HEC-RAS 2D Sediment Users Manual (Beta) PDF

HEC-RAS Two-Dimensional
Sediment Transport User’s

Manual
December 2020
Approved for Public Release. Distribution Unlimited. RD-??
REPORT DOCUMENTATION PAGE Form Approved OMB No. 0704-

0188
The public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing
instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of
information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing
this burden, to the Department of Defense, Executive Services and Communications Directorate (0704-0188). Respondents should be aware that
notwithstanding any other provision of law, no person shall be subject to any penalty for failing to comply with a collection of information if it does
not display a currently valid OMB control number.
PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ORGANIZATION.
1. REPORT DATE (DD-MM-YYYY) 2. REPORT TYPE 3. DATES COVERED (From - To)
December 2020 Research Document
4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER
HEC-RAS Two-Dimensional Sediment Transport 5b. GRANT NUMBER
Users Manual
5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S) 5d. PROJECT NUMBER
Alejandro Sanchez, CEIWR-HHT 5e. TASK NUMBER
5F. WORK UNIT NUMBER

Stanford Gibson, CEIWR-HHT
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION REPORT
NUMBER
U.S .Army Corps of Engineers
Institute for Water Resources RD-??
Hydrologic Engineering Center (CEIWR-HEC)
609 Second Street
Davis, CA 95616-4687
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/ MONITOR'S ACRONYM(S)
11. SPONSOR/ MONITOR'S REPORT

NUMBER(S)
12. DISTRIBUTION / AVAILABILITY STATEMENT
Approved for public release; distribution is unlimited.
13. SUPPLEMENTARY NOTES
14. ABSTRACT
15. SUBJECT TERMS
HEC-RAS; River Analysis System; HEC; Hydrologic Engineering Center; two-dimensional;

R&D; research & development; United Kingdom; research project; models; hydraulic;
benchmarking; test; grid cell sizes; grid; bathymetry; cell face; terrain; elevation; flow
solver; time steps; result; water surface; velocities; computational interval
16. SECURITY CLASSIFICATION OF: 17. LIMITATION 18. NUMBER 19a. NAME OF
RESPONSIBLE
PERSON
a. REPORT b. ABSTRACT c. THIS PAGE OF OF
ABSTRACT PAGES
U U UU
UU 116 19b. TELEPHONE
NUMBER
Standard Form 298 (Rev. 8/98)

Prescribed by ANSI Std. Z39-18
HEC-RAS
Two-Dimensional Sediment
Transport Technical Reference
Manual
December 2020
U.S. Army Corps of Engineers
Institute for Water Resources
Hydrologic Engineering Center
609 Second Street
Davis, CA 95616
(530) 756-1104
(530) 756-8250 FAX
www.hec.usace.army.mil
RD-??
RD-?? Table of Contents
Table of Contents
Table of Contents
Table of Contents ............................................................................................ i
List of Figures................................................................................................ iii
List of Tables ................................................................................................. ix
Chapter 1 ..................................................................................................... 1
Introduction .................................................................................................. 1
Chapter 2 ..................................................................................................... 2
Performing a 2D Sediment Transport Analysis .................................................... 2
2.1 Best Practices for a 2D Sediment Model ........................................... 2
2.1.1 Hydraulic Warm Up ......................................................................... 2
2.1.2 Computational Mesh ....................................................................... 3
2.1.3 Hydraulic Equations ........................................................................ 7
2.1.4 Time Step ..................................................................................... 8
2.2 Sediment Data ............................................................................. 9
2.2.1 Initial Conditions and Transport Parameters ......................................10
2.2.2 Associated Bed Gradation Templates with Bed Material Layers .............18
2.2.3 Sediment Boundary Conditions ........................................................19
2.2.4 User-Defined Grain Classes .............................................................23
2.2.5 Cohesive Options ..........................................................................26
2.2.6 Transport Methods.........................................................................30
2.2.7 Transport Function Calibration and Modification .................................37
2.2.8 2D Options ...................................................................................38
2.2.9 Bed Mixing Options ........................................................................42
i
Table of Contents RD-??
2.2.10 Bed Gradations .........................................................................43
2.3 Viewing Sediment Results in RAS Mapper ....................................... 46
2.3.1 Plotting 2D Sediment Time Series ....................................................51
2.3.2 Plotting 2D Sediment Transects/Profile Lines .....................................52
2.4 Sediment Computation Options and Tolerances ............................... 54
2.4.1 Sediment Computation Options and Tollerances .................................55
2.5 Sediment Output Options ............................................................. 67
2.5.1 Output Level .................................................................................69
2.5.2 Output Interval Multiples ................................................................71
2.5.3 Customized Sediment Output Variables ............................................71
2.5.4 Sediment Hotstart .........................................................................75
2.6 Viewing Results........................................................................... 75
2.6.1 Exploring the HDF5 File ..................................................................75
2.6.2 Log Output ...................................................................................76
Chapter 3 .................................................................................................... 79
References ................................................................................................... 79
ii
RD-51 List of Figures
List of Figures
Page
No.
Figure 2-1. HEC-RAS Unsteady Computation Options and Tolerances editor. --------- 3
Figure 2-2. Example of a computational mesh with a simple refinement region within
the channel. ----------------------------------------------------------------- 4
Figure 2-2. Example computational mesh with cells in the channel aligned with the
direction of flow. ------------------------------------------------------------ 4
Figure 2-2. Example computational mesh with small faces indicated by ellipses. ----- 5
Figure 2-2. Mesh Tolerances section of the RAS Mapper Options window. ------------ 6
Figure 2-2. Example computational meshes with Minimum Face Length Tolerances of
5% (red) and 20% (black). ------------------------------------------------ 7
Figure 2-2. Example of an erroneous bed change result for a river using the Diffusion-
Wave Equation. ------------------------------------------------------------- 8
Figure 2-2. Sediment Data editor. ------------------------------------------------------- 9
Figure 2-3. Bed Gradation Template editor. ------------------------------------------- 11
Figure 2-4. Crating a Bed Material Layer in RASMapper. ----------------------------- 12
Figure 2-5. Sediment Bed Material Layer underlying a 2D mesh and in the RASMapper
tree under Map Layers. --------------------------------------------------- 13
Figure 2-6. Give the layer classifications names that will show up in the sediment editor
under Layer Properties. --------------------------------------------------- 14
Figure 2-7. RAS Mapper Manage Layer Associations editor. ---------------------- 15
Figure 2-8. Bed Gradation Template editor. ------------------------------------------- 16
Figure 2-9. Example of a Bed Layer Group definition in the Define Gradation Layers
editor. --------------------------------------------------------------------- 17
Figure 2-10. Example of a Bed Layer Group assigned to a computational cell. ------ 17
Figure 2-11. Simple, single gradation, 2D sediment specification with one Bed
Gradation associated with one Bed Material Type. ---------------------- 18
iii
List of Figures RD-51
Figure 2-12. More detailed bed material specification, where the Bed Material Layer
was imported from a shape file. ----------------------------------------- 19
Figure 2-13. Boundary Conditions tab in the Sediment Data editor. ----------------- 20
Figure 2-14. Sediment Rating Curve editor. ------------------------------------------- 20
Figure 2-15. Sediment Load Series editor. -------------------------------------------- 21
Figure 2-16. Opening the Unsteady Temperature editor from the Unsteady Flow Data
editor. --------------------------------------------------------------------- 22
Figure 2-17. Specifying a temperature time series. ----------------------------------- 23
Figure 2-18. Opening the User Defined Grain Classes editor from the Sediment Data
editor. --------------------------------------------------------------------- 24
Figure 2-19. User Defined Grain Classes editor.--------------------------------------- 24
Figure 2-20. Selecting the Density Method in the Define Grain Classes and Sediment
Properties editor. --------------------------------------------------------- 25
Figure 2-21. Opening the Cohesive Options editor from the Sediment Data editor. 27
Figure 2-22. Cohesive Options editor. ------------------------------------------------- 27
Figure 2-23. Selecting the Hwang (1989) floc settling velocity formula.------------- 28
Figure 2-24. Specifying a floc settling velocity as a function of suspended sediment

concentration. ------------------------------------------------------------- 29
Figure 2-25. Example consolidation curve of dry bulk density as a function of time. 30
Figure 2-26. Opening the Transport Methods editor from the Sediment Data editor. 31
Figure 2-27. Schematic of sediment and current velocity profiles. ------------------- 31
Figure 2-28. Setting the Total-load correction factor options in the Transport Model
editor. --------------------------------------------------------------------- 32
Figure 2-29. Setting the bed-load correction factor options in the Transport Model
editor. --------------------------------------------------------------------- 33
Figure 2-30. Setting the suspended-load correction factor method in the Transport
Model editor. -------------------------------------------------------------- 33
Figure 2-31. Specifying a constant total-load diffusion coefficient. ------------------ 34
Figure 2-32. Specifying a total-load diffusion coefficient based on weighted bed and
suspended load coefficients. --------------------------------------------- 35
Figure 2-33. Specifying a total-load adaptation length.------------------------------- 35
iv
Figure 2-34. Specifying a weighted total-load adaptation length. -------------------- 36
Figure 2-35. Specifying a suspended-load adaptation coefficient. ------------------- 36
Figure 2-36. Opening the Transport Function Calibration and Modification editor from
the Sediment Data editor. ------------------------------------------------ 37
Figure 2-37. Transport Function Calibration and Modification editor. ---------------- 38
Figure 2-38. Opening the 2D Options editor from the Sediment Data editor. ------- 39
Figure 2-39. 2D Options editor. -------------------------------------------------------- 39
Figure 2-40. Setting the hindered settling method to the Richardson and Zaki (1954)
method. ------------------------------------------------------------------- 41
Figure 2-41. Accessing the Bed Mixing Options editor from the Sediment Data editor.42
Figure 2-42. Bed Mixing Options editor. ----------------------------------------------- 42
Figure 2-43. Bed Gradation editor with an example grain size distribution. --------- 44
Figure 2-44. Setting sample specific cohesive parameters in the Bed Gradation editor.
----------------------------------------------------------------------------- 45
Figure 2-45. Right click on the Plan Description under Results and Select Create a New
Results Map Layer to add Sediment results (or non-default hydraulic
results). ------------------------------------------------------------------- 46
Figure 2-46. Two types of sediment results (Bed and Transport) in the Results Map
Parameter editor. --------------------------------------------------------- 47
Figure 2-47. Example use of the animation bar to view sediment concentrations in
time. ----------------------------------------------------------------------- 48
Figure 2-48. To edit the color ramp and display properties of the seidment maps, right
click on the map and select Layer Properties. Then Click Edit under the
Surface menu. ----------------------------------------------------------- 49
Figure 2-49. Setting the Max and Min to equal values (whichever abolute value is
larger) with opposite signs, centers the Bed Change plot, making all
depsoitin and erosion the same colors and the range of no change white.
----------------------------------------------------------------------------- 50
Figure 2-50. Save customized sediment color ramps that persist throughtou the project
with the Save Color Ramp button. ------------------------------------- 51
Figure 2-51. View a time series of a sediment (or hydraulic) result by selecting
(chekcing) it in the RAS Mapper result tree and then right clicking on the
cell. Select Plot Time Series and then the result map to launch the time
series for that cell or cell face. ------------------------------------------- 52
v
List of Figures RD-51
Figure 2-52. Right Click on the Feature menu to create a Polyline Layer. Give it a
name.---------------------------------------------------------------------- 53
Figure 2-53. Draw the polyline layer along the 2D sediment results transect. ------ 53
Figure 2-54. Plot a bed change transect with a polyline profile plot. ----------------- 54
Figure 2-55. Bed Gradation editor with an example grain size distribution. --------- 55
Figure 2-56. General tab of the Sediment Computation Options and Tolerances editor.
----------------------------------------------------------------------------- 56
Figure 2-57. Specifying the warmup periods for 2D simulations. -------------------- 57
Figure 2-58. Specifying the warmup method for 2D simulations. -------------------- 59
Figure 2-59. Outer loop Convergence Parameters.------------------------------------ 64
Figure 2-60. Calculation of computational bed layer thickness from a single user-
specified initial bed layer or bed gradation. ----------------------------- 65
Figure 2-61. Calculation of computational bed layer thickness and composition from
user-specified initial bed layers. ----------------------------------------- 65
Figure 2-62. Computational Sediment Layer Parameters section in the HEC-RAS

Sediment Computation Options and Tolerances editor.----------------- 66
Figure 2-63. Example of a power-law parameterization of the critical shear for erosion
as a function of the dry bulk density. ------------------------------------ 67
Figure 2-64. Opening the Sediment Output Options editor from the Unsteady Flow
Analysis editor.------------------------------------------------------------ 68
Figure 2-65. Sediment Output Options editor. ---------------------------------------- 69
Figure 2-66. Opening the User Defined Grain Classes editor from the Sediment Data
editor. --------------------------------------------------------------------- 70
vi
vii
RD-51 List of Tables
List of Tables
Table Page
Number No.
ix
List of Tables RD-51
x
RD-51 Chapter 1 – Introduction
Chapter 1
Introduction
The Corp’s Hydrologic Engineering Center River Analysis System (HEC-RAS) is designed to
simulate one-dimensional (1D) steady, unsteady flow. The latest release of HEC-RAS V6.0
also simulates unsteady two-dimensional horizontal (2D) sediment transport, and bed
change, sorting, and layering. Sediment transport is computed with a non-equilibrium total-
load formulation. The total-load transport equation is solved with implicit Finite-Volume
methods on the same unstructured polygonal mesh as the flow solver. Sediment transport
is coupled to the flow model at the time step level. One powerful feature of the 2D flow
solvers is that they use the subgrid topographic variations directly into the model thus
improving the accuracy of the solution and permitting the use of relatively coarse meshes
resulting in reduced computational times. The sediment transport model is designed to work
within the subgrid framework of the flow model, and computes subgrid erosion and
deposition rates, bed elevations, gradations, and bed layering.
This document discusses how to utilize the 2D sediment user-interface, the model input and
output, and how-to setup and run a 2D sediment transport model in HEC-RAS. The
document is intended as supplemental to the 1D Sediment Users Manual as many concepts
and features are covered in detail in that document. Most of the 1D sediment capabilities
are supported in 2D sediment and many new features have been added to 2D sediment
which are not available yet in 1D sediment. Some of the new sediment features include
variable density bed sorting and layering model, flocculation, consolidation, hiding and
exposure effects, multiple new transport potential formula. However, as a beta release,
there are still several computational and user-interface limitations and known issues
including the inability to hot-start sediment, the inability to couple 1D and 2D sediment, the
inability to modify terrains based on computed bed change, inability to visualize subgrid
output directly in HEC-RAS, the inability to specify avalanching parameters in the user-
interface, and the inability to specify subsidence in 2D areas. Lastly, the 2D sediment
transport feature in HEC-RAS V6.0 is a beta feature and should not be used for design
purposes.
1
Chapter 2 – Performing a 2D Sediment Transport Analysis RD-??
Chapter 2
Performing a 2D Sediment
Transport Analysis
2.1 Best Practices for a 2D Sediment Model

