Professional Documents
Culture Documents
1.1.1 Overview
In this tutorial group, you will learn how HEC-HMS is structured, and you will learn fundamental mechanics of the
program including creating data sets, computing simulation runs, and viewing results. This tutorial will provide you
with the necessary background for subsequent tutorials.
This tutorial group includes the following components:
1. Creating a New Basin Model(see page 2)
2. Creating a New Meteorologic Model(see page 10)
3. Creating New Control Specifications1
4. Creating and Computing a Simulation Run(see page 17)
5. Viewing Results2
6. Viewing Observed Flow(see page 22)
1.1.2 Background
The purpose of these grouped tutorials is to illustrate how to develop a simple HEC-HMS model from scratch. The
defined procedure does not make use of the GIS capabilities within HEC-HMS; the procedure shows how to create a
non-georeferenced project. This tutorial group is broken up to show how to create a basin model, create a
meteorologic model, create a control specification, create a simulation run, and how to visualize model results.
The Punxsutawney watershed, a part of the Allegheny River Basin located in western Pennsylvania, is used for
this and subsequent tutorials in this course. The primary conveyance system of this 158 sq. mi. watershed includes
the following streams: Stump Creek, East Branch Mahoning Creek, and Mahoning Creek. The confluence of
Stump Creek and East Branch Mahoning Creek is located east of the enclave of Big Run. Mahoning Creek is
downstream of the confluence.
Storm events in April 1994, May 1995, and May 1996 in this region are considered in this model. A number of
recording rainfall and stream flow gages in the region will be useful in modeling the basin response to various
ranges in rainfall intensity. The watershed will be modeled as 3 subbasins with incremental precipitation from
recording rainfall gages; user-specified gage weighting and inverse-distance weighting methods will be used. Initial
subbasin and reach parameters will be computed based on measured physical parameters and maps, where
available, and qualitative engineering judgment with emphasis on local knowledge of the region.
Initial modeling parameters will be refined through several calibration runs. After the calibration stage, the model
will adopt hydrologic parameters based on how well the simulated results reconstitute observed flow hydrographs.
The rectangular box in the first figure shows the study area location. The second figure shows the study area
including Stump Creek, East Branch Mahoning Creek, and Mahoning Creek at the town of Punxsutawney.
1 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsguides/creating-a-simple-model/creating-a-new-control-specifications
2 https://www.hec.usace.army.mil/confluence/display/HMSGUIDES/Viewing+HEC-HMS+Model+Results
1.2.1 Overview
In this tutorial you will create a new basin model consisting of three subbasins, two junctions, and a routing reach;
then you will connect the elements to form a hydrologic network, and finally enter parameter data for these
hydrologic elements.
1.2.2 Steps
1. Open HEC-HMS.
2. From the File | New menu, create a new project.
3. Name the project "Punxsutawney". Set the default unit system to U.S. Customary.
4. From Components on the menu bar, select the Basin Model Manager option. The Basin Model Manager
window will open.
5. Click the New… button to begin the process of creating a basin model. The Create A New Basin Model
window will open.
6. Change the default basin model name (Basin 1) to Punxsutawney and add the description Headwaters to
the City of Punxsutawney - Existing Conditions.
7. Click the Create button when all the information is correct. The new, empty basin model will be added to
the Watershed Explorer. Close the Basin Model Manager.
8. Set the default methods for subbasins and reaches that will be used when creating new elements. From the
Tools menu, select the Program Settings … option.
9. Select the Defaults tab and set the default loss method to Initial and Constant, the default transform
method to Clark Unit Hydrograph, the default baseflow method to Linear Reservoir, and the default
routing method to Lag. Click the OK button when you have finished making the selections.
10. Open the basin model by clicking on it in the Watershed Explorer.
11. Click the subbasin creation tool icon in the components toolbar.
12. Click on the basin map to create the first subbasin element. You can use the default name.
13. Add two more subbasin elements.
14. Switch to the reach creation tool and click on the basin map to create a reach element.
15. Switch to the junction creation tool and add two junction elements to the basin map. The elements
should be placed as shown in the figure below.
16. Connect the elements into a hydrologic network as shown in the figure below. Right click on each element
and select the Connect Downstream option. Use the mouse to identify the element that should be
downstream. A connection line is drawn to show the elements are connected.
Question: In this step modeling elements are graphically connected. What is another way to connect
elements?
Answer
Elements can be connected from the component editor for a given element by selecting a downstream
element.
17. Click on the Subbasin-1 icon in the Watershed Explorer and rename it to Stump Creek.
18. Similarly, click on the other subbasin elements and rename them too; Subbasin-2 becomes EB Mahoning
Creek and Subbasin-3 becomes Mahoning Creek Local as shown in the figure below.
