2D area of inundation is depicted by colored grid elements (mostly blue
and green). The accuracy of the of the FLO-2D predicted area of inundation is limited by the 250 ft grid
element size. Table2 compares the FLO-2D predicted area of inundation against the Corps shape file
estimates. The difference in the two estimates of the floodplain area of inundation is only 6 percent.
The first reach (Cochiti Dam to I-25) had the largest discrepancy because the Corps had difficulty
discerning the flooded area from the anatomizing channel in the aerial photos. In addition, the FLO-2D
model channel data did not accurately represent the complex channel braids in all areas because of lack
of cross section data.
31
Figure 16. FLO-2D Predicted Area of Inundation Compared with the Corps' Digitized Shape File for the
2005 Area of Inundation Monitoring Project
% Difference
155
207
33.8
I-25 to Belen
1746
1931
10.6
1013
999
1.4
4746
5026
5.9
7660
8163
6.2
Total
The FLO-2D prediction of the USGS gage hydrographs for the spring release flow monitoring period from
May 24 through June 1 are shown in Figures 17-20. The model had an initial ramp-up to the release
discharge as shown in each hydrograph plot. The FLO-2D model accurately replicated the discharge
hydrographs recorded at USGS gages for the June 2005 Cochiti Dam release, in part, because the
overbank storage volume has been accurately assessed. It should be noted that the overbank flows are
highly variable along the 128 mile reach of river and yet the model accurately predicts the discharge
within about 100 cfs out 6,000 cfs at all four gages located throughout the reach.
32
Discharge (cfs)
Discharge (cfs)
Figure 17. FLO-2D Predicted vs. USGS Gage Discharge at San Felipe May 24-June 1, 2005
Figure 18. FLO-2D Predicted vs. USGS Gage Discharge at Albuquerque May 24-June 1, 2005
33
Discharge (cfs)
Discharge (cfs)
Figure 19. FLO-2D Predicted vs. USGS Gage Discharge at Isleta Lakes May 24-June 1, 2005
Figure 20. FLO-2D Predicted vs. USGS Gage Discharge at San Acacia May 24-June 1, 2005
The model replication of the gage discharge deviates slightly from the measured data because the gages
are located in a sand bed channel and the gage rating curves shift over time. Scour and deposition in
the sand bed channel limits the accuracy of the reported gage data. Other causes of variability are
attributed to:
34
Evaporation during very hot or very cloudy days may vary significantly from the mean monthly
conditions, and this could result in several percent error in the discharge predictions at downstream
gages. It was concluded from these results that the model accurately replicated the 2005 overbank
monitoring data base. This project exemplifies that if the channel flood routing is accurate in time and
space, then the relationship between the volume of water in the channel and the predicted area of
floodplain inundation will be accurate.
3.3.2 Validity of the Computational Core - Channel Flow
The FLO-2D 1-D channel component has been validated in numerous studies using HEC-RAS and HEC-2
comparisons, flume studies and actual river data (as previously discussed for the Rio Grande). For FEMA
riverine Flood Insurance Studies, the Corps of Engineers prepared a certification letter in 2001
(presented in the Appendix) and the model has been used on numerous river FIS mapping projects. The
most difficult verification test is replicating gage data in both discharge magnitude (or stage) and timing.
Matching gage data is especially difficult for many rivers with dams because depth variable roughness
can play a role in timing of diurnal power releases. Three case studies are presented to validate the
channel routing hydraulics:
It should be noted that since the street routing algorithm (as a rectangular channel) in the FLO-2D model
uses the channel routing subroutine, it is not necessary to present separate validation tests for the FLO2D street component.
35
the start of the reach and 314.26 at the end of the reach. The FLO-2D predicted water surface
elevations were 317.28 and 314.31 for the inflow and outflow nodes. FLO-2D Version 2009 results in
Table 3 show consistent discharge throughout the reach. The predicted velocities and flow depths are
uniform and replicate the hydraulic operating conditions. The actual flow hydraulics deviate slightly
from the design conditions because of the bed slope variability that has evolved with the channel
subsidence (see Case Study 5). FLO-2D Version 2007 results are shown for comparison. The FLO-2D
Version 2009 predicted more consistent discharges and more accurate starting and ending water
surface elevations than Version 2007 because the Version 2009 stability criteria is more refined. This
application validated the FLO-2D dynamic wave routing algorithm for a large channel with a very mild
slope.
36
Version 2007
Version 2009
41902.
8150.00
3865
3960
4055
4150
4245
4340
4436
4532
4531
4530
4626
4721
4720
4719
4815
4911
4910
4909
4998
5086
5085
5084
5171
5258
5257
5256
5342
5341
5340
5425
5424
5423
5507
5506
5505
5587
5586
5585
5584
5664
5663
5662
5661
5738
5737
5736
5810
5809
5808
5807
5806
5805
5804
5803
5802
5801
5800
5799
5798
5797
5796
5795
5794
5793
5792
5791
5790
5789
5788
5787
5786
5785
5784
5783
5782
5706
5627
5626
5544
5543
5459
5458
5372
5285
5284
5283
5282
5281
5280
5279
5191
5102
5101
5012
5011
316.78
316.53
316.50
316.47
316.45
316.42
316.39
316.37
316.34
316.32
316.30
316.27
316.25
316.22
316.20
316.17
316.15
316.13
316.10
316.06
316.04
316.01
315.98
315.94
315.89
315.84
315.81
315.77
315.74
315.71
315.68
315.66
315.63
315.60
315.57
315.54
315.51
315.49
315.46
315.44
315.41
315.39
315.37
315.34
315.32
315.29
315.27
315.24
315.21
315.18
315.15
315.12
315.09
315.06
315.03
315.01
314.98
314.95
314.93
314.91
314.89
314.87
314.85
314.83
314.81
314.79
314.77
314.75
314.73
314.71
314.69
314.67
314.65
314.63
314.61
314.59
314.57
314.54
314.52
314.50
314.48
314.46
314.43
314.41
314.39
314.37
314.35
314.33
314.31
314.28
314.26
314.24
314.21
314.19
314.17
26.52
26.31
26.32
26.30
26.29
26.27
26.25
26.24
26.21
26.20
26.16
26.11
26.07
26.02
25.98
25.93
25.89
25.86
25.80
25.74
25.70
25.63
25.63
25.66
25.67
25.70
25.73
25.81
25.91
26.04
26.13
26.16
26.16
26.15
26.13
26.13
26.12
26.12
26.12
26.11
26.10
26.10
26.11
26.13
26.15
26.15
26.18
26.19
26.19
26.20
26.20
26.21
26.19
26.16
26.16
26.18
26.20
26.21
26.23
26.25
26.27
26.29
26.31
26.34
26.36
26.38
26.40
26.42
26.44
26.47
26.49
26.51
26.53
26.55
26.54
26.55
26.56
26.55
26.56
26.56
26.57
26.57
26.57
26.59
26.59
26.60
26.60
26.60
26.60
26.59
26.60
26.61
26.60
26.61
26.61
2.73
2.75
2.75
2.75
2.76
2.76
2.76
2.76
2.77
2.77
2.78
2.78
2.79
2.80
2.80
2.81
2.82
2.82
2.83
2.84
2.85
2.86
2.86
2.92
3.06
3.13
3.12
3.11
3.09
3.07
3.06
3.05
3.05
3.05
3.05
3.05
3.06
3.06
3.06
3.06
3.06
3.06
3.06
3.06
3.05
3.05
3.05
3.05
3.04
3.04
3.04
3.04
3.04
3.05
3.05
3.05
3.04
3.04
3.04
3.03
3.03
3.03
3.02
3.02
3.01
3.01
3.01
3.00
3.00
2.99
2.99
2.99
2.98
2.98
2.98
2.98
2.98
2.98
2.98
2.98
2.98
2.98
2.98
2.98
2.97
2.97
2.97
2.97
2.97
2.97
2.97
2.97
2.97
2.97
2.97
25153.