Sediment models are often more sensitive to hydraulic modeling choices that fixed
bed models. Adding sediment to a hydraulic model that runs and generates
reasonable results often exposes model problems hidden by the robustness of the
hydraulic equations or approximations that were sufficient for hydraulic analysis but
not for sediment modeling.
2.1.1 Hydraulic Warm Up

One of the most important hydraulic parameters for sediment modeling is the
hydraulic warm-up period. Unless the prototype starts dry and pushes a wetting
front through the domain, it is essential to use a hydraulic warm up period for the a
2D sediment model. A hydraulic model can often pass through a wetting phase at the
beginning of a simulation without affecting the final results (this is not best practice
but can produce acceptable results in some cases). If the sediment model runs
without a warmup period, it will scour as the wetting front initialized the model and
bias the result.
HEC-RAS will only run hydraulics during the first half of the warmup period, and then
will add sediment warm up during the second half. Choose a warmup period that
wets the mesh fully during the first half. HEC-RAS does not populate a warmup
period by default. Define this parameter by selecting OptionsComputational
Options and Tolerances… and selecting the 2D Flow Options tab.
Define the Initial Conditions Time (hrs) and the Ramp Up Fraction (the portion
of the Initial Conditions Time the model will take to gradually get up to the full warm
up flows – often 0.1). Run the hydraulic model (without sediment) to make sure this
time is sufficient to reach an initial steady-state flow condition.
2
RD-?? Chapter 3 – Viewing 2D Sediment Results
Figure 2-1. HEC-RAS Unsteady Computation Options and Tolerances editor.
2.1.2 Computational Mesh

Although HEC-RAS can compute on highly irregular meshes, poor mesh quality will
decrease the accuracy of the numerical solution and can lead to poor convergence
and numerical instabilities. Mesh quality affects both the flow and sediment, but it is
especially important when simulating sediment transport.
Mesh Alighnment
With 2D flow and sediment transport, the HEC-RAS will produce better results if the
mesh is oriented or aligned with the flow. This reduces numerical diffusion and
improved computational accuracy. In the example below, a refinement region is used
to increase the spatial resolution (25 ft) within the channel. The refinement region
allows for a more resolution within the channel while maintaining coarse resolution
outside of the channel, thus reducing the total number of computational cells and
computational costs. The refinement region also aligns the channel banks which also
improves the model results. However, the issue with this approach is the that the
cells are not aligned with the flow producing more numerical diffusion. In addition,
the regions in between the square or rectangular cells within the refinement polygon
and the boundaries of the refinement polygon can have relatively irregular cells and
relatively poor mesh quality.
3
Figure 2-2. Example of a computational mesh with a simple refinement region within the
channel.
To improve the mesh alignment with the flow, a break line is inserted along the
centerline of the channel as shown in the figure below. The breakline is added with 6
repeat points and a 30-ft spacing.
Figure 2-3. Example computational mesh with cells in the channel aligned with the direction
of flow.
4
Small Faces
When building a mesh avoid having cells with small faces. This is because small faces
can increase run times and introduce instabilities in both the hydraulic and sediment
model. Small faces increase the number of faces and therefore the computational
time and memory. In addition, small faces often have poor mesh quality.
Figure 2-4. Example computational mesh with small faces indicated by ellipses.
HEC-RAS has a feature which detects small faces and snaps together the face nodes
(also referred to as face points in HEC-RAS). The tolerance parameter for the
maximum face length is based on the ratio between the face length and the distance
between adjacent computational points. The tolerance parameter can be modified by
the user in RAS Mapper by
1. Select the menu Tools.
2. Click on Options…. This will open The RAS Mapper Options window
5
3. On the left side under Project Settings, select Mesh Tolerances.
4. Modify the value next to Minimum Face Length Tolerance (Percent) by
manuall typing a value or using the above and down arrows.
5. Click Apply
6. Click OK to the window
7. If necessary recompute 2D meshes.
Figure 2-5. Mesh Tolerances section of the RAS Mapper Options window.
An example two meshes with Minimum Face Length Tolerances of 5% (red) and 20%
(black) are shown in the figure below. In general the differences between the two
meshes occur at transition areas between resolution and near breaklines.
6
Figure 2-6. Example computational meshes with Minimum Face Length Tolerances of 5%
(red) and 20% (black).
2.1.3 Hydraulic Equations

2D sediment runtimes can be be long, so it can be attractive to select the Diffusion
Wave Equation (DWE) to try to reduce the runtime. However, with the DWE
sediment models of rivers often develop a distinctive pattern, eroding a deep sub-
channel and depositing in the rest of the channel (see figure below). This is not a
credible result. Most 2D river sediment transport models should be run with one of
the Shallow Water Equations (SWE). The DWE should only be used for watershed or
flooding type applications.
7
Figure 2-7. Example of an erroneous bed change result for a river using the Diffusion-Wave
Equation.
2.1.4 Time Step

The time step is the most common issue with 2D sediment models that do not run.
Modelers usually try to run the 2D sediment model with time steps that are too
large. The model setup should obey (or at least respect*) the Courant condition,
which roughly means that the water should pass though one cell per time step.
Users can calculate a limiting Courant condition for their model using the smallest
cell and estimating a maximum current velocity and then choose a fixed time step
that satisfies this condition. However, recent versions of HEC-RAS make time step
selection much easier.
The Advanced Time Step Control under the Unsteady Computation Options and
Tolerances includes options for dynamic time steps that will change during the
simulation. The option to Adjust Time Step Based on Courant is the most widely
used and is becoming standard practice for 2D modeling. Define a maximum and
minimum Courant condition and then the maximum halving or doubling steps
allowed from the base time step, and the model will compute the appropriate time
step throughout the model.
The Implicit Finite Volume solver in HEC-RAS 1D and 2D is not Courant limited. So
the Courant Condition is more of a stability guideline and time-step selection
support, than the hard limit it can be in other, explicit, solvers. Therefore, modelers
sometimes set their max Courant condition closer to 2, without introducing
computational issues or instabilities.
8
2.2 Sediment Data

2D sediment transport models in HEC-RAS require four files: an unsteady flow file, a
geometry file, a sediment file, and a plan file to tie them together. Sediment data
should be specified after the geometry. Some mandatory sediment parameters are
specified spatially and thus require knowledge of the geometry. If the geometry
changes, the sediment data may need updates before it can run.
Sediment results are very sensitive to hydraulic results (e.g. roughness parameters
and ineffective flow areas). An HEC-RAS sediment model will not be robust or
reliable if it is not built on a well-constructed hydraulic geometry (USACE, 1993,
Thomas and Cheng, 2008). Before adding sediment data, test the hydraulic model
over the expected range of flows using the HEC-RAS Unsteady Flow Analysis option,
calibrating the bottom roughness and if necessary, the turbulence coefficients. It is
recommended to identify hydraulics model problems before adding the bed change
complexity. Only add sediment data after crafting a careful, robust, hydraulic model.
This document focuses on the 2D sediment data. However, many of the 2D sediment
parameters and data inputs are the same as 1D. To enter sediment data, select
Sediment Data from the Edit menu or press the sediment data icon. The sediment
data editor will appear as depicted in Figure 2-8. The sediment data editor has four
tabs: Initial Conditions and Transport Parameters, Sediment Boundary
Conditions, the USDA-ARS Bank Stability and Toe Erosion Model (BSTEM),
and 2D Sediment (Beta). The first two and last tabs are mandatory for 2D
simulations. The third (BSTEM) is only for bank process computations and is not
required for a sediment transport model.
Figure 2-8. Sediment Data editor.
9
2.2.1 Initial Conditions and Transport Parameters

The Initial Conditions and Transport Parameters is the first tab in the Sediment Data
editor and opens by default when the editor launches. From this editor the user can specify
the transport function, sorting method, fall velocity method for the entire model. It is also
used to specify sediment data for 1D cross-sections. However, if the model does not include
cross-sections, then that information may be left empty.
Transport Function
Select a Transport Function from the drop-down box near the top of the editor. HEC-RAS
6.0 includes the following eleven transport functions:
1. Ackers and White (Ackers and White 1973; Day 1980; Proffitt and Sutherland 1983)
2. England and Hansen (Engelund-Hansen 1967)
3. Laursen-Copeland formula (Laursen 1968)
4. Meyer-Peter and Müller (1948)
5. Toffaleti (1968)
6. MPM-Toffaleti (Meyer-Peter and Müller 1948; Toffaleti 1968)
7. Yang (sand and gravel eqns.)
8. Wilcock and Crowe (2003)
9. Soulsby-van Rijn (Soulsby 1997)
10. van Rijn (1984a,b; 2007a,b)
11. Wu et al. (2000)
Modeling Note – Transport Function Sensitivity: Sediment transport functions simulate non-
linear transport processes and produce very different results. Model results are very
sensitive to selected function. Carefully review the range of assumptions, hydraulic
conditions, and grain sizes for which each method was developed. Select the method
developed under conditions that most closely represent the system of interest and calibrate
results to actual bed change.
Sorting Method
Transport functions compute transport potential without accounting for availability. The bed
Sorting Method (sometimes called the mixing or armoring method) keeps track of the bed
gradation which HEC-RAS uses to compute grain-class specific transport capacities and can
also simulate armoring processes which regulate supply. The Sorting Method drop down
menu in the Sediment Data editor only applies to 1D sediment transport. The 2D sediment
transport model always uses a method similar to the Active Layer method for 1D except
that it can have any arbitrary number of layers. The active layer thickness is set equal to
the d90 by default or a factor of the d90.
Modeling Note – Mixing Method Sensitivity: Sediment transport results can be as sensitive
to the mixing method selected as the transport function.
10
Fall Velocity Methods

The same sediment fall velocity formulas are available for 2D and 1D sediment transport.
The options include:
1. Rubey (1933)
2. Toffaleti (1968)
3. Van Rijn (1993)
4. Report 12 (Default method in HEC-6)
5. Dietrich (1982)
6. Soulsby (1997)
7. Wu and Wang (2006)
The fall velocity formula used to compute free particle settling velocity for both cohesive and
noncohesive sediments. Depending on the concentration and whether the user has selected
flocculation, cohesive sediments may also settle as flocs. Hindered settling of noncohesive
particles may also be simulated if selected by the user. In general, the fall velocity formula
should not be used as a model calibration parameter for morphology change since the
results are not very sensitive to the formula. Preference for the formulas is generally based
on the transport formula used, the range of grain classes, and/or the parameters utilized by
the fall velocity formulas. For example, the Wu and Wang formula considers the particle
shape.
Bed Gradations
Bed gradations are specified in the same manner as for 1D models. Instead of requiring
users to input bed gradations for each cell individually, HEC-RAS uses a Bed Gradation
Template concept similar to that in the Channel Modification Editor. Sediment Bed
Material Types are defined in a Sediment Bed Material Layer in RAS Mapper. The
Sediment Bed Material Types are regions defined as polygons in RAS Mapper in a
Sediment Bed Material Layer. The polygons can be overlapping to override regions.
Sediment Bed Material Layer Types are the associated to Bed Gradation Templates,
Non-erodible Surfaces, or Bed Layer Groups within the 2D Bed Gradations (Beta) tab
of the Sediment Data editor. Bed Layer Groups consist of one or more bed layers. Each
layer is assigned a Bed Gradation Template or a Non-erodible Surface. The
connectivity between these elements utilized for specifying the sediment bed material are
described in the schematic below.
Figure 2-9. Bed Gradation Template editor.
11
Non-erodible surfaces are surfaces such as bedrock or structures which cannot be eroded.
Non-erodible surfaces may be associated with a Sediment Bed Material are specified
within a Bed Layer Group. Non-erodible surfaces are specified at computational cells and
are not enforces at computational faces.
Defining Bed Material Layers in RAS Mapper

Users will associate bed gradations, layer groups, or non-erodible surfaces with Sediment
Bed Material Layers defined in RASMapper. In order to define gradations in the sediment
editor, the user must define Sediment Bed Material Layers in RAS Mapper first.
Sediment Bed Material Layers are independent of specific model geometries or terrains.
Like n-value Layers or Land Cover Layers (see 2D Hydraulic Manuals). Create new
Sediment Bed Material Layers by right clicking on the Map Layer node in RASMapper.
To Define Bed Material Layers, follow these five steps:
1. Create a New/Empty RAS Layer