19. Enter the area for each of the subbasins. First select basin model, then Click on the Parameters menu and
select the Subbasin Area option. Enter the area as shown in the table below
Name Area (mi2)
EB Mahoning Creek 41
Stump Creek 29
20. Set up Baseflow Parameters. From the Parameters menu, select the Baseflow and then Linear Reservoir
option.
21. Enter the values from the table below, then click the Close button when finished. Be sure to select the
correct type of initial condition: Discharge Per Area. Also be sure to select the correct number of layers:
Subbas Numb Initial GW 1 GW 1 GW 1 GW 1 GW 2 GW 2 GW 2 GW 2
in er of Type Initial Fracti Coeffi Step Initial Fracti Coefic Step
Layers on cient s on ient s
22. Set up Loss Parameters. From the Parameters menu, select the Loss and then Initial and Constant
option. Enter the values from the table below, then click the Close button when finished.
Subbasin Initial Loss (in) Constant Rate (in/hr) Impervious (%)
Input the Initial loss, Constant rate and Impervious parameters provided in the table above
23. From the Parameters menu, select the Transform and then Clark Unit Hydrograph option. Enter the
values from the table below, then click the Close button when finished.
Subbasin Time of Concentration (hr) Storage Coefficient (hr)
EB Mahoning Creek 7 11
Stump Creek 9 12
Input Time of Concentration and Storage Coefficients provided in the table above
Question: In steps 20-23 modeling methods are parameterized from the Parameters menu. What is an
alternative way to parameterize modeling methods in the HEC-HMS user interface?
Answer
The component editor is an alternative way to parameterize the model. Globally editing tends to be much
faster when parameterizing multiple elements of the same modeling method.
24. On the basin map in the desktop area, click the Reach-1 element to open its Component Editor. Click on
the Routing tab and then enter a lag value of 300 minutes.
1.2.3 Summary
You have now completed steps for adding hydrologic elements to a basin model, connected the hydrologic
elements to form a hydrologic network, and parameterized the methods used to simulate hydrologic processes.
The project below was created by following the steps in this tutorial.
1.3.2 Overview
In this tutorial you will create a meteorologic model. You will link the meteorologic model to a basin model and
then parameterize a method for computing basin average precipitation. This tutorial only shows one of the
precipitation methods in HEC-HMS. There are many additional methods for computing basin average rainfall,
applying gridded precipitation, and for simulating evapotranspiration, and snowmelt processes.
1.3.3 Steps
1. If you are starting this tutorial as a continuation of the previous tutorial, unzip the data folder into your
project folder and continue. Otherwise, open the HEC-HMS project named Punxsutawney.
2. Create a new precipitation gage. Go to the Components menu and select the Time-Series Data Manager
option. In the manager window, press the New… button to create a gage; the window for creating a gage
will open.
3. In the new gage window, change the default name (Gage 1) to DUJP. Press the Create button to create the
new gage; it will automatically be added to the Watershed Explorer. You can leave the manager window
open since it will be used again shortly.
3 https://www.hec.usace.army.mil/confluence/download/attachments/25133272/Final_Create_New_Basin.zip?
api=v2&modificationDate=1614798049143&version=5
4 https://www.hec.usace.army.mil/confluence/download/attachments/25133293/Initial_Create_New_Meteorologic.zip?
api=v2&modificationDate=1614795850875&version=5
5 https://www.hec.usace.army.mil/confluence/download/attachments/25133293/data.zip?api=v2&modificationDate=1608738093655&version=2
4. Set up data source : Go to the Watershed Explorer and click on the new gage you just created. Change the
data source from Manual Entry to Single Record HEC-DSS (HEC-Data Storage System).
5. Select the Correct External Data Source. You can click on the select button next to the filename field to
navigate to the file. The file is located in the data subdirectory of the project: /Punxsutawney/data/
observe.dss.
6. Select the Correct Pathname. Click on the select button next to the pathname input field to navigate to the
pathname. Use the B-Part filter in the search By Parts area near the top of the screen. Select the DUJP entry
in the B-part filter to show only pathnames with the correct pathname part shown in the figure below.
7. Change the Default Time Window to inspect some of the data. In the Watershed Explorer, click on the time
window under the DUJP gage icon. In the Component Editor, change start date to 28Apr1996, the start
time to 01:00, the end date to 04May1996, and the end time to 00:00. Click on the "Table" and "Graph" tabs
in the Component Editor to see the data.
8. You have finished setting up the DUJP time-series precipitation gage. We will use the gage later by referring
to it by name. Now repeat steps 2 to 7 to create the PNXP and MFFP gages. All of the precipitation data is
now ready to use. You can commence with building the meteorologic model.