8150.00
8150.54
8158.49
8159.69
8156.24
8160.54
8158.80
8156.42
8153.06
8154.96
8154.77
8154.92
8153.67
8151.70
8154.46
8154.53
8153.75
8154.63
8155.47
8157.89
8161.16
8160.14
8161.48
8160.33
8159.09
8159.27
8161.08
8160.43
8161.72
8162.23
8161.16
8159.82
8160.12
8159.51
8160.55
8159.09
8159.73
8160.62
8160.48
8164.11
8165.50
8164.84
8165.92
8167.59
8165.12
8165.93
8166.05
8164.24
8163.69
8165.67
8166.55
8165.55
8165.37
8163.04
8161.77
8159.96
8159.38
8158.63
8159.64
8161.95
8160.76
8159.55
8159.54
8163.16
8163.08
8159.30
8158.60
8158.71
8159.30
8160.53
8159.85
8156.63
8156.08
8156.58
8154.97
8158.40
8158.06
8158.95
8160.02
8161.15
8161.10
8162.95
8162.79
8162.18
8161.18
8160.86
8161.31
8162.84
8163.19
8161.24
8162.85
8166.84
8168.79
8168.72
8169.45
8171.22
3865
3960
4055
4150
4245
4340
4436
4532
4531
4530
4626
4721
4720
4719
4815
4911
4910
4909
4998
5086
5085
5084
5171
5258
5257
5256
5342
5341
5340
5425
5424
5423
5507
5506
5505
5587
5586
5585
5584
5664
5663
5662
5661
5738
5737
5736
5810
5809
5808
5807
5806
5805
5804
5803
5802
5801
5800
5799
5798
5797
5796
5795
5794
5793
5792
5791
5790
5789
5788
5787
5786
5785
5784
5783
5782
5706
5627
5626
5544
5543
5459
5458
5372
5285
5284
5283
5282
5281
5280
5279
5191
5102
5101
5012
5011
37
317.28
317.04
317.01
316.98
316.95
316.93
316.90
316.88
316.85
316.83
316.80
316.78
316.75
316.73
316.70
316.67
316.65
316.62
316.59
316.56
316.54
316.51
316.48
316.46
316.42
316.37
316.33
316.30
316.27
316.23
316.20
316.17
316.14
316.11
316.07
316.04
316.01
315.98
315.95
315.92
315.89
315.86
315.84
315.81
315.78
315.75
315.72
315.69
315.66
315.63
315.59
315.56
315.52
315.49
315.46
315.42
315.39
315.36
315.33
315.30
315.27
315.24
315.21
315.18
315.15
315.12
315.09
315.07
315.04
315.01
314.98
314.95
314.92
314.89
314.86
314.83
314.80
314.76
314.73
314.70
314.67
314.64
314.61
314.57
314.54
314.51
314.48
314.45
314.42
314.39
314.35
314.32
314.28
314.25
314.21
27.03
26.82
26.83
26.81
26.79
26.78
26.76
26.75
26.72
26.71
26.66
26.62
26.57
26.53
26.48
26.43
26.39
26.35
26.29
26.24
26.20
26.13
26.13
26.18
26.20
26.23
26.25
26.34
26.44
26.56
26.65
26.67
26.67
26.66
26.63
26.63
26.62
26.61
26.61
26.59
26.58
26.57
26.58
26.60
26.61
26.61
26.63
26.64
26.64
26.65
26.64
26.65
26.62
26.59
26.59
26.59
26.61
26.62
26.63
26.64
26.65
26.66
26.67
26.69
26.70
26.71
26.72
26.74
26.75
26.77
26.78
26.79
26.80
26.81
26.79
26.79
26.79
26.77
26.77
26.76
26.76
26.75
26.75
26.75
26.74
26.74
26.73
26.72
26.71
26.70
26.69
26.69
26.67
26.67
26.66
2.66
2.67
2.67
2.68
2.68
2.68
2.68
2.69
2.69
2.69
2.70
2.70
2.71
2.72
2.72
2.73
2.74
2.74
2.75
2.76
2.77
2.77
2.78
2.84
2.97
3.03
3.03
3.01
3.00
2.98
2.96
2.96
2.96
2.96
2.96
2.96
2.97
2.97
2.97
2.97
2.97
2.97
2.97
2.97
2.97
2.97
2.96
2.96
2.96
2.96
2.96
2.96
2.97
2.97
2.97
2.97
2.97
2.97
2.96
2.96
2.96
2.96
2.96
2.95
2.95
2.95
2.95
2.95
2.94
2.94
2.94
2.94
2.94
2.93
2.94
2.94
2.94
2.94
2.94
2.94
2.94
2.94
2.94
2.94
2.94
2.95
2.95
2.95
2.95
2.95
2.95
2.95
2.96
2.96
2.96
8150.00
8150.05
8150.01
8150.00
8150.00
8150.00
8149.99
8149.99
8149.98
8149.98
8149.97
8149.97
8149.96
8149.96
8149.94
8149.95
8149.94
8149.94
8149.92
8149.93
8149.92
8149.91
8149.89
8149.89
8149.89
8149.88
8149.87
8149.86
8149.86
8149.85
8149.84
8149.84
8149.83
8149.82
8149.82
8149.81
8149.81
8149.80
8149.79
8149.78
8149.78
8149.77
8149.76
8149.75
8149.74
8149.74
8149.73
8149.73
8149.73
8149.72
8149.72
8149.71
8149.71
8149.70
8149.70
8149.69
8149.69
8149.69
8149.69
8149.68
8149.67
8149.66
8149.66
8149.66
8149.65
8149.64
8149.64
8149.63
8149.62
8149.62
8149.62
8149.61
8149.61
8149.60
8149.61
8149.59
8149.58
8149.57
8149.57
8149.56
8149.55
8149.55
8149.53
8149.54
8149.55
8149.53
8149.53
8149.53
8149.53
8149.54
8149.51
8149.52
8149.51
8149.52
8149.51
38
Table 4. WSPG Hydraulic Results for the Camarillo Hills Drain Los Posas Road to Highway 101
39
1015765
1017063
1018358
1019653
1020946
1022236
1023525
1024814
1026100
1027385
1028667
1029949
1031229
1032506
1033783
1035059
1036332
1037604
1038874
1040143
1041388
1042614
1043820
1045005
1046168
1047312
1048434
1049536
1050618
1051679
1052721
1053741
1054740
1055720
1056678
1057617
1058535
1059432
1060309
1061164
1061999
1062813
1063609
1064382
1065134
1065868
1066580
1067271
1067942
1068593
1069223
1069832
1070421
1070989
1071536
1072063
1072062
1072569
1073055
1073054
1073520
1073965
1073964
1074388
1074791
1074790
1075173
1075537
1075536
1075879
1076201
1076200
1076501
1076785
1076784
1077066
1077065
1077344
1077620
1077619
1077894
1078166
1078165
1078434
1078700
1078699
1078964
ELEVATION