2. Import or Draw Polygons in the Created RAS Layer
3. Give the Polygon Classifications Sediment Material Names that will show up in the
Sediment Editor
4. Go to Manage Geometry Associations and associate the Bed Material Layers with
geometry files
5. Associate each Bed Material Classification with a bed gradation, bed layers, or define
it as a non-erodible surface, in the 2D sediment editor.
Defining these bed material layers have two main workflows, which are not mutually
exclusive. Users will generally import a pre-existing shape file or draw overlapping
polygons. Often users will do both, importing a base bed material shape file, but then
drawing in channel features, sand bars, and/or non-erodible features on top of them as new
layers.
Figure 2-10. Crating a Bed Material Layer in RASMapper.
12
The Sediment Bed Material Layer will show up under the Map Layers tree in RASMapper.
Figure 2-11. Sediment Bed Material Layer underlying a 2D mesh and in the RASMapper tree
under Map Layers.
Whether you create your Sediment Bed Material Layer by importing a shape file, creating
polygons, or both, there are two additional steps before the bed materials are ready for the
sediment model.
RAS Mapper will create a Classification Polygons node under your map layer and will add
a feature to the geo-referenced database for each polygon you import or create. You will
have to give each of these polygons a (preferably descriptive) name which will show up in
the sediment editor, where you will associate the polygon(s) that share this classification
with a Bed Material or Bed Layer Group.
Name the layer classifications by right clicking on the Classification Polygons node (in
editable mode ) and selecting Layer Properties. Then select the Features tab and give
each of the Layer Classifications a Name that will show up in the Sediment Data editor.
13
Figure 2-12. Give the layer classifications names that will show up in the sediment editor
under Layer Properties.
When the Sediment Bed Material Layer and Classifications are complete, Stop Editing
and save.
Before these layers and classifications become available in the Sediment Data editor,
however, they must be associated with a geometry file. Just like n-values, Land Cover, and
Terrains must be associated with a geometry file, Sediment Bed Material Layer Groups
must be matched to one-or-more geometries to connect the Mapper classifications to the
other files in HEC-RAS.
Right Click on the Geometries node in RAS Mapper and Select Manage Geometry
Associations…
Select the Sediment Bed Material Layer that goes with each Geometry.
14
Figure 2-13. RAS Mapper Manage Layer Associations editor.
It is worth noting, that this associates the Bed Gradation Layers with the geometry, not the
sediment file. Sediment files depend on geometry files for some of their data structures
(e.g. cross sections in 1D). The sediment file will inherit the bed gradation layers and
classifications from the active geometry file.
These layer classifications will become available in the 2D Sediment tab of the Sediment
Data Editor (see the 2D gradation selection section).
Bed Gradation Templates

Bed Gradation Templates contain the sediment grain class sizes, grain fractions by
weight, and optionally bulk cohesive parameters. Users define sediment Bed Gradations
Templates in a database with no spatial data. They can be considered simply as a database
of different sediment types. In many applications, Bed Gradation Templates will
correspond to individual bed samples taken in the project. Templates are created and edited
by pressing the Define/Edit Bed Gradation button, which will launch the dialog depicted
in Figure 17-17.
15
Figure 2-14. Bed Gradation Template editor.
Multiple bed gradations may be entered for 2D sediment models in the same as for 1D
sediment models.
To create a Bed Gradation Template, first create a new bed gradation template by
selecting the New Bed Gradation Sample button: and entering a name for the sample.
(Alternatively, data for several different samples can be entered at once by clicking on the
button and entering the data in the Multiple Bed
Gradation Table.
The gradation of the bed sample can be input in either of two forms by toggling between the
radio buttons at the bottom of the form:
• % Finer: % Finer defines the sample using as a cumulative bed gradation curve
with percent finer defined by the upper bound of each grain class. The diameter listed for
each grain class is the upper bound of that grain class and values should be entered as
percent values. (e.g. since this is specified in Percent Finer, 50% should be input as 50 and
NOT as 0.5)
• Grain Class Fraction/Weight: the sample fraction of each grain class is specified.
These values will be normalized so values do not have to add up to one or 100% and can be
input as simple masses if preferred. (e.g. if 20% of the sample is fine sand, input the value
0.2 or 20 as long as the rest follow that convention). The upper and lower bound grain
diameter is associated with each grain class to delineate the range of the class.
16
Bed Layer Groups

User specified Bed Layer Groups allow the user to specify vertical bed layers with different
bed gradations and different bulk properties. Bed Layer Groups are specified by selecting
the button in the main Sediment Data editor which will open the Define
Gradation Layers editor (Figure 2-15).
Figure 2-15. Example of a Bed Layer Group definition in the Define Gradation Layers editor.
In HEC-RAS 2D sediment transport, the depositional layer thickness is not utilized. The
parameter is only utilized for 1D sediment transport.
The layer thickness specified for non-erodible surfaces is not actually used. Bed gradations
are specified for each computational cell. Faces do not have bed gradations and only bed
elevations. All subareas within a cell have the same Sediment Bed Material. Therefore, if
a Bed Layer Groups is assigned to a cell, the bed layers will be at different elevations for
the different subareas. For example, the figure below shows a computational which has
been assigned a Bed Layer Group three layers. In the example below, the cell has three
subareas. The bed layers have the same thickness in each subarea but start at a different
elevation.
Figure 2-16. Example of a Bed Layer Group assigned to a computational cell.
17
2.2.2 Associated Bed Gradation Templates with Bed

Material Layers
Once the user has created Bed Gradation Templates, and a Sediment Bed Materials
Layer in RAS Mapper these can be associated in the 2D Bed Gradations (Beta) tab of the
Sediment Data editor.
After user defines the sediment Bed Gradation Templates, they are available in a drop-
down pick list under the Bed Gradation column of the Sediment Data grid. Clicking on a
cell of the Bed Gradation column generates a drop-down list of the defined bed sample
templates. A single bed sample is frequently associated with multiple Sediment Bed
Material Layers. Therefore, once selected; a sample can be easily copied into multiple cells
by placing the mouse pointer over the bottom right corner of the selected cell and dragging
vertically.
The drop-down list always includes all of the user specified bed gradations and layer groups
as well as a non-erodible option. If the Bed Material Layer represents a hard surface,
engineered element (e.g. bank protection, wing wall, bed rock, or river engineering
element), selecting the non-erodible option will allow the model to deposit on top of them,
but not erode below the starting terrain surface.
A simple example of a dataset is shown below which only has one Bed Material Type and
one Bed Gradation. Simple datasets like this are most easily created by starting with an
empty Sediment Bed Materials Layer in RAS Mapper and creating polygons for different
regions of different bed composition. Each polygon corresponds to a Bed Material Type and
can then be associated with a Gradation in the 2D Bed Gradations tab of the Sediment
Data editor.
Figure 2-17. Simple, single gradation, 2D sediment specification with one Bed Gradation
associated with one Bed Material Type.
A more complicated dataset is shown in the figure below for which the Sediment Bed
Material Layer was created by importing a shape file. The shape file defines the polygon
types in RAS Mapper with integers, that show up under Bed Material Type. The user then
18
defined and associated bed gradations with these Sediment Bed Material classifications.
Users can rename these imported polygons under the Layer PropertiesFeatures menu
of the RAS Mapper Layer.
Figure 2-18. More detailed bed material specification, where the Bed Material Layer was
imported from a shape file.
2.2.3 Sediment Boundary Conditions

The second tab on the Sediment Data editor defines sediment boundary conditions (see
figure below). The editor automatically lists external model boundaries. HEC-RAS requires a
boundary condition at all external boundaries. If any external boundary (i.e. boundary
condition line) is left unspecified, then an equilibrium sediment load boundary condition is
assumed. The types of boundary conditions which may be specified in a 2D sediment model
are:
1. Rating curve
2. Sediment load series
3. Equilibrium load
4. Clear water (no sediment)
The other boundary condition options available from the interface (which are mostly
associated with 1D flow splits) are not allowed at 2D sediment transport boundaries. Only
inflow sediment boundary conditions are required. When the model computes flow out of
the domain the sediment can simply leave the domain. At 2D boundaries that where flux is
always out of the domain, users can select the equilibrium boundary condition may or
simply leave the sediment boundary condition empty. Any boundary condition lines left
empty are assigned an Equilibrium Load boundary condition. This is necessary so a
boundary condition type is available if the flow were to reverse direction even locally during
the simulation.
19
Figure 2-19. Boundary Conditions tab in the Sediment Data editor.
Rating Curve
The rating curve specifies the sediment load in tons/day or sediment concentration in mg/l
as a function of flow discharge. The fractional composition of the incoming sediment load is
specified for each grain class. The fractional loads are then specified to each cell using the
cell sediment capacity as a weighting function. For 2D sediment transport, the user has the
option to no specify the fractional bed-load composition. In this case, the cell sediment
capacities (equilibrium concentrations) are used to compute the fractional sediment loads at
each cell. This option is useful when the rating curve sediment gradations are not known.
Figure 2-20. Sediment Rating Curve editor.
20
Sediment Load Series

The Sediment Load Series boundary condition specifies the sediment load in tons per time
increment (see figure below). The sediment gradation may be specified with a Gradation
Rating Curve or if this data is left empty, the gradation is determined by weighting the
local (cell) equilibrium sediment loads. The Gradation Rating Curve specifies the
gradation of the inflow sediment as a function of total load in tons/day. It is important to
keep in mind that any grain specified in the gradation curve is automatically added to the
computed grain classes.
Figure 2-21. Sediment Load Series editor.
Equilibrium Load
The Equilibrium Load boundary condition specifies the inflow sediment load as the
equilibrium sediment load. The equilibrium sediment load is computed as the equilibrium
sediment concentrations at the boundary cells times the face flows. This approximation
essentially assumes a zero-gradient concentration normal to the boundary. The equilibrium
boundary condition should be used whenever data is not available or when first setting up a
model to quickly get simulation results or to compare results from other boundary condition
types which may be suspect.
21
Clear Water
The Clear Water boundary condition specifies a zero load/concentration at the boundary.
This boundary is not very commonly used except for very specific situations.
Unsteady Temperature
Temperature is the only data the unsteady flow editor required for sediment transport
analyses. Specify temperature for an unsteady sediment transport model in the Unsteady
Flow Editor. Select the Water Temperature (for Unsteady Sediment)… option from
the Options menu (see figure below).
Figure 2-22. Opening the Unsteady Temperature editor from the Unsteady Flow Data editor.
The Unsteady Temperature time series editor, is similar to the unsteady flow and stage
editors (see figure below). In the absence of temperature data in the unsteady flow file,
HEC-RAS will assume 55.4ºF. The water temperature is used to compute the water
kinematic and dynamic viscosities, and the water density which are then used by the
sediment transport model.
22
Figure 2-23. Specifying a temperature time series.
2.2.4 User-Defined Grain Classes

HEC-RAS divides the sediment material into multiple grain classes. Default grain classes
sub-divide the range of transportable material, (0.002 mm to 2048 mm) into 20 grain
classes or bins, each including adjacent, non-overlapping fractions of the grain size
spectrum. Default grain classes follow a standard log base 2 scale where the upper bound of
each class is twice its lower bound, the upper bound of the smaller, adjacent class. The gain
class represents all particles they contain with a single, representative grain size. HEC-RAS
uses the geometric mean of the grain class to represent the grain size for each bin. Grain
boundaries (and labels) are editable.
The user may change the size range for the grain classes by going to the Sediment Data
editor and selecting Options | User Defined Grain Classes… (see figure below).
23
Figure 2-24. Opening the User Defined Grain Classes editor from the Sediment Data editor.
In the User Defined Grain Classes editor (see figure below), the user may change the
diameter limits, names of the grain classes, and sediment properties including the Specific
Gravity (SG), Porosity (n), dry Bulk Density (BD). The label (name) for each grain class may
be changed in the table. The labels are used in the sediment output to identify variables
associated with each grain class such as fractional concentrations and particle fall velocities.
It is important to note that for 2D sediment transport, not all the grain-classes in the User
Defined Grain Classes editor are used. Only the grain classes which exist in the
Sediment Bed Gradations and/or the sediment boundary conditions are used. This saves
computational time and memory since only the grain classes which need to be computed
are used. Therefore, the number of grain classes in 1D and 2D sediment may be different.
Figure 2-25. User Defined Grain Classes editor.
24
The grain size classes can be viewed as analogous so the grid resolution. The table below
shows typical specific gravity ranges for different common minerals and rocks.
Table 2-1. Specific gravity ranges for different minerals and rocks.
Material Specific Gravity
Quartz 2.6 – 2.7
Limestone 2.6 – 2.8
Basalt 2.7 – 2.9
Magnetite 3.2 – 3.5
Coal 1.3 – 1.5
The grain specific gravity should not be used as calibration parameter. It is a parameter of
the sediment particles and should be specified as accurately as possible. However, in
general the results are generally not very sensitive to the specific gravity for natural
sediments in rivers.
Density Methods
The Density Method refers to the methods used to specify the grain class dry
density/porosity/unit weight:
1. Bulk Density (All Classes)
2. Porosity (All Classes)
3. UW-Cohesive/Porosity-Cohesionless
The default Density Method is Bulked Density (BD). To specify a different Density
Method, the user may select one of the options in the drop down menu as shown in the
figure below.
Figure 2-26. Selecting the Density Method in the Define Grain Classes and Sediment
Properties editor.
25
Depending on the choice selected the appropriate columns will be enabled or grayed out in
the User-defined Grain Classes editor. If the Bulk Density method is selected the BD
(for dry bulk density) column is enabled. The program computes a grain class porosity from
the specific gravity (column labeled SG) which may not be the same as the values in the
grayed-out porosity (column labeled n). If the Porosity (All Classes) method is selected,
the specific gravity and porosity columns are enabled. The program internally computed a
dry bulk density from the specific gravity and porosity, which may not be the same as that
which is displayed in the editor. The last option is the UW-cohesive/Porosity-
Cohesionless method. In this method the dry bulk density is specified for cohesive grain
classes, and the porosity is specified for the cohesionless (noncohesive) grain classes.
The treatment of the dry bulk density in HEC-RAS 2D sediment is different from that of
HEC-RAS 1D. In 2D the bulk density is treated as a property of the sediment mixture and
there is not always a one-to-one relationship between the grain class fractions and the bulk
density. This is necessary to simulate processes such as packing and consolidation.
If all the utilized grain class (cohesive and noncohesive) have the same bulk density and
consolidation is not enabled, a constant bulk density is assumed, which significantly
simplifies the computations. If the specified bulk densities or porosities for the noncohesive
grain classes are variable, a variable bed density model is utilized with the bed porosities
calculated from the Colby (1963) formula. The variable bed density model is more
computationally expensive because it requires several more calculations and steps during
the simulation. If the dry densities or porosities are left empty, then the Wooster et al.
(2008) porosity formula is utilized based on the sediment sorting. Dry densities generally
vary between 300 to 1,600 kg/m3.
2.2.5 Cohesive Options

Most of the sediment transport equations were developed with sand and/or gravel data.
Therefore, most silt and all clay particles are outside of the range of applicability of the
sediment transport functions implemented in HEC-RAS. In most systems, these particles are
wash load, material only found in the bed in trace amounts, because transport capacity
always exceeds supply. Some modelers will just ignore fines as throughput load, arguing
that if fines never interact with the bed in the model reach, the model is insensitive to them
and they add unnecessary complexity and parameters to the model. However, sometimes
fines must be modeled explicitly. In reservoirs and other backwater or low energy zones, silt
and clay can deposit and clay lined channels, both natural and engineered, can erode,
causing local and downstream problems.
Fine sediment transport is further complicated by electrostatic and electrochemical forces.