9. Create a New Meteorologic Model. Click on the Components menu and select the Meteorologic Model
Manager option; the meteorologic model manager window will open. Press the New… button to create a
new meteorologic model.
10. Change the default name (Met 1) to May 1996 and enter a description of Precipitation for the Storm of May
1st, 1996. Press the Create button to create the new meteorologic model.
11. Go to the Watershed Explorer and click on the new meteorologic model. Switch to the
Component Editor and make sure the precipitation method is set to Specified Hyetograph. The
evapotranspiration and snowmelt methods should be set to --None--.
12. Connect the meteorologic model to the subbasins in the basin model. Go to the Basins tab in the
Component Editor. For the Punxsutawney basin model, change the Include Subbasins choice to Yes.
13. Select a precipitation gage for each subbasin: Return to the Watershed Explorer and click on the
precipitation node under the meteorologic model. Referencing the table below, select the correct gage for
each subbasin using the Component Editor.
You do not need to enter total depth values for the subbasins.
Subbasin Precipitation Gage
1.3.4 Summary
In this tutorial you created a meteorologic model for the project. The meteorologic model uses the specified
hyetograph precipitation method. Precipitation gages were added to the project and linked to the meteorologic
model.
The project below was created by following the steps in this tutorial.
1.4.2 Overview
In this tutorial you will create a control specifications that sets the time window and time interval for a simulation.
1.4.3 Steps
1. Open the HEC-HMS project named Punxsutawney.
2. Create a new control specifications: Click on the Components menu and select the Control
Specifications Manager option. The manager window will open. Press the New… button to create a new
control specifications.
3. Change the default name (Control 1) to May 1996 and enter a description of Storm of May 1st 1996. Press
the Create button to create the new control specifications.
4. In the Watershed Explorer, click on the control specifications you just created. Enter the correct start and
end times in the Component Editor. The start date should be 28Apr1996 at 01:00. The end date should be
04May1996 at 00:00. Select 15 Minutes for the time interval as shown in the figure below.
Question: The control specifications specify a time interval of 15 minutes. What will happen if
precipitation data in the meteorologic model is hourly?
Answer
HEC-HMS will equally distribute the hourly precipitation total over 15 minute intervals. If the hourly total
was 1 inch, each 15-minute interval would be 1/4 of an inch.
6 https://www.hec.usace.army.mil/confluence/download/attachments/25133293/Final_Create_New_Meteorologic.zip?
api=v2&modificationDate=1614796003723&version=4
7 https://www.hec.usace.army.mil/confluence/download/attachments/25133300/Initial_Create_Control_Specs.zip?
api=v2&modificationDate=1614796193988&version=3
1.4.4 Summary
In this tutorial you control specifications were created that will dictate the duration and time interval of a
simulation.
The project below was created by following the steps in this tutorial.
1.5.2 Overview
In this tutorial you will create and compute a simulation run.
1.5.3 Steps
1. Open the HEC-HMS project named Punxsutawney.
2. Click on the Compute menu and select the Create Compute → Simulation Run... option. A wizard window
will open to guide you in creating a new simulation run. In the first step you must choose a name; Use May
1996 instead of the default.
3. In second step, choose the Punxsutawney basin model. In third step, choose the May 1996 meteorologic
model. In Step 4, choose the May 1996 control specifications. Press the Finish button to complete the
process of creating a simulation run.
4. Compute the simulation run: A simulation run must be selected before it can be computed. The tool bar
includes a selection list that shows all of the simulation runs that have been created in the project. Click on
the selection list and choose RUN: May 1996. Once a simulation run has been selected, click on the
Compute button immediately to the right of the selection list to perform the compute. The tool bar
button has an icon of a raindrop whenever a simulation run is selected. A compute progress window will
open to show the advancement of the simulation. The simulation may abort if errors are encountered. If this
happens, read the messages and fix any problems; then compute the simulation run again. Close the
progress window when the run computes successfully.
8 https://www.hec.usace.army.mil/confluence/download/attachments/25133300/Final_Create_Control_Specs.zip?
api=v2&modificationDate=1614796337873&version=3
9 https://www.hec.usace.army.mil/confluence/download/attachments/25133307/Initial_Create_Compute_Simulation.zip?
api=v2&modificationDate=1614796454557&version=4
Question: In this step the simulation was computed from the Compute toolbar. What is another way to
compute the simulation?
Answer
You can select the Compute Run option at the bottom of the Compute menu (the name of the selected run
will be shown within brackets next to the Compute Run menu option). You can compute from the Compute
tab by right-clicking a simulation node and selecting Compute. Simulations can be computed from global
editors whenever a simulation is selected in the compute toolbar.
1.5.4 Summary
In this tutorial you created and computed a simulation run. The simulation run is composed of a basin model,
meteorologic model, and control specifications.
The project below was created by following the steps in this tutorial.