DEPTH
102.93
99.35
98.33
97.82
97.53
97.32
97.17
97.04
96.94
96.85
96.76
96.69
96.62
96.55
96.50
96.45
96.39
96.34
96.29
96.24
96.19
96.14
96.10
96.06
96.01
95.97
95.92
95.89
95.85
95.81
95.76
95.73
95.69
95.65
95.62
95.59
95.56
95.53
95.50
95.47
95.45
95.43
95.41
95.39
95.38
95.37
95.35
95.33
95.30
95.28
95.25
95.22
95.19
95.16
95.13
95.10
95.07
95.04
95.01
94.98
94.94
94.90
94.86
94.82
94.78
94.75
94.71
94.67
94.63
94.59
94.55
94.52
94.48
94.45
94.41
94.38
94.34
94.31
94.27
94.24
94.20
94.16
94.13
94.09
94.05
94.01
93.98
13.83
10.30
9.33
8.85
8.60
8.42
8.31
8.21
8.15
8.09
8.04
8.00
7.97
7.93
7.92
7.90
7.88
7.86
7.85
7.83
7.82
7.80
7.80
7.79
7.78
7.77
7.76
7.76
7.76
7.75
7.74
7.74
7.73
7.73
7.73
7.74
7.74
7.75
7.75
7.76
7.77
7.79
7.80
7.82
7.84
7.87
7.88
7.90
7.90
7.92
7.92
7.93
7.93
7.94
7.94
7.95
7.95
7.96
7.96
7.96
7.96
7.95
7.95
7.94
7.94
7.94
7.94
7.93
7.93
7.92
7.92
7.92
7.92
7.92
7.92
7.92
7.92
7.92
7.92
7.92
7.92
7.91
7.92
7.91
7.91
7.90
7.91
VELOCITY OUT
10.72
11.54
13.72
14.45
15.00
15.34
15.57
15.74
15.88
15.99
16.08
16.16
16.23
16.29
16.34
16.38
16.41
16.44
16.48
16.51
16.54
16.57
16.59
16.61
16.63
16.65
16.66
16.67
16.68
16.69
16.71
16.72
16.74
16.74
16.74
16.73
16.72
16.71
16.69
16.68
16.66
16.63
16.59
16.56
16.51
16.47
16.43
16.40
16.38
16.36
16.34
16.32
16.31
16.30
16.29
16.29
16.27
16.27
16.25
16.25
16.26
16.27
16.27
16.28
16.29
16.29
16.30
16.31
16.31
16.32
16.33
16.34
16.34
16.34
16.34
16.34
16.34
16.34
16.34
16.34
16.34
16.34
16.35
16.35
16.36
16.36
16.36
4786.00
DISCHARGE OUT
4786.00
4786.04
4786.18
4786.21
4786.13
4786.10
4786.08
4786.07
4786.07
4786.06
4786.06
4786.06
4786.05
4786.05
4786.05
4786.05
4786.05
4786.05
4786.05
4786.05
4786.04
4786.04
4786.05
4786.05
4786.05
4786.04
4786.04
4786.04
4786.05
4786.05
4786.04
4786.04
4786.03
4786.04
4786.03
4786.03
4786.03
4786.03
4786.03
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.02
4786.01
4786.01
4786.02
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.02
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
4786.01
Table 5. FLO-2D Results for Camarillo Hills Drain 100-year Design Discharge
40
41
Figure 21. 1997 High Flow Season Hydrograph at Jensen, Utah (FLO-2D vs. Measured Discharge)
42
Figure 22. November 1998 Power Plant Discharge Release from Flaming Gorge Dam
43
Figure 23. FLO-2D Predicted vs. Measured Discharge, Jensen USGS Gage for Power Plant Releases at Flaming Gorge Dam
44
Figure 24. Version 2009 FLO-2D Predicted vs. Measured Discharge, Jensen USGS Gage for Power Plant Releases at Flaming Gorge Dam
45
46
47
3.4 Hydrology
3.4.1 Overview
Water inflow to the FLO-2D model can include an unlimited number of inflow hydrographs, rainfall,
stage-discharge or time-stage data. Rainfall can be uniform or spatially variable using either depth area
reduction values or NEXRAD radar data. For uniform rainfall, the storm is discretized as a cumulative
percent of the total rainfall. There are a number of options to simulate variable rainfall including a
moving storm, spatially variable depth area reduction, or grid based rainfall data from an actual storm
event. A rainfall distribution can be selected from a number of predefined distributions. The rainfall
runoff can be routed off the watershed and into to the channel system in the same flood simulation.
3.4.2 Hydrology Validation
Table 6. SUMMARY.OUT file for the Diamond Project to Verify Rainfall Volume Conservation
Once the rainfall on the grid system has ponded to a depth greater than the tolerance value (TOL in
TOLER.DAT file), the water can be exchanged between grid elements as overland flow.
48
Figure 26. NEXRAD Cell Total Precipitation July 31, 2006 on the Catalina Mountain Watersheds
(Provided by the USGS email correspondence August 7, 2007)
49
Since the infiltration losses were minor because of the high antecedent moisture conditions, it was not
necessary to calibrate the FLO-2D predicted rainfall runoff. As a result, the FLO-2D generated flood
hazard maps closely replicated the aerial photo taken on the morning of July 31, 2006 (Figure 27). The
predicted flood paths in blue in Figure 28 can be precisely correlated with the green flow paths in Figure
27. It was concluded that the FLO-2D model hydrology component accurately predicted the flood
inundation as a result of the detailed temporal and spatial discretization of the storm event.
Figure 27. Lower Soldier Canyon Alluvial Fan Aerial Photo. Agua Caliente Wash crosses
Soldier Trail (center photograph). Photograph by Terry Hendricks (PCRFCD; 7-31-06).