These particles are not just outside of the empirical range of the equations, but they often
erode and deposit by fundamentally different processes. These forces cause fine particles,
particularly clay, to flocculate and “stick” to the bed surface, so that fine erosion and
deposition are often not primarily functions of sediment size. These processes make fine
deposition and erosion fundamentally different than the cohesionless sand and gravel
transport.
In HEC-RAS, the cohesive options can be set by selecting Set Cohesive Options… under
Options menu of the Sediment Data editor (see figure below).
26
Figure 2-27. Opening the Cohesive Options editor from the Sediment Data editor.
The Cohesive Options editor is shown in the figure below. The editor allows the user to
select how the erosion is computed for cohesive sediments.
Erosion Parameters
HEC-RAS includes three cohesive erosion methods; applying the standard transport
equations, or two different implementations of the Krone and Partheniades approach. In
HEC-RAS 2D sediment transport, the HEC 6T Capacity Method is not available. When
utilizing the 2D sediment transport model, it is generally not recommended to use the
transport functions for the cohesive grain classes, especially clay. If the cohesive grain
classes represent a very small fraction of the bed gradation, then it is better to simply
remove them from the bed gradation then it is better to model them as cohesive.
Figure 2-28. Cohesive Options editor.
27
The treatment of cohesive sediments and mixtures of cohesive and noncohesive sediments
is slightly different for HEC-RAS 1D and 2D. HEC-RAS 1D considers the smallest five grain
classes ‘fine sediment’. HEC-RAS 1D applies the same cohesive erosion method selected to
these grain classes. By default, the first five grain classes are the clay and silt classes, and
are all finer than 0.0625 mm. However, if the user edits the diameters sizes for these, the
cohesive erosion methods will still apply to the first five grain classes, regardless of their
size. HEC-RAS 2D has diameter thresholds that determine whether a grain class forms flocs
and behaves cohesively on the bed. The reason for having these separate thresholds is to
allow for the erosion to be calculated with the user-selected transport function while still
allowing fine grain classes to form flocs. The upper threshold for grain classes to form flocs
is 0.0625 mm. The size tolerance for grain classes are behave cohesively on the bed is
either set to 0.0 mm if the Use Selected Transport Functions for All Grain Sizes option
is selected or 0.0625 mm if the Krone/Partheniades formulas are selected.
In HEC-RAS 1D if more than 20% of the active layer is cohesive, then the model considers
the sediment active layer as cohesive and computes the erosion for all grain classes using
the cohesive method. However, in 2D the erosion of mixed cohesive/noncohesive sediments
is computed
Ariathurai (1974) parameterized Partheniades (1962) results into a formula with an erosion
rate coefficient M and critical shear for erosion 𝜏𝜏𝑐𝑐 . M and 𝜏𝜏𝑐𝑐 include the stochastic nature of
both the sediment bed composition and surface and the bed shear stress. The critical shear
for erosion is usually in the range of 0.2 to 0.8 Pa, while M is usually in the range widely
from 0.01 to 50 Pa/hr (Winterwerp and van Kesteren 2004). Dahl et al. (2018) obtained
good results in simulating the lower Mississippi river using 𝜏𝜏𝑐𝑐 = 0.02 lb/ft2, 𝜏𝜏𝑚𝑚𝑚𝑚 = 0.04 lb/ft2,
M = 0.05 lb/ft2/hr, and M = 0.03 lb/ft2/hr.
Flocculation
The settling velocity of cohesive particles is computed using a flocculation settling velocity.
The flocculation settling velocity takes into account the formation and destruction of flocs as
a function of the sediment concentration. HEC-RAS has two methods for estimating the
settling velocity of flocs (see figure below). The first is the method of Hwang (1989). The
method represents the settling velocity as 4 separate zones: (1) free settling, (2)
flocculation, (3) hindered settling, and (4) negligible settling. The method requires 4
empirical coefficients.
Figure 2-29. Selecting the Hwang (1989) floc settling velocity formula.
Example coefficient values for the Hwang (1989) formula are shown in the table below.
28
Table 2-2. Example coefficient values for the Hwang (1989) formula from literature.
Reference Location a b m n
Krone (1962) San Francisco Bay, CA 0.048 25.0 1.00 0.40
Owen (1970) Severn River, UK 0.100 10.0 1.30 1.0
Nichols (1984) James River, VA 0.039 3.8 1.32 1.52
Hwang (1989) Lake Okeechobee, FL 0.080 3.5 1.88 1.65
Costa (1989) Hangzhou Bay, China 0.100 6.20 1.60 1.20
Marván (2001) Ortega River, FL 0.160 4.50 1.95 1.70
Ganju (2001) Loxahatchee River, FL 0.19 5.80 1.80 1.80
The second method available to estimate the floc settling velocity is a user-specified
Flocculation Curve. The Flocculation Curve defines the settling velocity as a function of
the suspended sediment concentration. An example of a user-specified flocculation curve is
shown in the figure below. The free-particle settling velocity is used when it is higher than
the user-specified fall velocity for concentrations less than the concentration at the peak
settling velocity. When the concentration is high than the maximum concentration specified
in the flocculation curve, the settling velocity corresponding to the maximum concentration
is used.
Figure 2-30. Specifying a floc settling velocity as a function of suspended sediment

concentration.
29
In addition to taking into account the sediment concentration, HEC-RAS corrects the floc
settling velocity for the water temperature by applying a correction which is a function of
the water dynamic viscosity. This assumes that the user-specified curve and the coefficients
in the Hwang method correspond to a standard water temperature of 55.4ºF.
Consolidation
Consolidation curves are developed for a specific bed material. Since HEC-RAS only allows
the user to specify one consolidation curve, a representative consolidation curve is used for
the all of the modeling domain. It is assumed that the consolidation curve was developed
for cohesive sediments. Until more than one consolidation curve is added to HEC-RAS a
simple correction is computed to account for the presence of noncohesive sediments which
can significantly affect the consolidation curve.
250
)
3
200
150
Dry Bulk Density (kg/m
100
50
-1 0 1 2 3 4 5
10 10 10 10 10 10 10
Time (day)
Figure 2-31. Example consolidation curve of dry bulk density as a function of time.
2.2.6 Transport Methods

The Transport Methods editor is where the user may select the parameters and values for
several transport-related variables. These include the Load Correction Factor, Diffusion
Coefficient, and adaptation parameters. To open the Transport Methods editor, open the
Sediment Data editor, and select open the Options menu and select Transport
Methods… (see figure below).
30
Figure 2-32. Opening the Transport Methods editor from the Sediment Data editor.
Load Correction Factor

HEC-RAS 2D sediment transport has the option to approximate the current velocity and
concentration profiles with approximate semi-analytical profiles, as well as to utilize an
empirical formula for the bed-load velocity (see figure below). If these options are enabled,
a load correction factor is included in the temporal term of the transport equation (i.e.
advection-diffusion equation). The load correction factor accounts for non-uniform vertical
profiles of the concentration and current velocity as well as the bed-load velocity. Since
most of the sediment concentration is typical near the bed where the current velocities are
slower, the load correction factor is generally less than one and produces a temporal lag
between the flow and the sediment concentrations.
Figure 2-33. Schematic of sediment and current velocity profiles.
The load correction factor options are specified in the Transport Model and AD
Parameters editor which can be accessed by opening the Sediment Data editor and
selecting the menu Options and selecting Transport Methods…
The load correction factor options are specified within the section Total-Load Correction
Factor section of the first tab of editor called AD Parameters. By default, HEC-RAS does
not compute any load correction factors. To turn on the load correction factor the user may
check the box labeled Total-load Correction Factor. One this checkbox is selected the
user should select the methods used to compute the bed and suspended-load correction
factors.
31
Figure 2-34. Setting the Total-load correction factor options in the Transport Model editor.
The bed-load correction factor method selects the formula for computing the bed-load
velocity. The available bed-load velocity formulas are:
1. No correction
2. Van Rijn
3. Van Rijn-Wu
4. Phillips and Sutherland
Selecting No Correction assumes the bed-load velocity is equal to the depth-averaged

velocity. The van Rijn and van Rijn-Wu are similar formulas. The only difference is that the
van Rijn-Wu formula has coefficients that have been recalibrated with a larger dataset of
measurements.
32
Figure 2-35. Setting the bed-load correction factor options in the Transport Model editor.
The suspended load correction factor methods select vertical sediment concentration profile.
The three options are:
1. No correction
2. Rouse sediment concentration profile
3. Exponential sediment concentration profile
When selecting the Rouse and Exponential sediment concentration profiles, it is assumed
that the current velocity follows a logarithmic profile. If No Correction is selected, the
suspended-load correction factor is set to one.
Figure 2-36. Setting the suspended-load correction factor method in the Transport Model
editor.
33
Generally, for practical applications there is not enough data to calibrate the bed and
suspended-load correction factors. Their use is a compromise between accuracy and
computational time which are both relatively minor. In general, the morphology change is
much more sensitive to other parameters and options such as the transport function and
adaptation parameters than the load correction factors.
Diffusion Coefficient
Horizontal turbulent mixing and dispersion is modeled in HEC-RAS with a Fickian diffusion
model. Turning on horizontal mixing is important in simulations with high resolution and
sharp variations in sediment concentration capacities and concentrations, and when using
high-resolution advection schemes. When the computational grid is relatively coarse or
when using a first-order advection scheme the numerical diffusion may be so large that
adding horizontal mixing is unnecessary. In general, sediment diffusion should not be used
in combination with the Diffusion Wave Equation (DWE), since this can result and overly
diffusive results.
The methods for computing the horizontal diffusion coefficients are in the AD Parameters
tab of the Transport Model and AD Parameters editor which can be opened from the
Sediment Data editor under Options and selecting Transport Methods... HEC-RAS has
two options for the total-load diffusion coefficient.
1. Constant
2. Weighted bed-load and suspended-load diffusion coefficients
The default in HEC-RAS is for total-load diffusion of zero (i.e. no horizontal diffusion).
Figure 2-37. Specifying a constant total-load diffusion coefficient.
34
Figure 2-38. Specifying a total-load diffusion coefficient based on weighted bed and
suspended load coefficients.
Erosion Parameters
The Erosion Parameters tab of the Transport Model and AD Parameters editor
contains the settings for the noncohesive sediment transport erosion formulation. The total-
load sediment erosion is computed as a function of the total-load adaptation coefficient. In
HEC-RAS there are two methods for computing total-load adaptation coefficient:
1. Total-load length
2. Weighted bed-load and suspended-load adaptation coefficients
The total-load adaptation length computes the total-load adaptation length as a function of
the unit discharge and sediment fall velocity (see figure below). The total-load adaptation
length is the simplest and most computationally efficient of the two options.
Figure 2-39. Specifying a total-load adaptation length.
35
The weighted bed-load and suspended adaptation coefficients requires specifying methods
for the bed-load and suspended-load adaptation coefficients and computing the fraction of
suspended sediments. However, it is the most physically accurate method.
Figure 2-40. Specifying a weighted total-load adaptation length.
Figure 2-41. Specifying a suspended-load adaptation coefficient.
When first setting up a sediment transport model, it is recommended to use a constant total
load adaptation length for simplicity. Once the user has a stable model producing
reasonable results, it is recommended to perform a sensitivity of the adaptation length by
adjusting the adaptation length. In many cases, the results are not found to be sensitive.
This is usually for relatively coarse grid simulations under mild to moderate forcing.
However, if the results are found to be sensitive to the adaptation length then more tests
are necessary in determining to optimal adaptation method and parameters.
36
2.2.7 Transport Function Calibration and Modification

The Transport Function Calibration and Modification editor allows the user to scale or
calibrate different transport formula in various ways. It is opened from the Sediment Data
editor by opening the Options menu and selecting Calibrate Transport Function (see
figure below).
Figure 2-42. Opening the Transport Function Calibration and Modification editor from the
Sediment Data editor.
The preferred method for calibrating the transport function is by means of the scaling
factors. The total-load scaling factor is applied the same to all the transport functions. The
mobility factor multiplies by the threshold for incipient motion (albeit a critical shear,
reference shear, or Shields number). Therefore, a mobility factor larger than 1 reduces the
mobility of a sediment. The option is also available to calibrate specific coefficients and
exponents of several transport formula.
37
Figure 2-43. Transport Function Calibration and Modification editor.
2.2.8 2D Options
There are several sediment transport options which are only available in 2D. The goal is for
both 1D and 2D sediment transport to have the same options but for now the 2D only
options have been places in a single editor called 2D Options which is available from the
Sediment Data editor under the Options menu (see figure below).
38
Figure 2-44. Opening the 2D Options editor from the Sediment Data editor.
The 2D Sediment Options editor is shown in the figure below. The editor is used to specify
the options for sheet and splash erosion, the morphologic acceleration factor, the base bed-
slope coefficient, and hindered settling.
Figure 2-45. 2D Options editor.
Sheet and Splash Erosion

The first option available in the editor is the Sheet and Splash Erosion. The 2D sediment
transport model has the option to use the Wei et al. (2009). The formulation has an
erodibility coefficient which for simplicity is set to a single value in HEC-RAS. The units of
the erodibility coefficient are kg∙m-3.644∙s0.644. Wei et al. (2009) reported values for the sheet
and splash erosion coefficient between 1124 and 2555 kg∙m-3.644∙s0.644 for 3 grassland
rangeland plots in Arizona (all variables in the International System of Units). However, its
value can vary by orders of magnitude for different soil types and cover characteristics.
39
Morphologic Acceleration Factor