1.6.2 Overview
In this tutorial, you will view graphical, time-series, and summary results of subbasins, reaches, and junctions.
10 https://www.hec.usace.army.mil/confluence/download/attachments/25133307/Final_Create_Compute_Simulation.zip?
api=v2&modificationDate=1608739999402&version=3
11 https://www.hec.usace.army.mil/confluence/download/attachments/25133313/Initial_View_Results.zip?
api=v2&modificationDate=1614796752307&version=2
1.6.3 Steps
1. Open the HEC-HMS project named Punxsutawney.
2. Compute the simulation run: May 1996.
3. To view the global summary table, click on the Results menu and select the Global Summary Table.
Alternately, you can click the View Global Summary Table button, , on the toolbar; the button shows a
table with a globe superimposed on it.
4. To view results for each basin model element, click on an element in the basin map with the right mouse
button. A context menu is displayed with several choices including View Results. Under View Results you
can see a graph, summary table or time-series table. You can obtain the same results by selecting the
element, then clicking on the Results menu and selecting the appropriate option.
5. To plot a graph for the Stump Creek subbasin, click on the Stump Creek subbasin in the basin map with the
right mouse button, then click the View Results menu and select the Graph option.
The time-series shown in a graph are listed in the legend. The graph includes the incremental precipitation
(blue in upper plot) and the precipitation loss (red in upper plot). Incremental precipitation is computed by
the meteorologic model for each subbasin. The precipitation excess is the incremental precipitation minus
the losses computed by the selected loss method. The graph also includes the baseflow (brown in lower
plot) and the total subbasin outflow (blue in lower plot).
6. Summary statistics are automatically computed and shown in the summary table. Click on the Stump Creek
subbasin in the basin map with the right mouse button, then click the View Results menu and select the
Summary Table option.
The summary table shows that the peak discharge is 553.4 cfs. The table also shows that the peak flow
occurred on 30 April 1996 at 14:15 hours. When the table initially opens, it shows the total discharge volume
is 1.60 inches as shown in the figure below. By clicking on the AC-FT radio button, the discharge switches
units. The discharge volume is equivalent to 2478.8 AC-FT as shown in the second image below.
Question: What is the peak discharge for subbasin Mahoning Creek Local?
Answer
1786.0 cfs
7. The computed discharge at a specific time is best determined from the time-series table for the element.
Click on the Stump Creek subbasin in the basin map with the right mouse button, then click the View
Results menu and select the Time-Series Table option. The image below shows a time-series results table.
Question: What is the baseflow for subbasin EB Mahoning Creek on May 1, 1996 at 12:00?
Answer
333.1 cfs
8. To plot results for Reach-1, as shown in the figure below, click on the Reach-1 reach in the basin map with
the right mouse button, then click the View Results menu and select the Graph option. The time-series
shown in a graph are listed in the legend. The graph includes the combined inflow to the reach (dashed blue)
and also the computed outflow (solid blue). No matter how many elements connect to the upstream side of
the reach, only the combined inflow is shown.
9. To plot results for Junction-2, click on the Junction-2 junction in the basin map with the right mouse button,
then click the View Results menu and select the Graph option. The time-series shown in a graph are listed in
the legend. The outflow from the Punxsutawney Local subbasin (dashed blue) is shown along with the
outflow from Reach-1 (dotted blue). If an observation time series is available at the location, a dotted black
line is displayed. The total outflow is also shown (solid blue). All inflows into the junction from elements
immediately upstream will be shown along with the total outflow, but no combined inflow is shown. Only
1.6.4 Summary
In this tutorial graphical, time-series, and summary results of subbasins, reaches, and junctions were
demonstrated.
The project below was created by following the steps in this tutorial.
1.7.2 Overview
In this tutorial you will add observed flow and view results for a hydrologic element.
1.7.3 Steps
1. From the Components menu, select Time-Series Data Manager
2. For Data Type, select Discharge Gages
3. Select New... to create a new discharge gage
4. Name the gage PNXP and create
12 https://www.hec.usace.army.mil/confluence/download/attachments/25133313/Final_View_Results.zip?
api=v2&modificationDate=1614796841699&version=6
13 https://www.hec.usace.army.mil/confluence/download/attachments/43819676/Initial_View_Observed.zip?
api=v2&modificationDate=1614796937472&version=2
5. Open the component editor for the discharge gage PNXP. For DSS Filename, select /Punxsutawney/data/
observe.dss. For the DSS pathname, select the FLOW record at location PNXP.
1.7.4 Summary
In this tutorial an observed flow record was as a discharge gage, then linked the discharge gage to a basin element
as observed flow. Observed flow was viewed in relation to other element results.
The project below was created by following the steps in this tutorial.