50
Figure 28. Soldier Canyon Alluvial Fan FLO-2D Predicted Area of Inundation for the July 31, 2006 Flood
In a second NEXRAD rainfall project, the Flood Control District of Maricopa County (FCDMC) applied the
FLO-2D model on the Rainbow Wash watershed (Figure 29) to replicate a 2005 storm. The watershed
was gaged and had measured discharge data for the storm. FCDMC funded the development of the GDS
tool that interpolates the NEXRAD data to FLO-2D grid elements. The FLO-2D model results prepared by
the FCDMC were compared with the measured data and with comparable HEC-1 model results (Figure
30). The HEC-1 model was the previously required project hydrologic model for the FCDMC. The FLO2D model is also now an accepted hydrologic model for FCDMC projects. The discrepancy between the
gage data and the FLO-2D predicted recessional limb of the basin outflow hydrograph were attributed to
gage cross section scour. Modeling the concentrated flow as channel discharge later also improved the
timing on the rising limb. The FCDMC (2010) stated, "(a)s part of this study, temporal-based NEXRAD
radar data was used with FLO-2D and HEC-1 to model a significant storm that occurred over the 18
51
square mile Rainbow Wash watershed in 2005. Using the FCDMC standard modeling parameters
without calibration, FLO-2D very closely replicated the measured runoff hydrograph storm event, while
HEC-1 significantly over predicted the runoff volume and peak discharge." Figure 31 indicates that the
FLO-2D accurately replicated the area of inundation for this 2005 flood event.
Figure 29. FLO-2D Predicted Maximum Depth for Rainbow Wash (FCDMC, 2010)
7,000
Observed
FLO-2D Saxton 2005
HEC-1 2005
6,000
Discharge, cfs
5,000
4,000
3,000
2,000
Volume Comparisons:
a. HEC-1 1993 Runoff Volume: 458 % of measured
b. Saxton 2005 Runoff Volume: 138 % of measured
1,000
Time
Figure 30. FLO-2D Predicted Hydrograph vs. Gage Data and HEC-2D Results (FCDMC, 2010)
52
Figure 31. FLO-2D Predicted Inundation Area vs. the Observed Flooding from 2005 Storm (FCDMC, 2010)
the initial and final soil saturation conditions. The FLO-2D Green-Ampt code was provided to the
FCDMC and was manually verified by FCDMC engineering staff using HEC-1 and an Excel spreadsheet
with interactive code.
The FCDMC's Rainbow Wash simulation demonstrates the accuracy of the Green-Ampt infiltration
method. Nevertheless, a simple infiltration model was created to validate the Green-Ampt component.
Using the Diamond project that demonstrated rainfall volume accuracy, a uniform rainfall of 0.5 inches
in one hour was applied along with a Green-Ampt infiltration hydraulic conductivity of 0.5 inches/hr.
There was zero abstraction, no soil suction and no impervious area. The correct result of this model is
no flow depth, 0.5 inches of rainfall and 0.5 inches of infiltration as shown is Table 7.
The SCS curve number parameters can be assigned graphically in the GDS to allow for spatially variable
rainfall runoff. Shape files can used to interpolate SCS-CNs from ground cover and soil attributes. The
SCS-CN method can be combined with the Green-Ampt infiltration method to compute both rainfallrunoff and overland flow transmission losses. For the combined loss routines, the SCS-CN method will
be applied to grid elements with rainfall during the model computational timestep and the Green-Ampt
method will compute infiltration for grid elements that do not receive rainfall during the timestep. This
enables transmission losses to be computed with Green-Ampt on alluvial fans and floodplains while the
SCS-CN is used to compute the rainfall loss in the watershed basin. The FLO-2D SCS method calculations
were verified for Pima County by JE Fuller Hydrology and Geomorphology and Stantec Consultants with
hand computations and simplified project applications. Application of a CN = 80 for the Diamond
Project with a total rainfall (P) of 3.1 inches results in a total volume of actual runoff Q:
Q = (P - 0.2 S)2/ (P + 0.8 S) = (3.1 - 0.2 * 2.5)2/(3.1 + 0.8 * 2.5) = 1.33 inches
where:
S = 1000/CN - 10 = 2.5
which in Table 8 below is the equivalent of the total point rainfall minus the infiltrated and abstracted
water 3.1 inches - 1.78 inches = 1.32 inches. Applying this value to the volume conservation in Table 5,
there are 10374 grid elements 300 ft square:
Volume of Runoff = 1.32 inches/12 inches/ft * 10374 * 300 * 300 ft2/43560 ft2/acre = 2,357.73 acre-ft
The sum of the floodplain storage plus the floodplain outflow = 2,316.01 + 41.66 = 2,357.67 acre-ft as
shown in Table 8.
Figure 32. FLO-2D Model Predicted Depth and Velocities for White Tanks Fan 36 (JE Fuller, 2010)
56
Figure 33. FLO-2D Mode Predicted Depth and Velocities for White Tanks Fan 36
The PFHAM report and methodology was submitted to a panel of experts ("Blue Ribbon Panel) for peer
review which met on June 2-3, 2010. The Blue Ribbon Panel consisted of experts from a variety of
engineering, scientific, and regulatory disciplines associated with alluvial fan flood hazard assessment.
The Blue Ribbon Panel concluded that the methodology was "...reasonable, defensible, and scientifically
sound..." and that two-dimensional modeling was "... strongly recommended for alluvial fan flood
hazard assessment" including flooding attenuation. It was also concluded that the proposed hazard
assessment methodology using the FLO-2D Mapper program flood hazard delineation method of low,
moderate and high hazard was acceptable.
3.4.3 Summary of the Hydrology Component
The hydrologic component in FLO-2D is critical to accurate flood hazard delineation because the area of
inundation is primarily a function of both the flood inflow hydrographs plus effective rainfall on the
project area. Rainfall can occur directly on ponded areas, rivers, streets, buildings and infiltration losses
can be computed for any surface. The rainfall runoff from urban areas or watersheds can be routed
over the grid system to the water courses and then routed downstream in the channel component all in
one model simulation. The combined FLO-2D hydrologic and hydraulic flood routing is a significant
advancement over the traditional 1-D lumped parameter unit hydrograph method watershed approach.
The FCDMC (2010) stated that the District was "...having great success with the application of the FLO2D hydrologic and hydraulic modeling in support of mapping and managing flood hazard areas within
Maricopa County." "...(T)his year FLO-2D is also being used for areas within our ADMP study areas to
57
model watersheds that would normally be modeled with HEC-1. FCDMC would not be using it as the
primary modeling tool on this year's ADMP's if it were not a valuable and appropriate tool." The PFHAM
report by JE Fuller (2010) recommended that the "...engineering tool for hydrologic modeling of active
alluvial fans and alluvial plains in Maricopa County is FLO-2D".
3.5 Component Applicability
The FLO-2D model can be used to simulate the complete hydrologic cycle including rainfall, infiltration,
evaporation, groundwater-surface water exchange and flood routing. Model components used in the
hydrologic and hydraulic simulations include channel and street routing, urban flooding, alluvial fan
flooding, sediment transport, mud and debris flows, levee confinement, levee overtopping and levee
and dam breach. Each flood project is unique in its combined utilization of these components. These
combinations lead to an infinite range of potential variability in a given model application and not all of
these components are required for every flood hazard delineation project. For FEMA Flood Insurance
Studies (FIS), the flooding is generally of short duration so components such as evaporation,
groundwater-surface water exchange are not required. In addition, for FIS studies sediment transport
modeling is not required. As such, these particular components along with the mud and debris flow
component and the dam breach erosion component are not discussed in this document. The FLO-2D
model does not simulate the following flood conditions:
Pressure flow is not simulated directly, but it can be simulated by a third party program such as
a storm drain model and applied as a hydraulic structure rating table.