The Morphologic Acceleration Factor is a parameter which can be used to turn off the
bed change or as the name implies, to accelerate the bed change. The factor is directly
multiplied by the mass bed exchange rates at every time step. A value of zero will turn off
the bed elevation and bed composition change.
The Morphologic Acceleration Factor can be utilized in several ways. The first and most
obvious is to simply turn off updating the bed elevation and bed composition by setting it to
a value of zero. This saves computational time and is useful for idealized situations are
when debugging a problem with a sediment model.
Another way the factor can be used is as a scaling factor to the morphological change. In
other words, the factor can be used to simulate a time period which represents the
morphological change of a time period equal to the simulation period times the
Morphologic Acceleration Factor. This approach is commonly used in coastal applications
with tidal boundary conditions. As an example, a 5-year simulation can be run with
Morphologic Acceleration Factor of 20 to simulate 100 years of change which greatly
reduces the computational time. However, it should be noted that this approach changes
the order of events (i.e. storms and tidal forcing with respect to morphological features)
which can have a negative impact on the accuracy of the approach.
Lastly, the Morphologic Acceleration Factor can also be used as a time speed-up factor.
This approach is more appropriate in river applications or when simulating single events or
sequences of events. As an example, a 20-day simulation can be reduced to 2 days by
setting the Morphologic Acceleration Factor to 10 and importantly also speeding up the
boundary conditions by the same factor. However, just as in the previous example, care
must be taken as to not speed-up the time so much that the hydrodynamics change
substantially and thereby the morphological change.
The Morphologic Acceleration Factor should be used with caution as it can lead to
misleading results or instabilities. The best practice for using the factor is to test the validity
of the factor by doing a full or partial length of a simulation with the factor and without it
and compare the results. If the results agree reasonably well, then longer or alterative
simulations can usually be done with the Morphologic Acceleration Factor thus saving
time. It is generally not recommended to use a factor larger than 30 to 50, and values
between but not equal to 0 and 1. The Morphologic Acceleration Factor can be verified
by reducing its value (e.g. by a factor of 2) while also increasing the simulation time by the
same factor and verifying the simulation results do not significantly change.
Since, the Morphologic Acceleration Factor is applied to the bed exchange rates, the
approach does not change total-load transport rates or concentrations. In addition, the
approach is not locally mass conservative but approximately globally conservative.
Finally, it is important to emphasis that the Morphologic Acceleration Factor should NOT
be used as a calibration parameter. It should be used carefully and validated to make sure
the results are not sensitive to the Morphologic Acceleration Factor. When used carefully
the Morphologic Acceleration Factor can be a powerful and useful tool in numerical
studies.
40
Base Bed-Slope Coefficient

The Base Bed-Slope Coefficient specifies the maximum value of the bed-slope coefficient.
The coefficient is then reduced based on the skin and critical shear stresses. As the ratio
between the skin and critical shear stress increases bed-load particles are less influenced by
the bed slope. The coefficient is used to compute an additional sediment flux in the
downslope direction which is a function of the bed slope, the bedload, and the bed-slope
coefficient. For further details on the formulation, the user is referred to the HEC-RAS 2D
Sediment Technical Reference manual (HEC 2020). The Base Bed-Slope Coefficient
typically has a value between 0.1 and 1. Increasing the coefficient has the effect of
smoothing the bathymetry. Measured bed change can be used to calibrate the base bed-
slope coefficient. However, its effect is significantly less than the transport formula,
transport scaling factors and mobility scaling factor. The Base Bed-Slope Coefficient will
tend to improve model stability by smoothing out small scale instabilities. However, if the
value is too large the numerical scheme can become unstable. In these situations the user
can either reduce the computational time step or reduce the value of the Base Bed-Slope
Coefficient.
Hindered Settling
Hindered settling is the condition in which the settling velocity of particles or flocs is reduced
due to a high concentration of particles. Hindered settling is primarily produced by particle
collisions and the upward water flow equal to the downward sediment volume flux. Hindered
settling occurs to both cohesive and noncohesive particles. However, the hindered settling
correction described here only applies to noncohesive particles. The hindered settling of
cohesive particles is accounted for in the floc settling method. Currently, the 2D sediment
has the option to use the Richardson and Zaki (1954) formula for hindered settling (see
figure below). For simplicity the exponent in the Richardson and Zaki (1954) is set to 4.0
and cannot be modified in the user interface.
Figure 2-46. Setting the hindered settling method to the Richardson and Zaki (1954)
method.
41
2.2.9 Bed Mixing Options

The Bed Mixing Options editor contains both 1D and 2D sediment bed mixing options and
other parameters. The editor is opened from the Sediment Data editor by going to the
Options menu and selecting Bed Mixing Options … (see figure below).
Figure 2-47. Accessing the Bed Mixing Options editor from the Sediment Data editor.
An example of the Bed Mixing Options editor is shown in the figure below. The sections
which are applicable to 2D sediment are Hiding Functions and Active Layer Options.
Figure 2-48. Bed Mixing Options editor.
42
Hiding Functions
In non-uniformly sized sediment beds, smaller particles are hidden from the flow and
physically trapped by larger particles whereas larger particles are more exposed to the flow
and less impeded to move freely. The hiding (and exposure) function computes a correction
to the incipient motion variable such as a shear stress or velocity to account for the hiding
and exposure of particles. The hiding functions available in HEC-RAS are:
1. None (not correction)

2. Ashida and Michiue (1971)
3. Day (1980)
4. Egiazaroff (1965)
5. Hayashi et al. (1980)
6. Parker et al. (1982)
7. Proffitt and Sutherland (1983)
8. Wilcock and Crowe (2003)
9. Wu et al. (2000)
The hiding and exposure function can have a big impact on the results. Some of the hiding
and exposure functions were developed for in conjunction with specific transport potential
functions. For example, the Wu et al. (2000) hiding function was developed for use with
bed- and suspended-load transport potential functions published in the same paper.
Similarly, the Day (1980) and Proffitt and Sutherland (1983) hiding function where
developed specifically for the Ackers and White (1973) transport potential formula. Lastly
the Wilock and Crowe (2003) hiding function was developed specifically for the Wicock and
Crowe formula (Wilcock 2001; Wilcock and Crowe 2003).
Active Layer Options

Within the Active Layer Options section of the Bed Mixing Options editor are the input
options controlling the active layer thickness and the Exchange Increment Method:
2.2.10 Bed Gradations

Instead of specifying the sediment size distribution and potentially the cohesive sediment
properties at every 2D computational cell (and cell subarea), HEC-RAS uses a template
concept similar to that of the Channel Modification Editor. In HEC-RAS a Bed Gradation
Template describes the grain size distribution and cohesive sediment properties of a
sediment mixture. Bed Gradation Template are first defined in a database and then
associated to Sediment Bed Materials which are specified at each computational cell
within RAS Mapper. The Bed Gradation editor, is opened from the Sediment Data editor
by clicking on the button . An example grain size distribution in the Bed
Gradation editor is shown in the figure below.
To create a new Bed Gradation the user may click on the “New” button . Give it a name
(e.g. “Nonuniform”). Enter the gradation information as shown in the figure below.
43
Figure 2-49. Bed Gradation editor with an example grain size distribution.
The user may enter the size distribution as percent finer values or grain class percent
values. The current mode is indicated by the radio buttons under the table on the left side
of the editor names % Finer and Grain Class %. The user may also convert the values in
the table from one mode to the other by clicking on the button .
The 2D sediment transport model will only compute grain classes which are utilized in the
initial bed gradations or any boundary conditions. This reduces the computational costs and
speeds up the model. Consequently, it is important for the user to be mindful of how many
grain classes are utilized in the input. For example, HEC-RAS has 5 cohesive grain classes.
Except for the particle fall velocity, all the cohesive grain classes are treated the same,
therefore using 5 cohesive grain classes or just 1 will have little impact on the results but a
big impact on the computational costs.
Optionally, HEC-RAS allows users to associate bulk bed properties with bed gradations.
More specifically, the sediment cohesive properties can be specified for each bed gradation.
This option is activated by selecting the checkbox at the bottom of the Bed Gradation
editor labeled Set Sample Specific Cohesive Parameters (see figure below).
44
Figure 2-50. Setting sample specific cohesive parameters in the Bed Gradation editor.
45
2.3 Viewing Sediment Results in RAS Mapper

HEC-RAS and RAS Mapper do not write sediment results by default. The default results
include Depth, Velocity, and Water Surface Elevation (WSE). Users must add sediment
results to the Results tab like they would add any other hydraulic paramters.
Expand the Results node in the RAS Mapper tree. Each simulated plan has a node under
Results. Right click on the plan and select Create a New RAS Map Layer to add
variables.
Figure 2-51. Right click on the Plan Description under Results and Select Create a New
Results Map Layer to add Sediment results (or non-default hydraulic results).
The Results Map Parameters editor has two main Map Types: Hydraulics and Sediment.
Expand the Sediment tree in the left pane of the editor to see the available sediment
parameters. The figure below includes the default sediment results avialble. Select
additional 2D sediment results by selecting Customized Variables in the Sediment
Output Options menu.
46
Figure 2-52. Two types of sediment results (Bed and Transport) in the Results Map
Parameter editor.
2D sediment maps fall into two main catagories: Sediment Bed Maps and Sediment
Transport Maps. Sediment Bed maps include results like bed change, bed elevation, bed
gradation, and other bed properties. Sediment Transport includes results like concentration,
capacity, and other results dynamic, water column or flux results. Any results that HEC-RAS
also computes by grain class will also include an expandable tree of results (like Total-load
Concentration in the example above). Select the variable node (e.g. Total-load
Concentraiton in the example above) to add the Total and all the grain classes to RAS
Mapper. Or, click on one of the sub nodes (e.g. Total, VFG, MG) to add only the total or only
one grain class at at time. In the example bleow the user added Bed Change and all of the
Total-load Concentration results but is only diplaying the Fine Gravel Concentration (FG).
Simply clicking the checkbox to turn sediment results on often does not display any results
because sediment reuluts are often not initialized in the first time step.
Use the animation bar to view results through time.
47
Figure 2-53. Example use of the animation bar to view sediment concentrations in time.
HEC-RAS 6.0 cannot plot cell face results yet.
Modeling Note – Subgrid Visualiation - Because the 2D model updates the bed on the
subgrid scale but RAS Mapper does not display subgrid results yet, RAS Mapper will
map average results over the wet, dry, or total cell. “Wet” cell results will average
the bed properties and bed change properties from the sub-grid portions of the cell
below the computed water surface elevation and “dry” results plot the sub-grid
results for the cell above the computed water surface. Sediment bed results that are
not specified as wet or dry average results ove the whole cell.
After adding the map or maps, Close the Parameter Selection editor. The selected map(s)
should appear under the plan heading in the Results node (see Figure below). It is almost
always useful to edit the default color ramp to visualize sediment results, and there are
some standard practices for some variables. To edit layer properties, right click on the map
and select Layer Properties (see figure below).
This will launch the Layer Properties editor (see figure below). To edit the color ramp
press the Edit button under Surface.
48
Figure 2-54. To edit the color ramp and display properties of the seidment maps, right click
on the map and select Layer Properties. Then Click Edit under the Surface
menu.
RAS Mapper includes a standard Bed Change color ramp designed to display this common
sediment result map (see figure below). However, this display works best if it is symetrical,
where white is “no/minor change” and aggredation and degredation get their own colors. To
make bed change semetrical, turn of the Use Dataset Min/Max to make the Min and Max
editable, and then set them both equal to whichever value is larger, making sure to keep
the Min negative (e.g. if the Max in the example below is 0.091 and the Min is -0.073, set
the Min to -0.091 to make the bed change display symetrical).
49
Figure 2-55. Setting the Max and Min to equal values (whichever abolute value is larger)
with opposite signs, centers the Bed Change plot, making all depsoitin and
erosion the same colors and the range of no change white.
RAS Mapper has two default sediment color maps: Bed Change and Concentration.
Modelers will develop preferred color ramps and viusaliztion approaches. RAS Mapper allows
users to add customized color ramps and visualization parameters to the color ramp editor
(see figure below). Edit the color ramp to optimize the visualization, then press Save Color
Ramp. Give the customized color ramp a name, and it will show up at the bottom of the
color ramp list for any result in this project.
50
Figure 2-56. Save customized sediment color ramps that persist throughtou the project with
the Save Color Ramp button.
2.3.1 Plotting 2D Sediment Time Series

Right click on a cell or cell face to visualize sediment time series within the model domain.
Right clicking on a model cell or face generates the vidualization menu depeicted in the
figure below. Select Plot Time Series . The Plot
Time Series option expands into a sub-menu with that includes all of the Results maps
(default and added) included in the active plan(s). Only the Results Maps selected (checked)
in RAS Mapper are avialble, however. In the example below, WSE, Velocity, Depth, and
Courant are not selected (checked) in RAS Mapper, so they are greyed out. But bot Bed
Change and Total-Load Concentration are checked, so they are availbe. The example below
slected Cell Bed Change, requesting the depicted time series at this cell.
If sediment results have multiple grain classes, (e.g. Total-load Concentration below) these
variables have additional sub-menus where users can choose if they want to plot the Total
result or the result from individual grain classes. Again, only the sub-class results selected
in the main RAS Mapper pane will be available.
51
Figure 2-57. View a time series of a sediment (or hydraulic) result by selecting (chekcing) it
in the RAS Mapper result tree and then right clicking on the cell. Select Plot
Time Series and then the result map to launch the time series for that cell or
cell face.
2.3.2 Plotting 2D Sediment Transects/Profile Lines

RAS Mapper can also display results for multiple cells along a transect. HEC-RAS callse
transects of 2D results Profile Lines. To create a results transect Right Click on Features
menu and select Create a New LayerPolyline Layer. Name the Profile Line and press
OK (see figure below).
52
Figure 2-58. Right Click on the Feature menu to create a Polyline Layer. Give it a name.
This adds the transect to Features and activates tehe drawing tool. Draw the polyline that
you would like to visualize results along. Dobule click when it is complete. Then right click
on the new Feature (Bed Change Transect in the figure below) and select Stop Editing.
Click Yes when asked to save.
Figure 2-59. Draw the polyline layer along the 2D sediment results transect.
Now, right clicking on this feature, generates a menu similar to the menu RAS Mapper
provides when right clicking a cell or cell face. Users can select a profile plot and animate it
with time. A profile will plot a longitudinal transect of the seleted result for the displayed
time step. As with all of the time series and profile plots in RAS Mapper, press the Table tab
to retrieve the numerical results behind this plot.
53
Figure 2-60. Plot a bed change transect with a polyline profile plot.
2.4 Sediment Computation Options and Tolerances

The sediment transport user-specified computational options, parameters, tolerances, and
methods are set in the Sediment Computation Options and Tolerances editor. To open
the editor, open the Unsteady Flow Analysis editor and select the Options menu and
click on the menu item Sediment Computation Options and Tolerances (see figure
below).
54
Figure 2-61. Bed Gradation editor with an example grain size distribution.
The Sediment Computation Options and Tolerances editor has two tabs: General and
2D Computational Options. These tabs are described below.
2.4.1 Sediment Computation Options and Tollerances