14 https://www.hec.usace.army.mil/confluence/download/attachments/43819676/Final_View_Observed.zip?
api=v2&modificationDate=1614797034339&version=2
2.1.1 Overview
In this tutorial group you will use the HEC-HMS GIS tools to create an HEC-HMS basin model. This tutorial group
contains the following components:
• Creating and linking a terrain dataset to a basin model(see page 28)
• Applying the terrain preprocessing tools to develop an initial watershed delineation(see page 33)
• Refining the watershed delineation through merging and splitting elements(see page 39)
• Renaming hydrologic elements(see page 49)
There are questions at the end of each tutorial.
2.1.2 Background
In this tutorial group, you will be working with a terrain model of the Mahoning Creek watershed. The terrain data
was downloaded from the USGS National Map Viewer. After the data was downloaded, it was projected to the
Albers equal area projection, the horizontal units were set to feet. The terrain data was checked for any missing
data grid cells and the vertical units were set to feet as well. Additional GIS layers obtained for this example include
a topographic map, a shapefile showing published streams, and shapefiles showing USGS stream gage and
precipitation gage locations.
The Mahoning Creek watershed at Punxsutawney, PA is part of the Allegheny River Basin located in western
Pennsylvania. The primary conveyance system of this 158 sq. mi. watershed includes the following streams: Stump
Creek, East Branch Mahoning Creek, and Mahoning Creek. The confluence of Stump Creek and East Branch
Mahoning Creek is located east of the enclave of Big Run. Mahoning Creek is downstream of the confluence. The
watershed will be modeled as 3 subbasins. The figures below show the study area.
2.2.1 Overview
In this tutorial you will create a new HEC-HMS project, create a terrain dataset, and load the terrain dataset into a
Basin Model. As shown in the steps below, best practice is to copy all external files, including GIS data layers, into
the HEC-HMS project directory.
2. Use File Explorer to copy the GIS Data folder (contained in the "Initial" zip file) into the …
\Punxsutawney\maps directory. Now, the external GIS files are local in the project directory.
3. From the Components menu, select Terrain Data Manager.
15 https://www.hec.usace.army.mil/confluence/download/attachments/41418774/Initial.zip?api=v2&modificationDate=1606157265364&version=1
4. Within the Terrain Data Manager, Click New… and name the terrain data "NED_10m".
5. In the next step, select the terrain file and choose the vertical units. Within the Filename entry, select the
folder icon and navigate to the "terrain_ft_albersft.tif" file (should be in the …\Punxsutawney\maps\GIS
Data folder). Set the Vertical Units to Feet.
6. Click the Finish button. A new element is added to the Watershed Explorer, in the Terrain Data folder.
Model.
2. Within the Basin Model Manager, click New… and name the basin model Mahoning Creek.
3. From the Watershed Explorer, expand the Basin Model node and select the Mahoning Creek basin model.
In the Basin Model's Component Editor, select NED_10m for the Terrain Data.
5. You will be prompted to select a coordinate system for the Mahoning Creek Basin Model.
Press Skip and the program will set the Basin Model's coordinate system based on the terrain data.
6. The terrain data will now appear in the map window. Note: To access map layer controls (e.g. visibility,
color scales, etc.), select View | Map Layers, highlight the desired layer, and select Draw Properties. This
will allow display modifications to be made to the selected layer.
7. Check the coordinate system of the basin model by selecting the GIS | Coordinate System menu option.
You cannot change the Basin Model's coordinate system once it has been set.
8. Load additional map layers into the basin model. Select the View | Map Layer menu option to open the Map
Layers window. Use the Add… button to select the rain_gages_albersft, NHD_River_albersft, and
flow_gage_albersft shapefiles.
9. Modify the draw properties for the three shapefiles so the Basin Model Map looks like the following figure.
2.2.4 Questions:
1) What happens if the basin model coordinate system is already set before the terrain dataset is linked/added to
the basin model?
Answer
HEC-HMS will project the terrain data into the basin model's projection. All GIS layers created during the
delineation process will be created in the basin model's projection.
2.3.1 Overview
In this tutorial you will use the GIS delineation tools to delineate subbasin and reach elements. The delineation will
be based on the stream threshold. The next tutorial shows how to customize the delineation using split and merge
tools.
16 https://www.hec.usace.army.mil/confluence/download/attachments/41418774/Final.zip?api=v2&modificationDate=1606157525953&version=1
2. From the GIS menu select Preprocess Sinks. This tool creates two new layers, Sink Fill and Sink Locations.
These layers will be automatically added to the Basin Model Map. You can turn them off from the Map Layers
dialog (select View | Map Layers to open the Map Layers dialog). The Sink Locations dataset shows the
amount of fill required for water to move from cell to cell (shows how much fill is required to remove pits in
the terrain). The Sink Fill dataset is the hydrologically correct terrain dataset and is used for delineation.