Levee confinement and levee overtopping are the remaining FLO-2D components left for discussion.
Claim 18: Levee confinement is accurately simulated by
FLO-2D in eight potential flow directions.
Substantiation of Claim 18. Levee assignment occurs along the grid element boundary in any of the 8
potential flow directions. The user assigns the levee blocked direction and the crest elevation. As long
as the crest elevation is greater than the water surface elevation, the levee will stop flow from being
exchanged between grid elements. Figure 34 displays the Middle Rio Grande floodplain flooding
confined by the levee system on both sides of the river. This project is one of the FLO-2D example
projects and is used as a benchmark test for model revisions.
58
Figure 34. Rio Grande Overbank Flooding Confined by Levees (red outline in Maxplot)
59
Figure 35. Simple Levee Example Project with 1,000 cfs Steady Discharge
Figure 36. Corps' Monroe Alluvial Fan Project with a Defined Channel Using Version 2007
Figure 37. Corps' Monroe Alluvial Fan Project Difference in Maximum Depth
(all the differences in maximum depth are 0.2 ft or less)
62
MAXIMUM VELOCITY
FPS OR MPS)
2946
2922
2968
2985
2998
2793
2753
2830
2863
2894
1172
2666
2619
1414
2567
1011
1253
931
1092
1334
369
289
2455
2394
14.93
14.93
13.81
12.06
11.19
11.16
11.16
10.31
9.82
9.72
8.98
8.83
8.83
8.79
8.77
8.62
8.59
8.41
8.37
8.36
8.33
8.33
8.28
8.28
TIME OF OCCURRENCE
(HRS)
2.59
2.59
3.02
2.50
2.31
3.04
3.04
3.04
3.03
3.03
3.48
3.06
3.06
4.23
3.13
3.47
3.47
3.19
3.03
2.71
3.83
3.83
3.11
3.11
Table 10. Difference in Maximum Velocity Between Version 2007 and Version 2009 for the Corps'
Monroe Project
When reviewing the Monroe project along with the Flume Case Study and the Camarillo Hills Drain, it is
demonstrated that project results can vary with model enhancements to improve speed and stability
while still converging to known analog solution results.
A source of model version differences are changes or enhancements to the model detail components.
This may include changes in variables, new components or adding spatially variable components. A
complete list of the model revisions between Version 2007 and Version 2009 are presented in the
Appendix. Some recent model revisions include:
Courant stability parameter previously hardwired at 1.0 is now a user assigned variable.
Rainfall runoff from buildings that are completely blocked.
Spatially variable limiting Froude numbers.
The sheer volume of computations during a flood simulation can also be a source of noise error
generation and slight variability in results. Not all of the variables and arrays in the FLO-2D model are
coded as double precision numbers. Real numbers only track 8 significant digits. When real numbers
are employed instead of double precision numbers, small differences between large numbers can lead
to variation in hydraulic results with changes in timestep magnitude.
63
Overland flow
Channel flow
Channel/overland flow exchange
Street flow
Buildings (ARF/WRF values)
Hydraulic structures
Rainfall/Infiltration
Levees and levee overflow
Levee and dam breach
Sediment transport
A batch file program has been created that runs all the benchmark tests sequentially. Each test is
automatically compared against a known or previous set of results and a message is displayed indicating
that either the test passed or failed for a specific component. The benchmark tests are run to verify that
the model revision produces the anticipated results. All benchmark tests are used to verify that any
code revisions have either negligible effect on the model output, generate differences within acceptable
limits, or meet anticipated variation in the results as in the case of model enhancements. The FLO-2D
benchmark testing program is designed to provide model consistency and reliability. Baseline projects
are also used in the testing program to check component interaction. The baseline projects are actual
flood projects that are considerably more complex than the individual component benchmark tests and
encompass a broad range of flow conditions and component interaction.
It is the intent of FLO-2D Software, Inc. to be transparent with its modeling practices. When a modeling
issue is identified, the workflow of the FLO-2D staff is as follows (Table 11):
Event/Action
1.
2.
3.
4.
5.
6.
Outcome
64
FLO-2D model revision and update since June 2010 posted on the FLO-2D website (www.flo-2d.com).
Future model revisions will be tagged with a version and build number identified with a numeric code in
the format:
VV-YR.MM.BB
where:
-VV is a two digit number indicating the version number (e.g. 11 for 2011)
-YR is a two digit number of the year that major new components or routines were implemented
(e.g. 12 for 2012);
-MM is a two digit number of the month that major new components or routines were implemented
(e.g. 06 for June);
-BB is a two digit number that identifies the executable build. A change in the build number
represents a significant bug fix or model revision. Build numbers are not changed for typos,
minor revisions or enhancements.
For example, 11-12.04.02 indicates that the 2th build or compiled revised executable of Version 2011
was released to the public and posted on the website in April of 2012. The model writes this tag to the
SUMMARY.OUT file and other places. A FLO-2D Model Revisions document is updated every time there
is a bug fix or enhancement in the model or processor programs. In the document, each update release
to the public will have build tag number listed to identify the model revisions with that build. Finally in
all new versions, the model build number will be displayed on the FLO-2D model runtime introductory
message to the screen, on the screen graphics, in the About dialog box in the GDS, in the CONT.DAT file
line one and in SUMMARY.OUT file.
65
Substituting:
Q = 8,000 cfs = (1.486/0.05) *d2/3 * (0.003)1/2 * 200 ft * d
d = (8,000/325.57)3/5 = 6.83 ft
V = Q/A = 8,000 cfs /(6.83 ft*200 ft) = 5.86 fps
This case study validated that the FLO-2D model computes overland flow accurately. Variations of this
example has been used repeatedly by consultants and studies to challenge the basis of the eight flow
direction FLO-2D model. In fact, physical model flumes studies have been undertaken in universities to
test the FLO-2D flume results.
66
Channel routing
Channel-floodplain exchange (both overbank flow and return flow to the channel)
It is noteworthy that this project marked the end of the use of the channel power regression
relationships for the channel cross sections. If the actual surveyed cross section data were used in the
model instead of the power regression relationship, the predicted stage at the New Vista gage would be
an almost perfect match. Only power regression relationship were used to represent the cross sections
when the model was initially developed for this project.
67
Reno Gage
Figure 38. Sparks Truckee Meadows Area General Peak Flow Path (Google Earth, 2011)
68
Figure 39. City of Camarillo 10 ft FLO-2D Grid System with the Building Shape File
Figure 40. Detail of the City of Camarillo 10 ft FLO-2D Grid System with LiDAR Data
69
There were several areas that had limited predicted flooding in the 10 ft model compared to the 100 ft
model based on the higher resolution topography, streets, and the assignment of building area/width
reduction factors. In many areas, the flood inundation was confined to the streets. The area between
Lewis Road and Ponderosa Drive depicts the refined model resolution and the importance of flood
conveyance in the streets (Figure 42).