General
The General tab of the Sediment Computation Options and Tolerances editor contains
the 1D computational options but also includes some options which the 1D and 2D sediment
models share (see figure below). The 2D sediment transport model uses the Bed
Roughness Predictor, the Computation Multiplierand Warmup Periods.
55
Figure 2-62. General tab of the Sediment Computation Options and Tolerances editor.
Bed Roughness Predictor

The Bed Roughness Predictor options include three methods:
1. Limerinos
2. Brownlie
3. Van Rijn
The bed roughness predictor computes bed roughness dynamically based on the
hydrodynamics and bed materials. The sediment transport model and hydraulic models use
the computed roughness, then update it based on the new hydraulics and bed materials.
The 1D model allows users to define sub-reaches to apply the bed roughness predictor, but
the 2D model will apply the selected algorithm to the all wet cells.
When using bed-roughness predictor it is recommended to also output the cell bed
roughness’s and evaluate the computed bed roughness to maker sure results are realistic.
56
Warmup Periods
Warmup periods are simulation time windows added before the start of the simulation
(negative time) during which the boundary conditions are held constant and certain state
variables are updated. The Sediment Warmup Periods are computed after the unsteady
flow Initial Conditions Time and before the unsteady flow warmup period. During the
Initial Conditions Time, the flow and stage boundary conditions are “ramped up” or
increased from dry conditions to their initial value. During this period, each 2D area is
running separately. The unsteady flow warmup period is used to reach an initial condition
for hydraulics when running 1D and 2D flow simulations or with multiple coupled 2D areas.
Since 2D sediment does not support coupling of sediment across 1D and 2D or multiple 2D
areas, the unsteady flow warmup period should not be used with 2D sediment.
There are three Sediment Warmup Periods run in sequence: (1) Concentration, (2)
Gradation, and (3) Bathymetry. The start of each warmup period activates a portion of the
sediment computations as described in the table below.
Table 2-3. Sediment Warmup Period
Sediment Warmup Period Variables Updated

Concentration Total-load sediment concentration and all
advection-diffusion related variables
Gradation Grain class fractions, bed layer thickness, and
bed variables such bulk densities
Bathymetry Sediment and flow bed elevations at cells and
faces as well as hydraulic property tables
The Sediment Warmup Periods are specified at the bottom of the General tab of the
Sediment Computation Options and Tolerances window (see figure below). The
Sediment Warmup Periods are optional and any field left empty is assigned a zero value.
Figure 2-63. Specifying the warmup periods for 2D simulations.
During the Concentration Warmup Period, the 2D total-load advection-diffusion equation

is solved for sediment concentrations while holding all boundary conditions, bed gradations,
and bathymetry constant. The Concentration Warmup is only available in 2D sediment
transport, and avoids the initial erosion associated with starting a model with clear water
throughout the domain.
57
After the concentration warmup, the 2D model warms up the bed gradations. Starting with
the cell concentrations at the end of the concentration warmup, HEC-RAS will run the
unsteady, 2D, sediment model with constant flow, stage, and sediment boundary condtions
(those associated with the first time step in the simulation window). During the gradation
warmup, HEC-RAS will not adjust the bed, the elevation-volume curves for the cells, or the
elevation-area curves of the cell faces. It will only ajust the bed gradations, allowing them
to fine or (more often) coarsen in response to the flow field. The 2D sediment model will
also begin to develop vertical gradational stratigraphy in this warmup period, even bed
layers were not initially specified.
After the Gradation warmup, the 2D model will warmup the bathymetry. In this warmup
period, the model will still hold they hydraulic and sediment boundary conditions constant,
but will allow the bed elevations, and the elevation-volume/elevation-area curves to adjust
in response to transport. This phase will tend to fill local depresions, scour local raised
areas, plane out bed forms, highlight problems with the terrain, and, generally smooth the
terrain.
2D Computational Options
The 2D Computational Options tab of the Sediment Computation Options and
Tolerances editor contains the computational options which are utilized for 2D sediment
transport (see figure below). The editor has options and parameters related to the subgrid
erosion and deposition calculation methods, the advection scheme used in the transport
solver, the matrix solver, outer loop convergence tolerances,
58
Figure 2-64. Specifying the warmup method for 2D simulations.
Subgrid Erosion and Deposition Methods

One of the key features of HEC-RAS 2D is the subgrid modeling. HEC-RAS 2D sediment has
several methods for computing the subgrid erosion and deposition with different levels of
complexity and computational requirements.
The erosion rate is computed as the erosion potential times the bed availability (grain
fractions). Therefore, if the subgrid bed gradations vary, the subgrid erosion rates will also
vary. However, the subgrid erosion potential may be calculated with one of four methods
available (see table below).
59
Table 2-4. Summary of subcell erosion potential approaches.
Method Depth-Weighting Bed Properties Hydraulics

Constant None Cell Wet-Average Cell Wet-Average
Depth-Weighted Yes Cell Wet-Average Cell Wet-Average
Variable Bed None Subcell Cell Wet-Average
Full Subgrid None Subcell Subcell
The Constant erosion method is the simplest and most computationally efficient. For every
cell, a single erosion potential is computed using wet-area averaged hydraulics, and bed
properties. The bed properties include erosion coefficients, roughness, and grain size
distribution, etc. The Depth-Weighted erosion method also computes a single erosion
potential for every cell but applies a depth-weighting to compute the subgrid erosion
potentials. Therefore, the Depth-Weighted method uses averaged bed and hydraulics to
compute a cell erosion potential, but which is then weighted using the water depths to
compute the subgrid erosion potentials. Again, subgrid erosion rates always take into
account the subgrid grain fractions (material availability). The Depth-Weighted method is
only slightly more computationally expensive than the Constant method. The Variable
Bed erosion method does not perform depth-weighting but computes subgrid erosion
potentials utilizing the subgrid bed properties. The Variable Bed method utilizes average
hydraulic variables for the wetted area of the cell. The Full Subgrid erosion method does
not perform depth-weighting but computes subgrid erosion potentials utilizing the subgrid
bed properties and hydraulics. The method requires estimating subgrid hydraulic variables
such as current velocities, shear stresses, shear velocities, depths, etc. depending on the
transport potential formula utilized. The Full Subgrid method is the most computationally
expensive of all the methods and will generally lead to the largest variations in subgrid
erosion compared to the other methods. All of the erosion methods assume that a constant
adaptation coefficient within the cell.
HEC-RAS 2D sediment has three subgrid deposition methods (see table below). All the
deposition methods assume that the sediment concentration and adaptation coefficient are
constant within a cell. The first is the Veneer method. The Veneer method applies a
spatially constant deposition rate over the wet portion of the cell. The Veneer method is the
simplest and most computationally efficient of the methods. The Veneer deposition method
is analogous to the Veneer method utilized in HEC-RAS 1D sediment transport. The Depth-
Weighted deposition method is similar to the Reservoir method in HEC-RAS 1D sediment
transport. The Capacity-Weighted deposition method is similar to the Capacity-
Weighted erosion method.
Table 2-5. Summary of subcell deposition approaches.
Method Depth-Weighting Capacity-Weighting

Veneer (Constant) None None
Depth-Weighted Yes None
60
Capacity-Weighted None Yes
Advection Scheme
HEC-RAS 2D sediment transport has the option to choose between four advection schemes:
1. Upwind
2. Exponential
3. Minmod
4. Harmonic
The selected advection schemes are designed to off the use a wide range of options. The
Upwind and Exponential schemes are first-order difference schemes (Patankar 1980). The
schemes are linear and therefore have the benefit of not requiring additional outer loop
iterations. In addition, difference schemes are relatively simple, and provide smooth
bounded solutions. The Exponential scheme is based on the analytical solution of the
steady 1D advection-diffusion equation (without sources and sinks) and is less diffusive than
the Upwind scheme. However, when the horizontal mixing is turned off, the Exponential
scheme reduces to the Upwind scheme.
The Minmod and Harmonic schemes are high-resolution Total Variance Diminishing (TVD)
schemes (Harten 1983). High-resolution TVD schemes allow for second order or higher
accuracy in smooth and first-order accuracy in sharply varying regions without producing
spurious oscillations (bounded solutions). The TVD schemes are implemented here using a
flux-limiter formulation and a deferred correction approach which requires outer loop
iterations to converge. The Minmod scheme has good convergence characteristics but is
the most diffusive of the two TVD schemes. The Harmonic scheme is less diffusive,
produces good convergence characteristics in implicit schemes, and is equivalent to the
Hybrid Linear/Parabolic Approximation (HLPA) scheme of Zhu (1992).
Matrix Solvers
HEC-RAS 2D Sediment Transport solves an implicit Advection-Diffusion (transport) equation
for the fractional total-load concentrations. The discretization produces a linear system of
equations which may be represented by a sparse-matrix problem. The sparse matrix solver
is an important component of the computational options because a large portion of the
computational time can be spent solving the sparse matrices for each grain class transport
equation. HEC-RAS has three options for the matrix solver:
1. PARDISO
2. FGMRES-SOR
3. FGMRES-ILU0
The PARDISO solver is a high-performance, robust, memory efficient and easy to use
solver for solving large sparse symmetric and non-symmetric linear systems of equations on
shared memory and distributed-memory architectures. Here the Intel Math Kernel Libraries
(MKL) PARDISO solver is utilized. The solver utilizes a combination of parallel left- and
right-looking supernode pivoting techniques to improve its factorization process (Schenk
et al. 2004; Schenk et al. 2011).
61
Iterative solvers require an initial guess to the solution. Iterative solvers generally require
less memory for because unlike with direct solvers, the structure of the matrix does not
change during the iteration process. In addition, iterative solvers utilize matrix-vector
multiplications which can be efficiently parallelized. The main drawback of iterative solvers
is that the rate of convergence depends greatly on the condition number of the coefficient
matrix. For poorly conditioned matrices, the iterative solver may not converge at all.
Therefore, the efficiency of iterative solvers greatly depends on the size and condition
number of the coefficient matrix.
Iterative solvers may be classified into stationary and projection methods. In stationary
methods the solution for each iteration is expressed as finding a stationary point for the
iteration. The number of operations for iteration step for stationary methods is always the
same. Stationary methods work well for small problems but generally converge slowly for
large problems. Projection methods extract an approximate solution from a subspace.
Generally, projection methods have better convergence properties than stationary methods
but because each iteration is generally more computationally demanding than stationary
methods, they tend to be more efficient for medium to large systems of equations. The
main disadvantage of iterative solvers is their lack of robustness.
HEC-RAS 2D sediment has the option of two iterative solvers: FGMRES-SOR and FGMRES-
ILU0. The first part FGMRES refers to the matrix solver while second parts SOR and ILU0
refer to the preconditioner utilized. FGMRES (Flexible Generalized Minimal RESidual) is a
projection method which is applicable to coefficient matrices which are non-symmetric
indefinite (Saad, 1993). The “flexible” variant of the Generalized Minimal RESidual method
(GMRES) allows for the preconditioner to vary from iteration to iteration. The flexible variant
requires more memory than the standard version, but the extra memory is worth the cost
since any iterative method can be used as a preconditioner. For example, the SOR could be
used as a preconditioner with different relaxation parameter values each time it is applied.
The FGMRES (and GMRES) method becomes impractical for large number of iterations
because memory and computational requirements increase linearly as the number iterations
increases. To remediate this the algorithm is restarted after iterations with the last solution
used as an initial guess to the new iterative solution. This procedure is repeated until
convergence is achieved. The FGMRES solver with restart is often referred to as FGMRES(m)
where m is the restart parameter. However, here the shorter name FGMRES is used for
simplicity. If m is too small, the solver may converge too slowly are even fail completely. A
value of m that is larger than necessary involves excessive work and memory storage.
Typical restart values are between 5 and 20. Here m is set to 10 which works well for the
most practical applications.
The Successive-Over-Relaxation (SOR) is a stationary iteration method based on the Gauss-

Seidel (GS) method. When ω = 1, the SOR method reduces to the Gauss-Seidel method. In
addition, for ω < 1, the method technically applies under-relaxation and not over-relaxation.
However, for simplicity and convenience the method is referred to as SOR for all values of
ω. Kahan (1958) showed that the SOR method is unstable for relaxation values outside of
0 < ω < 2. The optimal value of the relaxation factor is problem specific. The SOR method is
guaranteed to converge if either (1) if 0 < ω < 2, and (2) the matrix is symmetric positive-
definite, or strictly or irreducibly diagonally dominant. However, the method sometimes
converges even if the second condition is not satisfied. For simplicity, the relaxation
parameter is set to 1.3 here and cannot be adjusted by the user. The value of 1.3 has been
found to work reasonably well for a wide range of problems. A simple parallel version of the
SOR is utilized here which is referred to as the Asynchronous SOR (ASOR) which uses new
values of unknowns in each iteration/updates as soon as they are computed in the same
iteration (see Chazan and Miranker 1969; Leone and Mangasarian 1988). The ASOR is part
of a class of iterative solvers known as chaotic relaxation methods. Since the order of
62
relaxation is unconstrained, synchronization is avoided at all stages of the solution.