3. From the GIS menu select Preprocess Drainage. This tool creates two new layers, Flow Direction and Flow
Accumulation. The flow direction dataset shows the direction of flow from one grid cell to the next. The flow
accumulation grid shows the number of upstream grid cells. Zoom-in to a part of the watershed to display
the details of the grid cells that make up the flow accumulation grid. Each grid cell value is computed by
summing the total number of grid cells that flow into the grid cell. Grid cells with a value of 0 are located on
17 https://www.hec.usace.army.mil/confluence/download/attachments/41418806/Initial.zip?api=v2&modificationDate=1614732491633&version=2
the topographic divide and grid cells with a large value make up the stream network.
4. From the GIS menu select Identify Streams. A pop-up window will prompt you to specify a threshold for
stream delineation. This value is the minimum drainage area for the program to identify a stream. Smaller
values will result in more subbasins and reaches. For this tutorial, enter 5 square miles as the threshold
value. This step creates a new Identified Streams layer.
You might need to turn off the NHD_River_albersft shapefile to see the Identified Streams layer in the Basin
Model Map. The "Diverging BlueBrown" color ramp might make the identified streams show up better in the
map.
layer. This tutorial will use the Mahoning Creek at Punxsutawney stream gage to define the outlet. The
Mahoning Creek at Punxsutawney stream gage is contained in the flow_gage_albersft map layer. Make sure
this layer is turned on (turn off the rain_gages_albersft map layer). Make sure the Identified Streams layers
is visible.
2. At least one break point must be added to the basin model. The program uses the most downstream break
point as the outlet of the watershed. The break point must be located on top of a grid cell within the
Identified Streams layer. To place a break point, select the Break Point Creation Tool on the toolbar,
. Place a break point on the identified stream next to the flow_gage_albersft stream gage. Name the break
point Outlet.
3. From the GIS menu select Delineate Elements to delineate the watershed. A Delineate Elements Options
dialog will open with options for setting prefixes for the element names. Choose "Subbasin" for subbasin
elements, "Reach" for reach elements, and "J" for junction elements (make sure the Insert Junctions
option is set to Yes). Finally, set the Convert Break Points option to Yes. This option will automatically
assign computation points to the break points. Click the Delineate button. Subbasins, reaches, and
junctions will be generated and automatically assigned names.
4. The delineated watershed should look like the following figure. The subbasin and reach network is based on
a stream threshold of 5 square miles.
2.3.4 Questions:
1) What is the largest cell value from the Flow Accumulation grid? What is the total drainage area for the largest
Flow Accumulation cell (cells are 34.31ft * 34.31ft)?
Answer
The largest flow accumulation value is 3887288. This value represents a drainage area of 3887288 * 34.31 ft *
34.31 ft / 5280 (ft/mi) / 5280 (ft/mi) = 164.14 square miles.
2) Why might you add multiple break points.
Answer
The most downstream break point is required to identify the outlet of the watershed. Upstream break points
can be used to add specific subbasin outlets. Flow gage locations, boundary conditions for other models, or
analysis points could be reasons to add additional break points.
3) How many Subbasins are there in the basin model? Why do some of the subbasins have a drainage area larger
than 5 square miles?
Answer
There are 17 subbasins when using a default stream threshold of 5 square miles. The stream threshold does
not necessary translate to subbasin area. HEC-HMS will only identify streams that exceed a drainage area of 5
square miles. Subbasins are then delineated for each stream segment. Long, skinny watersheds might not
have any tributaries that exceed the drainage area threshold; therefore, there could be a large area with no
reach segments. Look at the Stump Creek watershed. The Identified Stream layer has a stream segment
starting at 5 square miles, but there are no additional reach segments until the Confluence with the East
Branch Mahoning Creek.
2.4 Using Merge and Split Tools to Customize Subbasin and Reach
Delineation
2.4.1 Overview
In this tutorial you will customize the subbasin and reach delineation that was created by the automated
delineation. Both the merge and split tools will be used.
18 https://www.hec.usace.army.mil/confluence/download/attachments/41418806/Final.zip?api=v2&modificationDate=1606162259411&version=1
19 https://www.hec.usace.army.mil/confluence/download/attachments/41418818/Initial.zip?api=v2&modificationDate=1614732753100&version=2
downstream so that the outlet for the merged sub-basin is the same as the lower subbasin.
1. Open the Punxsutawney project and then open the Mahoning Creek Basin Model.
2. To merge subbasins or reaches, select the first element by clicking on it with the arrow tool, . Holding
down the SHIFT key, select the second element.