Figure 42. Flooding Differences Between the Models (100 ft model left and 10 ft model right)
Some of the predicted subdivision flooding in the 100 ft model was not predicted in the higher
resolution 10 ft model (Figure 43).
Figure 43. Flooding along Ponderosa Drive (100 ft model left and 10 ft model right)
The refined model had better resolution of grid element elevation along the highway crest. In the 100
ft model, flooding crosses Highway 101 (Figure 44), but the flood is contained by highway topography in
the 10 ft model (Figure 45). The interpolation of lower elevations on each side of the highway reduced
the highway element elevations in the 100 ft model. The 10 ft model predicts that the flooding would
be confined to elements north of the east bound lanes.
70
71
The digital terrain model (DTM) data base was compiled from six different mapping efforts by the Corps
and Bureau of Reclamation over six years. Each DTM data set represented a specific reach of the Middle
Rio Grande. The DTM data sets were provided in various formats and had different reference elevation
datum. All the data sets were compiled and converted to a consistent datum using the New Mexico
State Plane (Central) North American Datum (NAD) 1988 horizontal and North American Vertical Datum
(NAVD) 1988 vertical reference. The resolution of the DTM data varies by reach. The resolution of the
original DTM data sets was adequate to generate 2 ft contour maps.
Over 480 cross sections had been surveyed throughout the Middle Rio Grande (MRG) from Cochiti Dam
to Elephant Butte Reservoir. Most of the cross sections were surveyed in conjunction with the Bureau
of Reclamations river maintenance program. These cross sections were distributed and interpolated to
the 3,241 channel elements in the model. Levees constrain the potential flooding through most of the
MRG valley and the levees on both sides of the river are incorporated into the model. In the MRG
model, the inflow hydrograph was assigned to a channel element representing the Cochiti Dam outlet
works and outflow nodes were assigned to grid elements in the headwaters of Elephant Butte Reservoir.
The FLO-2D MRG model was used to predict the area of inundation, water surface elevation and
discharge anywhere in the MRG system.
FLO-2D model calibration was accomplished by reach. The primary data base that was used for the
model calibration was the Corps 2005 overbank monitoring study. The various USGS gage data were
used to replicate the flow hydrographs. The model discharge routing (compared with gage
hydrographs) and model spatially variable flood inundation (compared with infrared color photography)
were reviewed with each calibration run. Water surface elevations were measured at specific times at
various cross section locations. FLO-2D predicted water surface elevations were written to a file along
with the measured water surface elevations at the simulated survey time (12 noon, June 1, 2005).
Model calibration was a balance between the hydrograph calibration, water surface calibration and
matching the area of inundation.
Two validation runs were made for the period from April 3-11, 2004 (Figure 46) and May 5-22, 2004
(Figure 47) from San Acacia to San Marcial for discharges in the range of 1,500 cfs to 3,000 cfs. No
adjustments to the data files were made for these additional simulations. For the April 2004
hydrograph, FLO-2D under-predicts the reported discharge at the USGS San Marcial gage and for the
lower-flow, May 2004 hydrograph, the model over-predicts the reported discharge at the San Marcial
gage. This demonstrates the variability of the gage-discharge relationship and frequent mobile bed
shifts at the San Marcial gage. Using the identical model but different inflows based on the San Acacia
gage, the reported discharge is about 20 percent different from the predicted discharge for a portion of
the hydrograph, but high in one case and low in the other. In both hydrographs, however, the predicted
and reported discharges converge near the end of the hydrograph. This epitomizes the difficulty that
flood modelers have with field data bases associated with mobile bed river systems. Frequent gage
calibration discharge measurements are required to reset the reference stage. Once a flood routing
model has been calibrated with gage data, it generally will produce more consistent and accurate
discharge predictions than the gage measurements on the sand bed river.
72
Figure 46. FLO-2D Predicted vs. USGS Gage Discharge at San Marcial April 3-11, 2004
00
Figure 47. FLO-2D Predicted vs. USGS Gage Discharge at San Marcial May 5-22, 2004
model because of the very mild slope (5.7 x 10-5), high flow depth and low Froude numbers. In addition,
the bedslope is variable because of land subsidence associated with agricultural groundwater pumping
(Figures 50). The variable bed slope altered the design discharge operational conditions.
Arroyo Pasajero
California Aqueduct
Figure 48. Aerial Image of the FLO-2D Reach of the California Aqueduct Near Huron
(from Google Earth, 2011)
Figure 49. California Aqueduct Near Huron (from Kewaneh, Google Earth, 2011)
74
downstream flooded bottomlands and backwater areas, they quickly grow in the warmer nutrient rich
water. If they are flushed out of this habitat back into the river, fish drift into Lake Powell downstream
of the Colorado River confluence and are preyed upon by nonendemic lake fish species. To assess the
variability on backwater habitat from the power plant releases, the FLO-2D model simulated the diurnal
power generation hydrograph shown in Figures 52-54. The location of the Jensen gage is shown in
Figure 51.
Ouray Gage
Jensen Gage
Colorado River
Confluence
End of Uinta Valley
Figure 51. Green River Bed Slope from Flaming Gorge Dam to the Colorado River Confluence
While the integrity of the Flaming Gorge power release spikes is sustained through the steep Green
River Canyon in Dinosaur National Park to the Jensen Gage (Figure 24 repeated below), the spikes are
largely attenuated by the middle of the mild slope Uinta Valley at the Ouray gage (Figure 52) and they
are essentially melded into one hydrograph crown by the end of Uinta Valley as shown in Figure 53. By
the Colorado River Confluence, the two spike releases are a single hump as shown Figure 54.
Figure 24. FLO-2D Predicted vs. Measured Discharge, Jensen USGS Gage for the Flaming Gorge Releases
76
Figure 52. Flaming Gorge Power Plant Spike Release Hydrograph at the Ouray Gage
Figure 53. Flaming Gorge Power Plant Spike Release Hydrograph at the end of Uinta Valley
Figure 54. Flaming Gorge Power Plant Spike Release Hydrograph at the Colorado River Confluence
77
V. References
Cunge, J.A., F.M. Holly Jr., and A. Verwey, 1980. Practical Aspects of Computational River Hydraulics,
Pittman Advanced Publishing Program, London, UK.
Fletcher, C.A.J., 1990. Computational Techniques for Fluid Dynamics, Volume I, 2nd ed., Springer-Velag,
New York.
Flood Control District of Maricopa County, 2010. Letter to Jim O'Brien Re: "The Use of FLO-2D for
Hydrologic Modeling in Maricopa County", Phoenix, AZ.
Green, W.H. and G.A. Ampt, 1911. "Studies on soil physics, part I: The flow of air and water through
soils," J. of Agriculture Science.
International Association of Hydro-Environment Engineering and Research, 1994. "Guidelines for
Documenting the Validity of Computational Modelling Software", Madrid, Spain, www.iahr.com.
James, M.L., G.M. Smith, and J.C. Wolford, 1985. Applied Numerical Methods for Digital
Computation, Harper and Row, New York, N.Y., Third Edition.