However, the convergence behavior can be slightly different for different number of threads.
The ASOR solver has been parallelized with OpenMP.
The SOR preconditioner is based on the SOR solver except that no convergence checking is
done during the iteration process (DeLong 1997). This is done for simplicity and to avoid
additional computations associated with the determining the convergence status. The SOR
preconditioner is utilized for non-symmetric matrices. The SOR preconditioner matrix is
given by
The ILU0 (Incomplete Lower Upper with Zero Infilling) is part of a large class of
preconditioners which utilize incomplete factorizations of the coefficient matrix. The
effectiveness of the preconditioner depends on how well the sparse matrix by factored into
lower and upper sparse matrices. ILU0 is one of the most common types of incomplete
factorizations. In this factorization, all non-zero values of the exact factorization which are
located in zero value positions are discarded. The advantage of the ILU0 preconditioner is
that it preserves the structure of the original matrix. Another advantage of ILU0 versus
other similar factorizations is that it does not require specifying additional tolerances and
settings.
Outer Loop Convergence Options

The outer loop refers to the loop in which the transport equations are coupled to the bed
change and sorting equations whereas the inner loop is the iterative solver loop. The outer
loop is also necessary for updating the deferred corrections from the high-resolution
advection schemes. If a direct sparse matrix solver is utilized, there is no inner loop. The
maximum number of outer loop iterations may be set by user (see figure below). The
maximum number of outer loop iterations is a compromise between computational
efficiency and accuracy. If the sediment model is iterating every time step and going to the
maximum number of iterations, this is an indication, the either the time step is too big or
the convergence tolerances have been set too small. Outer loop convergence is monitored
by change in the sediment concentrations and active layer bed fractions between outer loop
iterations. The convergence of the fractional total-load sediment concentrations is assessed
by means of two tolerances. The first is the maximum value of the absolute differences
(errors) in concentrations between outer loop iterations. The second is the root-mean-
square of the differences (errors) between outer loop iterations. For the first iteration, the
error is approximated by comparing the current concentrations with the previous time step
values. The convergence of the active bed grain fractions is assessed by means of a
maximum absolute error tolerance. This approach works well for both constant density and
variable density simulations.
63
Figure 2-65. Outer loop Convergence Parameters.
Whenever, the sediment concentrations or active layer grain fractions do not converge or
reach the maximum number of iterations, a message is printed to the log file. In addition,
the convergence status is written to the HDF5 file at the Mapping Output Interval. This log
information can be used to optimize the convergence parameters and computational time
step.
When simulating Non-erodible Surfaces, the active layer grain fractions can vary
significantly in a time step and even within a time step (outer-loop iterations). Therefore, it
is expected that when simulating Non-erodible Surfaces with multiple-grain classes, at
least one or two iterations will be needed for the active layer fractions to converge. In fact,
even when simulating a single grain class, iterations are still required because for the
sediment concentrations to convergence above Non-erodible Surfaces because the
erosion rates are limited before solving concentrations using an expression which includes
an estimate of the deposition rates from the previous iteration.
64
Computational Sediment Layer Parameters

There are two types of layers in HEC-RAS. The first type is the bed layers which are used to
specify the bed gradations and the second is the computational bed layers which are used in
the solution of the bed sorting and layering. Specifying initial sediment bed layers is
optional. If the user does not specify initial bed layers, a vertically uniform bed composition
is assumed, and computational bed layers are created based on the user-specified initial
bed layer thickness (see figure below).
Figure 2-66. Calculation of computational bed layer thickness from a single user-specified
initial bed layer or bed gradation.
If the user specifies initial bed layers, these are subdivided (discretized) such that the initial
bed layer thickness is equal to or less than then the user-specified initial bed layer thickness
(see figure below).
Figure 2-67. Calculation of computational bed layer thickness and composition from user-
specified initial bed layers.
65
Figure 2-68. Computational Sediment Layer Parameters section in the HEC-RAS Sediment
Computation Options and Tolerances editor.
66
0.06
0.05
0.04
0.03
Critical shear for erosion (Pa)
0.02
0.01
0 50 100 150 200 250
3
Dry bulk density (kg/m )
Figure 2-69. Example of a power-law parameterization of the critical shear for erosion as a
function of the dry bulk density.
2.5 Sediment Output Options

The sediment output options are entered in the apply named Sediment Output Options
editor which can be accessed from the Option | Sediment Output Options… menu in the
Unsteady Flow Analysis editor (see figure below).
67
Figure 2-70. Opening the Sediment Output Options editor from the Unsteady Flow Analysis
editor.
The Sediment Output Options editor allows the user to specify which variables to output
and at what frequency.
68
Figure 2-71. Sediment Output Options editor.
2.5.1 Output Level

The Output Level allows the user to control how much logging information is written to the
Computation Log File as well as select a predefined list of output variables. The Output
Level is set within the Sediment Output Options editor which is accessed by clicking on
the Options | Sediment Output Options menu within the Unsteady Flow Analysis
editor (see figure below).
69
Figure 2-72. Opening the User Defined Grain Classes editor from the Sediment Data editor.
Table 2-6. Output sediment variables as function of the Output Level.
Output Sediment Output Variable Units

Level
≥1 Cell average bed change ft or m
≥1 Cell total total-load concentration mg/L
≥2 Cell average bed elevation ft or m
≥2 Cell total total-load concentration capacity (i.e. mg/L
equilibrium total-load concentration).
≥3 Cell fractional total-load concentrations (for mg/L
each grain class)
≥4 Cell fractional total-load concentration mg/L
capacities (i.e. fractional equilibrium total-load
concentrations)
≥5 Cell total fraction of suspended sediments (not -
by grain class)
≥5 Cell total total-load transport potential lb/ft/s or
kg/m/s
≥6 Cell fractional fraction of suspended sediments -
(for each grain class)
≥6 Cell total bed shear stress lb/ft2 or
Pa
≥6 Cell grain bed shear stress lb/ft2 or
Pa
≥6 Cell critical shear stress for each grain class lb/ft2 or
Pa
≥6 Cell fall velocity for each grain class ft/s or
m/s
≥6 Cell 16th percentile diameter (d16) mm
70

2.5.2 Output Interval Multiples

Sediment transport output is classified into two groups referred to as Output Blocks: (1)
Sediment Transport and (2) Sediment Bed. The main reason for this classification is to
be able to easily control the output interval of variables which vary at similar time scales.
Generally, “transport variables” which are usually within the water column (e.g.
concentrations, transport rates, etc.) vary at time scales similar to the hydraulics, while
sediment variables on the bed (e.g. bed elevations and gradations) vary at longer time
scales. Allowing the user to specify different output intervals for these two output groups
helps reduce the size of the HDF5 Output File by allowing the user to specify a larger output
interval for variables which change slowly, compared to the hydraulics. The variables which
correspond to the Sediment Bed are the bed roughness’s, bed elevations, bed changes,
bed change rate, bed slopes, bed layer fractions, bed layer thicknesses, bed layer percentile
diameters, bed layer dry densities. All other variables correspond to the Sediment
Transport group. The precise Output Block for each output variable is noted in the captions
of Tables 2-6 through 2-14.
2.5.3 Customized Sediment Output Variables

There is a large number of variables that the user can select to output. The optional output
variables are selected within the Sediment Output Options editor which is accessed from
the Options | Sediment Output Options… menu from the Unsteady Flow Analysis
editor. The following Tables 2-6 through 2-14 list all of the optional output variables related
to 2D sediment transport and their units.
Table 2-7. Output hydraulic variables. These variables

correspond to the Sediment Transport Output Block.
Output Variable Units

Shear Stress Pa or lb/ft2
Shear Velocity m/s or ft/s
Velocity m/s or ft/s
2D Cell Hydraulic Depth m or ft
2D Cell Hydraulic Radius m or ft
2D Subcell Velocity m/s or ft/s
2D Subface Hydraulic Radius m or ft
71
Table 2-8. Output sediment concentrations. These variables


2D Cell Total-load Concentration (Total) mg/L
2D Cell Total-load Concentration (by GC) mg/L
2D Cell Suspended Concentration (Total) mg/L
2D Cell Suspended Concentration (by GC) mg/L
2D Cell Bed-load Concentration (Total) mg/L
2D Cell Bed-load Concentration (by GC) mg/L
2D Cell Total-load Conc Capacity (Total) mg/L
2D Cell Total-load Conc Capacity (by GC) mg/L
2D Cell Suspended Conc Capacity (Total) mg/L
2D Cell Suspended Conc Capacity (by GC) mg/L
Table 2-9. Output sediment transport potentials. These variables


2D Cell Total-load Transport Potential (Total) kg/m/s or lb/ft/s
2D Cell Total-load Transport Potential (by GC) kg/m/s or lb/ft/s
2D Cell Bed-load Transport Potential (Total) kg/m/s or lb/ft/s
2D Cell Bed-load Transport Potential (by GC) kg/m/s or lb/ft/s
2D Cell Suspended-load Transport Potential (Total) kg/m/s or lb/ft/s
2D Cell Suspended-load Transport Potential (by GC) kg/m/s or lb/ft/s
Table 2-10. Output sediment transport variables. These variables


2D Cell Fraction of Suspended Sediments (Total) -
2D Cell Fraction of Suspended Sediments (by GC) -
2D Face Total-load Diffusion Coefficient (by GC) m2/s or ft2/s
72
2D Face Bed-load Diffusion Coefficient (by GC) m2/s or ft2/s

2D Face Suspended-load Diffusion Coefficient (by GC) m2/s or ft2/s
2D Cell Total-load Correction Factor (by GC) -
2D Cell Bed-load Correction Factor (by GC) -
2D Cell Suspended-load Correction Factor (by GC) -
2D Cell Total-load Adaptation Length (by GC) m or ft
2D Cell Total-load Adaptation Coefficient (by GC) -
2D Cell Bed-load Adaptation Length (by GC) m or ft
2D Cell Suspended-load Adaptation Coefficient (by GC) -
2D Cell Erosion Rates (by GC) kg/m2/s or lb/ft2/s
2D Cell Deposition Rates (by GC) kg/m2/s or lb/ft2/s
Table 2-11. Output incipient motion variables. These variables


2D Cell Critical Shear Stress (by GC) Pa or lb/ft2
2D Cell Hiding and Exposure Correction Factor (by GC) -
Table 2-12. Output sediment settling variables. These variables


2D Cell Settling Velocity (Total) m/s or ft/s
2D Cell Settling Velocity (by GC) m/s or ft/s
It is noted that the units of density in the US Customary Unit System are represented as
lb/ft3. Here lb represent pounds mass and are obtained by multiplying slugs/ft3 by a
standard gravity of 31.174 ft/s2.
73
Table 2-13. Output sediment active layer variables. These variables

correspond to the Sediment Bed group (Output Block).

2D Cell Fractions Active -
2D Cell Layer Thickness Active m or ft
2D Cell Dry Density Active kg/m3 or lb/ft3
2D Cell d10 Active mm
Table 2-14. Output Sediment bed Composition and layering variables.

These variables correspond to the Sediment Bed Output Block.

2D Layer Fractions -
2D Layer Thickness m or ft
2D Layer Dry Density kg/m3 or lb/ft3
Table 2-15. Output bed roughness variables. These variables

correspond to the Sediment Bed Output Block.

2D Cell Manning Total s/m1/3
2D Cell Manning Grain
2D Cell Manning Bedform
2D Cell Manning Transport
74
Table 2-16. Output bathymetry variables. These variables correspond to

the Sediment Bed Output Block.

2D Cell Bed Elevation m or ft
2D Cell Bed Change m or ft
2D Cell Bed Change Rate m/s or ft/s
2D Face Bed Elevation m or ft
2D Face Bed Change m or ft
2D Face Bed Change Rate m/s or ft/s
2D Subcell Bed Elevation m or ft
2D Subcell Bed Change m or ft
2D Subface Bed Elevation m or ft
2D Subface Bed Change m or ft
2.5.4 Sediment Hotstart

Currently, the sediment hotstart feature is not supported by the beta release of 2D
sediment transport in HEC-RAS. This feature will be fully supported in the first official
release of 2D sediment transport. Currently, hotstart files (a.k.a. restart files) cannot be
used when using 2D sediment transport.
2.6 Viewing Results

2.6.1 Exploring the HDF5 File
The HEC-RAS HDF5 Output File may be explored by reading the file with a computer
language (e.g. MATLAB or Python) or by opening it with an external HDF5 utility such as
HDFView (https://www.hdfgroup.org/downloads/hdfview/) or Panoply
(https://www.giss.nasa.gov/tools/panoply/). The HEC-RAS output is grouped in the HDF5
file into Output Blocks. There are several output blocks including Base Output,
Computation Block, DSS Hydrograph Output, DSS Profile Output, among others.
HEC-RAS 2D sediment output is separated into two Output Blocks: Sediment Transport
and Sediment Bed. The path of the Output Blocks within the HDF5 file is
/Results/Unsteady/Output/Output Blocks/. The name of each output dataset follows a
specific convention as <Location> <Variable Name> for variables which are not specific to
each grain class and <Location> <Variable Name> – <Grain Class Name> for variables
which are output for each grain class. For all the output variables, the rows indicate the
output times and the columns indicate the variation location (i.e. cell, face, subcell, or
subface).
75
2.6.2 Log Output

Monitoring the convergence status of a model is important but often difficult due the
complexity of the sediment transport models. The sediment log output may be grouped
similar to the HDF5 spatial output as: (1) Sediment Transport (sediment concentrations),
(2) Sediment Bed (grain class fractions and dry densities in the active layer). In addition,
the logging output is written out to two different places: (1) the Computation Log File
(*.bco*), and (2) the HDF5 Output File (*.p*.hdf).
The amount of information which is written to the Computation Log File depends on the
sediment Output Level (see Section 2.4 Sediment Output Options) and the convergence
status. The log output in the Computation Log File is output only when a convergence
tolerance is exceeded for Output Levels less than 5. If the output level is 5 or 6, a log
record is written at a constant computation time step interval. For an Output Level of 5,
the log record only contains convergence information for the last outer-loop iteration. For an
Output Level of 6 the log record contains convergence information for all iterations.
The log output in the HDF5 Output File is written to the same location as the Sediment
Transport and the Sediment Bed output blocks and is therefore written at the
corresponding output intervals. The output interval for the Sediment Transport and the
Sediment Bed output blocks can be controlled by the user by specifying a multiple of the
Mapping Output Interval (see Section 2.4 Sediment Output Options for details). The log
output is written as floating point datasets instead of compound datasets or other forms for
computational speeds.
The log output for Sediment Transport is written in the HDF5 Output File under
Dataset Name: Log Sediment Transport

Dataset Path: /Results/Unsteady/Output/Output Blocks/Sediment
Transport/Unsteady Time Series/2D Flows Areas/<2D Flow Area Name>/
where <2D Flow Area Name> is the name of each 2D flow area.
Table 2-17. Sediment Transport Log Output. Note: HDF5 columns begin at 0.
Column Log Output Variable Units