3. From the GIS menu select Merge Elements. Only two elements can be merged at one time. Continue to
merge all subbasins within the Stump Creek, Mahoning Creek and Mahoning Creek Local subbasins. After
merging the subbasins, merge river reaches within the Mahoning Creek Local subbasin so that there is only
one reach. You might have a few unconnected junctions left over when merging subbasin elements. Delete
any unconnected junction. When finished, there should be one subbasin and one reach below the
directory. The following figure shows the location of the two flow gages, Point 1 and Point 2.
1. Make sure the Identified Streams layer is visible and zoom into Point 1 in the Basin Model Map. You will
notice this point falls on top of the existing Identified Streams layer. Using the Arrow tool, select the Stump
Creek subbasin element (Subbasin1) and then select the GIS | Split Element menu option. You will notice
your mouse curser turns into a cross hair. Click on top of the Point 1 location, which should be located within
an identified streams grid cell. Make sure you are zoomed in far enough to click within an identified stream
grid cell. A message dialog will open asking if you want to proceed with the delineation. Click Accept.
2. Zoom into Point 2. You will notice Point 2 does not fall on top of the Identified Stream layer. You will not be
able to delineate a subbasin at this point because the Identified Stream layer does not extend to this
location.
3. Re-run the Identify Streams tool and set the threshold to 2 square miles. As shown below, the Identified
Stream layer extends through Point 2 when the stream threshold is 2 square miles.
Note
Re-running the identify streams tool will not automatically change the existing delineation. The
following message pops up when re-running the identify streams tool. The message states only
new or additional delineations will be impacted by the new identified streams, which includes re-
running the Delineate Elements or Split tools.
4. Use the Split Elements tool to add a subbasin outlet at Point 2. The program will delineate a new subbasin
with an outlet at Point 2 but will also delineate additional subbasins for each of the reach segments within
the selected subbasin.
5. Merge the extra subbasins and reaches in order to re-create the delineation shown below. After merging
elements, there should be one subbasin upstream of Point 2, one subbasin between Point 1 and Point 2, and
2.4.4
Question:
1) Is there a faster way to reduce the number of elements than merging two elements at a time?
Answer
You can re-run the Identify Stream tool and use a larger stream threshold. Then you would re-run the
Delineate Elements tool. The program will re-delineate the subbasin and stream network each time the
Delineate Elements step is run.
2.5.1 Overview
In this tutorial you will complete the initial basin model setup by refining element names, checking element
connectivity, and verifying drainage area.
a. Rename the top-left subbasin Stump Creek, the bottom-right subbasin East Branch Mahoning
Creek, and the subbasin below the confluence Mahoning Creek Local.
b. Rename the reach Mahoning Creek.
c. Rename the junction at the confluence of Stump Creek and East Branch Mahoning Creek Big Run.
20 https://www.hec.usace.army.mil/confluence/download/attachments/41418818/Final.zip?api=v2&modificationDate=1606162397229&version=1
21 https://www.hec.usace.army.mil/confluence/download/attachments/41418834/Initial.zip?api=v2&modificationDate=1614733182778&version=2
d. Rename the sink element at the outlet Mahoning Creek Gage at Punx.
3. All elements should already be connected to one another and the elements can be sorted hydrologically in
the Watershed Explorer. Right click on top of the Basin Model name in the Watershed Explorer and choose
the Resort Elements Hydrologically option. Elements will be sorted from upstream to downstream.
The table shows the element connections (you can see this information within the element's
Component Editor too). Sorting the elements hydrologically will confirm element connectivity is correct.
2.5.4 Questions:
1) Can two basin model elements have the same name?
Answer
No, it is not possible to have elements with the same name in HEC-HMS. Each element name must be unique.
2) What is the character limit for basin model element names?
Answer
You can use up to 28 characters when naming Basin Model elements. Use as many characters as possible to
define descriptive names.
3) What is a reasonable level of precision for model parameters?
Answer
Most hydrologic model parameters should be rounded to 1 or 2 places after the decimal. Some model
parameters, like subbasin or stream slope, will need more places after the decimal. The main point is that
model parameters can be estimated with significant precision from GIS datasets; however, the GIS datasets
were created with assumptions that do not support a large amount of precision. Also, the model parameters
developed from GIS datasets and other data sources should be treated as initial parameter estimates and
refined during model calibration and validation.