Jin, M. and D.L. Fread, 1997. Dynamic flood routing with explicit and implicit numerical solution
schemes, J. of Hyd. Eng., ASCE, 123(3), 166-173.
JE Fuller Hydrology and Geomorphology, 2010. " Refinement of Methodology: Alluvial Fan Flood Hazard
Identification & Mitigation Methods," Final Report submitted to the Flood Control District of
Maricopa County, Phoenix, AZ.
Michael Baker, Jr., 2001. "A Review of the FLO-2D Model for Riverine Applications," Letter Report
prepare for FEMA, by the Alexandria, VA office.
Miller, J.J., 1996. "Comparison of Models to Mitigate Flood Hazard to Transportation Alignments on
Alluvial Fans," Masters Thesis, Dept. of Geoscience, Univ. of Nevada, Las Vegas, NV.
Nalesso, M., 2009. "Integrated Surface-Groundwater Modeling in Wetlands with Improved Methods to
Simulate Vegetative Resistance to Flow," Ph.D. Dissertation, Florida International University, Miami,
FL
Ponce, V.M. and F.D. Theurer, 1982. Accuracy Criteria in Diffusion Routing, J. of Hyd. Eng., ASCE,
108(6), 747-757.
Ponce, S.M., 1989. Engineering Hydrology, Prentice Hall, Englewood Cliffs, New Jersey.
78
Appendix
U.S. Army Corps of Engineers Original Two Letters of Certification
Letters of Endorsement from:
Flood Control District of Maricopa County
Mohave County Flood Control District
Cochise County Flood Control District
Email Endorsements from:
California Department of Water Resources
U.S. Corps of Engineers, Sacramento District
Truck River Flood Project
List of FLO-2D Version 2009 Model Revisions
79
80
81
82
83
Letters of Endorsement
84
85
86
87
88
Email Endorsements
U. S. Army Corps of Engineers, Sacramento District
Senior engineering staff in the Hydraulic Design Section, Sacramento District, USACE, have
reviewed the document FLO-2D Model Validation for Version 2009 and Up, dated June 2011.
Based on our review of this document, as well as 15 years of successful application of FLO-2D
on numerous Sacramento District projects, we believe that FLO-2D v. 2009 is technically sound
for the purpose of floodplain mapping in support of FEMA Flood Insurance Studies.
Lea Adams, P.E.
Chief, Hydraulic Design Section
Sacramento District
US Army Corps of Engineers
1325 J St
Sacramento CA 95814
breaches of the Dams upstream of Reno-Sparks. These inundation maps are used by the
Regional Emergency Operations Center in running table top exercises of various flood
emergencies that could happen in the region. The flood project also monitors flood elevations
in critical flood pool areas to evaluate the adverse flood impacts that could happen if fill is
placed in these critical areas. Local ordinances require mitigation of floodplain storage volume
that may be displaced by a project.
Paul D. Urban, P.E. , Project Manager
Truckee River Flood Project
Reno, Nevada
90
91
92
93
17. For flow upstream in a channel an adjustment to the routing scheme was implemented to
increase the velocity. The coefficient for diffusive wave velocity was revised when the
momentum term shifted direction. (2/2/10).
18. The graphics display at runtime was fixed so that the interior channel elements display the
channel flow depth (2/2/10).
19. For channel routing, edited the channel read subroutine to stop when the end of the channel is
reach for displaying the interior channel elements (2/15/10).
20. Corrected the RGRIDTOT = 0 variable initialization for metric. This was already done for English
units (5/8/10).
21. Fixed a bug for metric mudflow with a channel. The stability criteria was being by-passed for low
flows based on 1.0 m instead of 1.0 ft in metric (6/5/10). This fix was released in BUILD No. 0910.06.01. The build number is listed at the end of the SUMMARY.OUT file and is being for the
first time with this bug fix.
22. The numerical solution of the momentum equation was expanded for the sign convention of
convective acceleration term for upstream channel flow in the case of tidal conditions
(7/14/10). BUILD No. 09-10.07.01.
23. The Courant number stability criteria which was previously hardwired and set to 1.0 was
redesigned to be an optional user variable as line 2 in TOLER.DAT. The default value is now 0.6
and the user can assign a value in the suggested range of 0.3 to 1.0 (7/14/10). BUILD No. 0910.07.01
24. Binary files of the output data are now generated for the last completed output interval
(multiple of TOUT in CONT.DAT). This means that the model can be restarted with the last
completed output interval for every simulation regardless of what cause the model to stop (i.e.
successful completion of the model, inadvertent computer termination, or user interrupted
event) (7/14/10). BUILD No. 09.10.07.01.
25. The HYCROSS.OUT had incorrect times listed when streets were included in the some but not all
of the prescribed floodplain cross sections (7/14/10). BUILD No. 09.10.07.01.
26. The infiltration porosity parameter POROS has been adjusted to accommodate the Flood
Control District of Maricopa County drainage manuals that specifies that the soil moisture
deficiency parameter DTHETA represent the volumetric soil moisture deficit which includes the
porosity. If POROS is assigned 0.0, then DTHETA includes the porosity and the DTHETA ranges
from 0.0 to 0.5. If DTHETA represents only the soil moisture deficit, set a reasonable soil
porosity in the range from 0.35 to 0.45 (11/15/10). BUILD No. 09-10.11.02.
27. The floodplain limiting Froude number is now spatially variable and is assigned in an optional
data file FPFROUDE.DAT. The values in this file will supersede the global assignment of the
limiting Froude number FROUDL in CONT.DAT. The format of the file is simply: F(line
character) Grid element number limiting Froude number. These values can be graphically
assigned in the GDS (11/15/10). BUILD No. 09-10.11.02
28. Fixed the binary file status (open) for the hydraulic structure data when continuing a previous
simulation (11/20/10). BUILD No. 09-10.11.02.
29. Expanded the hydraulic structure routine to allow multiple culverts to discharge to one culvert
(11/20/10). BUILD No. 09-10.11.02.
94
30. Corrected code for channel mudflow with hydraulic structures. This coded error was due to an
improper placement of an ENDIF statement that resulted in a fatal Fortran error that would not
allow the model to run (12/21/10). BUILD No. 09-10.12.02.
31. There was a metric conversion error for a Breach output parameter that would cause a fatal
error (1/24/11). BUILD No. 09-11.01.03.
32. Improved the depression storage initial routing for rainfall by decrementing the timestep for
negative flow depths instead of redistributing the flow. This was important for ARF reduced grid
elements (1/28/11). BUILD No. 09-11.01.03.
33. Fixed an array index for reading multiple inflow sediment supply size fraction groups (1/28/11).
BUILD No. 09-11.01.03.
34. Added code to identify and eliminate negative grid element storage when very small rainfall
begins to accumulate and runoff. This may occur when the flow depth first starts to exceed the
TOL value using large timesteps on grid elements with ARF reduced area. The volume
conservation error was minor (almost negligible), but the error was eliminated with this revision
(1/30/11). BUILD No. 09-11.01.03.