0 Overall convergence status flag -
1 Transport convergence status flag -
2 Number of outer-loop iterations -
3 Maximum total total-load concentration mg/L
4 Cell location of maximum total total-load -
concentration
5 Maximum fractional total-load concentration mg/L
6 Cell location of maximum fractional total-load -
concentration
7 Grain class corresponding to maximum fractional -
total-load concentration
76
8 Concentration maximum absolute correction for mg/L

all grain classes
9 Cell location of maximum absolute concentration -
correction for all grain classes
10 Grain class corresponding to maximum absolute -
concentration correction for all grain classes
11 Percent of active grid %
12 Minimum total inner loop iterations (only for -
iterative matrix solvers)
13 Grain class corresponding to minimum inner loop -
iterations (only for iterative matrix solvers)
14 Maximum total inner loop iterations (only for -
iterative matrix solvers)
15 Grain class corresponding to maximum inner loop -
iterations (only for iterative matrix solvers)
16 Maximum root-mean-squared residual of inner- -
loop iterations (only for iterative matrix solvers)
17 Grain class corresponding to maximum root- -
mean-squared residual of inner-loop iterations
(only for iterative matrix solvers)
Currently when the model diverges, it stops the simulation. In future versions, diverged
time steps will be repeated with subcycles in order to try to obtain a converged solution.
Usually divergence indicates a NaN (not a number for bed change, bed fractions, or
sediment concentrations). However, it could also mean an internal divergence tolerance was
exceeded. Future versions of the model will include more information on how the model
diverged.
Similarly, the logging output for the Sediment Bed is written in the HDF5 Output File
under
Dataset Name: Log Sediment Bed

Dataset Path: /Results/Unsteady/Output/Output Blocks/Sediment Bed/Unsteady
Time Series/2D Flows Areas/<2D Flow Area Name>/
where again <2D Flow Area Name> is the name of each 2D flow area.
The convergence status flags indicate different model states (see Table 2-17). A status is
computed each outer iteration for the sediment transport and bed calculations. The overall
convergence status takes into account transport and bed sorting.
77
Table 2-18. Convergence Status Values
Status Status Description

Value Name
-1 Diverged The solution has diverged. Divergence is determined
with a variety of tolerances and checks for many
variables.
1 Iterating Either the convergence tolerances or minimum number
of iterations have not been satisfied.
0 Converged Both the convergence tolerances and minimum number
of iterations have been satisfied.
2 Continue Either the convergence has stalled, or the maximum
number of iterations has been reached. The solution
will continue to the next time step.
Table 2-19. Sediment Bed Log Output. Note: HDF5 begin start at 0.
Column Log Output Variable Units

0 Overall status flag -
1 Sediment bed status flag -
2 Number of outer-loop iterations -
3 Maximum absolute value of active layer fraction correction %
(difference between outer-loop iteration values)
4 Cell location of maximum absolute value of active layer -
fraction correction
5 Subarea corresponding to maximum absolute value of -
active layer fraction correction
6 Grain class corresponding to maximum absolute value of -
active layer fraction correction
7 Number of avalanching iterations -
8 Number of avalanches at last iteration -
9 Maximum bed change due to avalanching ft or m
10 Cell location of maximum bed change to due avalanching -
78
RD-?? Chapter 4 – References
Chapter 3
References
Ackers, P., and White, W.R. 1973. Sediment transport: New approach and analysis,
Proceedings. ASCE, Journal of the Hydraulic Division, 99, HY11, pp. 2041-2060.
Ashida, K. and Michiue, M. 1971. An investigation of river bed degradation downstream of a

dam, Proc. 14th Congress of IAHR, Paris, France, 3, 247–256.
Ariathurai, C.R. 1974. A finite element model for sediment transport in estuaries, PhD
Thesis, University of California, Berkley, USA.
Armanini, A., and G. di Silvio. 1986. Discussion on the paper ‘A depth-averaged model for
suspended sediment transport’ by Galappattti, G., and Vreugdenhil, C.B. Journal of
Hydraulic Research. 24(5), 437–441.
Chazan, D., and Miranker, W. 1969. Chaotic relaxation. Linear Algebra and Its Applications,
2(2), 199–222.
Colby, B.R. 1963. Discussion of “Sediment transportation mechanics: Introduction and

properties of sediment”, Progress report by the Task Committee on Preparation of
Sediment Manual of the Committee on Sedimentation of the Hydraulics Division, V. A.
Vanoni, Chmn., Journal of Hydraulics Division., ASCE, 89(1): 266–268.
Costa, R.C.F.G. 1989. Flow-fine sediment hysteresis in sediment-stratified coastal waters,

MS Thesis, University of Florida, Gainesville, FL.
Dahl, T.A, Heath, R.E., Gibson, S.A., and Nygaard, C.J. 2018. HEC-RAS unsteady flow and
sediment model of the Mississippi River: Tarbet Landing to the Gulf, Mississippi River
Geomorphology & Potomology Program, Report No. 25, Engineer Research and
Development Center, Vicksburg, MS, USA.
Day, T.J. 1980. A study of the transport of graded sediments, Report N. IT 190, HR
Wallingford, U.K.
DeLong, M.A. 1997. SOR as a Preconditioner. PhD Dissertation. School of Engineering and
Applied Science, University of Virginia, 102 p.
Dietrich, W.E. 1982. Settling velocity of natural particles, Water Resources Research, 18(6),
1615–1626.
DHL 1980. Computation of Siltation in Dredged Trenches, Delft Hydraulics Laboratory,

Storm Surge Barrier Oosterschelde, Report 1267-V/M 1570.
Dixit, J.G. 1982. Resuspension potential of deposited kaolinite beds. M.S. thesis, University
of Florida.
79
Chapter 3 – References RD-??
Egiazaroff, I.V. 1965. Calculation of nonuniform sediment concentration, J. Hydr. Div.,

ASCE, 91(HY4), 225–247.
Engelund, F., and Hansen, E. 1967. A monograph on sediment transport in alluvial streams,
Teknisk Forlag, Copenhagen, 65.
Gallappatti, G., and Vreugdenhil, C.B. 1985. A depth-integrated model for suspended
sediment transport. J. Hydr. Res., 23(4), 359–377.
Ganju, N.K. 2001. Trapping organic-rich fine sediment in an estuary. MS Thesis, University
of Florida, Gaiesville, FL.
Harten, A. 1983. High resolution schemes for hyperbolic conservation laws, Journal of
Computational Physics. 49, 357–393.
Hayashi, T.S., Ozaki and Ichibashi, T. 1980. Study on bed load transport of sediment
mixture, Proc. 24th Japanese Conf. on Hydraulics, Japan.
Hwang, K.-N. 1989. Erodibility of fine sediment in wave dominated environments. MS

Thesis, University of Florida, Gainesville, FL.
Krone, R.B, 1962. Flume studies of transport of sediment in estuarial shoaling processes.
Final Report, Hydraulic Engineering Laboratory and Sanitary Engineering Research
Laboratory, University of California, Berkeley, CA.
Kahan, W. 1958. Gauss-Seidel methods of solving large systems of linear equations, PhD
thesis, University of Toronto.
Kuijper, C., Cornelisse, J.M., and Winterwerp, J.C. 1990. Erosion and deposition of natural
muds–sediments from the Western Scheldt (Breskens), Rep. 29, Rijkswaterstaat Delft
Hydraul. Cohesive Sed., Delft, Netherlands.
Leone, R.De., and Mangasarian, O.L. 1988. Asynchronous parallel successive overrelaxation
for the symmetric linear complementarity problem. Mathematical Programming, 42(1–
3), 347–361.
Liu, W.-C., Hsu, M.-H., and Kuo, A.Y. 2002. Modelling of hydrodynamics and cohesive
sediment transport in Tanshui River estuarine system, Taiwan, Marine Pollution Bulletin,
44. 1076–1088.
Marván, F.G. 2001. A two-dimensional numerical transport model for organic-rich cohesive
sediments in estuarine waters. PhD Thesis, Heriot-Watt University, Edinburgh, UK.
Nichols, M. 1984. Fluid mud accumulation processes in an estuary. Geo-Marine Letters, 4–5,
192–198.
Owen, M.W. 1970. A detailed study of the settling of an estuary mud. Report No IT 78, HR
Wallingford, Wallingford, UK.
Patankar, S.V. 1980. Numerical Heat Transfer and Fluid Flow. Hemisphere: Washington, DC,
197 p.
Partheniades, E. 1962. A study of erosion and deposition of cohesive soils in salt water.
Ph.D. thesis, University of California, Berkeley, 182 p
80
RD-?? Chapter 4 – References
Richardson, J.F. and Zaki, W.N. 1954. Sedimentation and fluidization, Part I. Transaction of
the Institute of Chemical Engineers., 32(1), 35–53.
Rubey, W.W. 1933. Settling velocities of gravel, sand, and silt particles, American Journal of
Science, 5th Series, 25(148), 325–338.
Saad, Y. 1993. A flexible inner-outer preconditioned GMRES algorithm. SIAM Journal

Scientific Computing (14): 461–469.
Schenk, O. and Gartner, K. 2011. Parallel Sparse Direct and Multi-Recursive Iterative Linear
Solvers: Pardiso User Guide 4.1.2.
Schenk, O. and Gartner, K. 2004. Solving unsymmetric sparse systems of linear equations
with PARDISO. Journal of Future Generation Computer Systems, 20(3):475–487.
Soulsby, R.L. 1997. Dynamics of marine sands. A manual for practical applications. Thomas
Telford Publications, London, England, 249 p.
Toffaleti, F.B. 1968. A procedure for computation of total river sand discharge and detailed
distribution, bed to surface, Technical Report No. 5, Committee on Channel
Stabilitzation, U.S. Army Corps of Engineers, November.
van Rijn, L.C. 1993. Principles of sediment transport in river, estuaries, coastal seas, and
oceans. International Institute for Infrastructural, Hydraulic, and Environmental
Engineering, Delft, The Netherlands.
Wooster, J.K., Dusterhoff, S.R., Cui, Y., Sklar, L.S., Dietrich, W.E., and Malko, M. 2008.
Sediment supply and relative size distribution effects on fine sediment infiltration into
immobile gravels, Water Resources Research, 44, W03424.
Wei, H., Nearing, M.A., Stone, J.J., Guertin, D.P., Spaeth, K.E., Pierson, F.B., Nichols, M.H.,
and Moffett, C.A. 2009. A new splash and sheet erosion equation for rangelands. Soil
Science Society of America Journal 73(4): 1386–1392.
Wu. W., Altinakar, M. and Wang, S.S.Y. 2006. Depth-average analysis of hysteresis
between flow and sediment transport under unsteady conditions, Int. J. Sediment Res.,
21(2), 101–112.
Wu, W., Wang, S.S.Y. and Jia, Y. 2000. Nonuniform sediment transport in alluvial rivers, J.
Hydr. Res., IAHR, 38(6), 427–434.
Wu, W. and Wang, S.S.Y. 2006. Formulas for sediment porosity and settling velocity,
Journal of Hydraulic Engineering, ASCE, 132(8), 858–862.
Zhu, J. 1992. On the higher-order bounded discretization schemes for finite volume
computations of incompressible flows, Computational Methods in Applied Mechanical
Engineering, 98, 345.
81

HEC-RAS 2D Sediment Users Manual (Beta) PDF

Uploaded by

Copyright:

Available Formats

HEC-RAS 2D Sediment Users Manual (Beta) PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

HEC-RAS 2D Sediment Users Manual (Beta) PDF

Uploaded by

Copyright:

Available Formats

HEC-RAS Two-Dimensional

Sediment Transport User’s

Approved for Public Release. Distribution Unlimited. RD-??

REPORT DOCUMENTATION PAGE Form Approved OMB No. 0704-

PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ORGANIZATION.

1. REPORT DATE (DD-MM-YYYY) 2. REPORT TYPE 3. DATES COVERED (From - To)

December 2020 Research Document

4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER

HEC-RAS Two-Dimensional Sediment Transport 5b. GRANT NUMBER

6. AUTHOR(S) 5d. PROJECT NUMBER

Alejandro Sanchez, CEIWR-HHT 5e. TASK NUMBER

5F. WORK UNIT NUMBER

U.S .Army Corps of Engineers

Institute for Water Resources RD-??

Hydrologic Engineering Center (CEIWR-HEC)

609 Second Street

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/ MONITOR'S ACRONYM(S)

11. SPONSOR/ MONITOR'S REPORT

12. DISTRIBUTION / AVAILABILITY STATEMENT

Approved for public release; distribution is unlimited.

13. SUPPLEMENTARY NOTES

15. SUBJECT TERMS

HEC-RAS; River Analysis System; HEC; Hydrologic Engineering Center; two-dimensional;

Standard Form 298 (Rev. 8/98)

Institute for Water Resources

Hydrologic Engineering Center

609 Second Street

(530) 756-8250 FAX

List of Figures................................................................................................ iii

List of Tables ................................................................................................. ix

Performing a 2D Sediment Transport Analysis .................................................... 2

2.1 Best Practices for a 2D Sediment Model ........................................... 2

2.1.1 Hydraulic Warm Up ......................................................................... 2

2.1.2 Computational Mesh ....................................................................... 3

2.1.3 Hydraulic Equations ........................................................................ 7

2.1.4 Time Step ..................................................................................... 8

2.2 Sediment Data ............................................................................. 9

2.2.1 Initial Conditions and Transport Parameters ......................................10

2.2.3 Sediment Boundary Conditions ........................................................19

2.2.4 User-Defined Grain Classes .............................................................23

2.2.5 Cohesive Options ..........................................................................26

2.2.6 Transport Methods.........................................................................30

2.2.7 Transport Function Calibration and Modification .................................37

2.2.8 2D Options ...................................................................................38

2.2.9 Bed Mixing Options ........................................................................42

2.2.10 Bed Gradations .........................................................................43

2.3 Viewing Sediment Results in RAS Mapper ....................................... 46

2.3.1 Plotting 2D Sediment Time Series ....................................................51

2.3.2 Plotting 2D Sediment Transects/Profile Lines .....................................52

2.4 Sediment Computation Options and Tolerances ............................... 54

2.4.1 Sediment Computation Options and Tollerances .................................55

2.5 Sediment Output Options ............................................................. 67

2.5.1 Output Level .................................................................................69

2.5.2 Output Interval Multiples ................................................................71

2.5.3 Customized Sediment Output Variables ............................................71

2.5.4 Sediment Hotstart .........................................................................75

2.6 Viewing Results........................................................................... 75

2.6.1 Exploring the HDF5 File ..................................................................75

2.6.2 Log Output ...................................................................................76