35. There was a bug that was introduced for writing channel hydraulic structure discharge to the
temporary file in the latest build. This did not affect any of the model computations, only the
writing of the structure discharge to the HYDROSTRUCT.OUT file. The discharge was reset to
zero before the final results were written to file (2/3/11). Build No. 09-11.02.03.
36. Writing the channel Froude number to the HYCHAN.OUT file was corrupted by failure to
initialize the variable prior to writing to file (2/7/11). Build No. 09-11.02.03.
37. In the channel variable regression relationships for computing the channel top width above the
assigned exceedance value, two errors were corrected. These errors were essentially reporting
errors that did not affect the actual flood routing because the top width is only used to assign
the right bank element (2/15/11). Build No. 09-11.02.03.
38. The run time of the model in SUMMARY.OUT was not being reported correctly. The runtime
was computed based on the CPU runtime. This was appropriate before parallel processors were
introduced. The runtime computation was revised to be based on the actual start and end of
the computation sequence in the model (2/21/11). Build No. 09-11.02.03 .
39. The floodway routine was improved for flow exchange between the floodway and floodway
fringe modifying the assignment for the floodway element as they filled with water (2/26/11).
Build No. 09-11.02.03.
40. The Karim Kennedy equation had a metric conversion error for the sediment D50 size (3/9/11).
Build No. 09-11.03.03.
41. The RAINCELL spatially variable rainfall assignment was re-initialized to enable the grid element
data to appear in random order in the RAINCELL.DAT file (4/6/11). Build No. 09-11.04.04.
42. The TIMETOPEAK output file was triggered by the BREACH time and if there was no dam or
levee breach simulation, the time was reported incorrectly (5/3/11). Build No. 09-11.05.04.
43. Expanded the cross section write format in the BASE.OUT file for larger discharges (5/4/11).
Build No. 09-11.05.04.
95
Revisions and bug fixes in the Processor Programs include (most recent at the bottom):
1. MAXPLOT The Version 2007 executable should be replaced with the 2009 in the FLO-2D folder.
The Version 2007 executable was accidentally installed in the model subdirectory. (10/25/2009).
2. GDS Removed the C:\TEMP\FLO-2D folder address. This is no longer needed (11/6/09).
3. MAPPER Rename the Mapper_2009.CHM to Manual_Mapper_2009.CHM so the Mapper
program can read the help files (11\6\09).
4. FLOENVIR Fixed the graphical display of the interior channel elements for more than one
channel segment (11/14/09).
5. MAXPLOT Fixed the graphical display of the interior channel elements for more than one
channel segment (11/14/09).
6. GDS Added warning messages for possible missing outflow nodes and for the channel
interpolation cross sections (11/18/09).
7. GDS Fixed the Breach Switch in the Levee Editor dialog box. The switch was required to be
turned on to save the levee data files even though a breach was not being simulated
(12/02/09).
8. To import AutoCAD DXF files into the GDS, the MO24rt.exe file may need to be reinstalled with
new control parameters: (/CFGHIJKM). This can be accomplished manually by performing the
following:
In Windows Explorer, locate the file C:\Program Files\FLO-2D\Mo24rt.exe. In the DOS command
window or from the Run command, execute the Mo24rt.exe file as follows:
C:\Program Files\FLO-2D\Mo24rt.exe/CFGHIJKM
9.
10.
11.
12.
13.
This will create all the files in the C:\Program Files\Common\ESRI folder (12/2/09).
MAXPLOT Fixed the channel flow depth display of non-bank interior elements (12/22/09).
FLO2DINTER.DLL Update for the GDS DLL library. Replace this file in the Windows\System32
folder (1/7/10).
FLOENVIR The display of the channel interior elements was fixed for channels with more than
one segment (2/2/10).
Mapper Some minor display changes and typos were fixed (2/2/10).
GDS A number of bug fixes and enhancements were made including (2/2/10):
New warning message for channel cross section interpolation.
Eliminated the GDS writing zeros in the INFLOW.DAT file when no hydrographs are
entered.
New warning message if RAIN.DAT is missing for RAIN.EXE and updates rainfall data
when all the *.DAT files are saved.
Deleted NOFLOCS outside the channel after channel realignment.
Levee duplicate assignments to the same boundary warning message.
Fixed the channel cross section plots to join the station lines.
Spelling errors were corrected.
14. GDS Setting the Total Rainfall in the RAIN dialog box was fixed so that the enter value was save
when assigning the rainfall distribution (5/8/10).
15. GDS Saving the global infiltration was corrected. The infiltration file INFIL.DAT was being
created without user initiation. Now INFILT.DAT is only generated when the user assigns either
96
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
global or individual parameters. The default global parameters are automatically assigned when
the user double clicks on an individual grid element to assign infiltration (5/8/10).
PROFILES The interpolation routine was adjusted for metric cross sections (6/15/10).
GDS A number of enhancements were incorporated for mega DTM data files including
(7/14/10):
Reading very large mega DTM data files exceeding 15 million points.
Reading new ASCII format DTM data files.
Reading multiple DTM data elevations files.
Null assignment of grid elements without DTM data.
Options for interpolation grid elements with null elevation assignment.
GDS Levee crest elevation can now be assigned from 3-D polylines data (7/14/10).
GDS Revised to read, assign and write optional Courant number in the TOLER.DAT (7/14/10).
GDS Expanded the HEC-RAS channel data interface extensively to enable geo-reference HECRAS channels to be automatically aligned an assigned cross sections.
rb_exe.exe channel right bank element assignment increased channel element numbers to 8
places, download and put this file in the FLO-2D folder (8/2/10).
rbcheck.exe channel right bank element check write to file, increased channel element
numbers to 8 places, download and put this file in the FLO-2D folder (8/2/10).
GDS Floodplain limiting Froude numbers are now spatially variable and can be assigned in the
GDS. The GDS was also modified to enable user defined Courant numbers (8/8/10).
PROFILES and HYDROG have been updated to handle grid systems greater than one million grid
elements (8/8/10).
GDS right bank element assignment rb_exe.exe program. The number of right bank extension
was improved for rectangular and trapezoidal channels. Download and put this file in the FLO2D folder (2/7/11).
HYDROG Expanded the cross section column width to handle grid element numbers greater
than 1,000,000 (2/14/11).
MAXPLOT Improved the file compare routine for plotting the differences between two output
files. Both graphical and computation improvements were made (2/15/11).
HYDROG Fixed a read format error introduced by a line change in the SUMMARY.OUT file
(3/2/11).
GDS Fixed an error in the levee profile tool (4/20/11).
FLOENVIR and MAXPLOT Fixed the levee plot graphics for flow directions for the case where
there was more than 200 consecutive grid elements in the first row or column that all had the
same coordinate value of either x or y (4/25/11).
GDS -- Fixed an error in the street realign tool. Default spatial street width was assigned a 1
instead of a 0. (5/2/11)
Revisions and bug fixes in the GUI program include (most recent at the bottom):
1. The version number was correct to match the FLO-2D model version number 2009.06 (5/8/10).
97