22 https://waterdata.usgs.gov/usa/nwis/uv?03034000+
23 https://www.hec.usace.army.mil/confluence/download/attachments/41418834/Final.zip?api=v2&modificationDate=1606162476978&version=1
3.1.1 Background
Establishing initial parameter estimates (i.e. loss, transform, and baseflow parameters) for large hydrologic
models can be an arduous task. There are several useful tools in HEC-HMS that can simplify the process. HEC-HMS
has the capability to calculate various subbasin and reach characteristics based on the terrain dataset associated
with the model. These characteristics are useful for quantifying physical traits of subbasins and reaches; they can
serve to identify homogenous characteristics/groupings of subbasins in a model, as well as to point out their
differences. Subbasin characteristics, along with common raster datasets can be accessed via the Expression
Calculator in HEC-HMS. Regression equations based on GIS traits such as river lengths, river slopes, land use
classification, soil types, etc. can be entered into the Expression Calculator to estimate initial transform, loss, and/or
baseflow parameters. These initial estimates are automatically populated into the global editors.
3.1.2 Overview
In this workshop, you will use HEC-HMS to compute subbasin and reach characteristics and you will use the built-in
Expression Calculator for parameter estimation. This workshop includes two components: 1) Computing Subbasin
and Reach Characteristics(see page 54) and 2) Estimating Parameters with GIS Datasets(see page 58).
3.1.3 Goals
In this workshop, you will:
• Learn how to compute subbasin and reach characteristics in HEC-HMS.
• Learn how to use the Expression Calculator in HEC-HMS to estimate transform and loss parameters.
3.1.4 Data
You will be using the Punxsutawney Watershed HEC-HMS model for this workshop. All model and parameter
components other than transform and loss parameters have been configured for you.
3.2.1 Overview
This tutorial will guide you through the steps of computing subbasin and reach characteristics within HEC-HMS.
3.2.2 Data
24 https://www.hec.usace.army.mil/confluence/download/attachments/43811618/Punx_BasinCharacteristics.zip?
api=v2&modificationDate=1614738337895&version=2
c. HEC-HMS will calculate various characteristics for each subbasin in the basin model.
2. For a quick overview of each of the subbasin characteristics, see the Basin Characteristics25 page of the HEC-
HMS Users Manual.
25 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsum/latest/geographic-information/basin-characteristics
b. HEC-HMS will calculate a few reach characteristics for each of the reaches in the basin model. In this
simple model, there is only one reach.
2. For a quick overview of the above reach statistics, see the Basin Characteristics26 page of the HEC-HMS Users
Manual.
26 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsum/latest/geographic-information/basin-characteristics
3.2.5 Questions
1. When comparing the longest flowpath, centroidal flowpath, and 10-85 flowpath slopes, which slope is the
steepest? Why might this be?
Click for answer...
For most subbasins, the longest flowpath slope is typically the steepest. Since longest flowpaths extend
from the subbasin outlet all the way to the subbasin divide, the higher elevations near the divide can result
in a steeper computed slope that is not necessarily representative of the other portions of the flowpath. For
this reason, the 10-85 slope is often more representative of the general flowpath slopes as seen throughout
most of the subbasin.
2. Which of the subbasin characteristics is an indicator of the general shape of a subbasin?
Click for Answer...
Elongation Ratio. . Given a circle with the same area as
the subbasin of interest, the elongation ratio can be thought of as the diameter of the circle (DC) divided by
the length of the longest flowpath (LLFP).
3. Subbasin A and Subbasin B both experience the same uniform rain event. Given only the below
characteristics, which subbasin would likely experience the higher peak discharge?
Longest Relief Elongation
Flowpath Ratio Ratio
Length
(MI)
the drainage area of Subbasin B is substantially larger than Subbasin A. If exposed to the same uniform rain
event, Subbasin B would receive more precipitation volume overall.
Arguments for Subbasin A → The Longest Flowpath is slightly shorter for Subbasin A meaning that water has
a shorter path to travel to get from the most hydraulically remote part of the watershed down to the outlet.
The Relief Ratio is larger for Subbasin A meaning that it has a greater relative elevation difference from the
subbasin divide to the subbasin outlet.
3.3.1 Overview
This tutorial will guide you through the steps of using the Expression Calculator in HEC-HMS to estimate initial
transform and loss parameters. Transform parameters will be estimated using Subbasin Characteristics and loss
parameters will be estimated using gridsets. Currently, the Expression Calculator tool can be launched from the
below global editors. The tool will be added to other global editors in the future.
• Loss: Deficit and Constant
• Loss: Green and Ampt
• Transform: Mod Clark
• Transform: Clark
• Transform: S-Graph
• Baseflow: Linear Reservoir
For this tutorial, the gridded data has already been prepared for you. A separate tutorial demonstrates how to use
the Grid Data Manager to import various GIS raster data for use in HEC-HMS.
3.3.2 Data
If you just completed the Computing Subbasin and Reach Statistics Tutorial, you may continue to use the same
model files.
27 https://www.hec.usace.army.mil/confluence/download/attachments/43811406/Punx_EstimatingParameters.zip?
api=v2&modificationDate=1614738478141&version=2
b. The Transform method chosen for this model is ModClark. Open the Global Editor for ModClark by
selecting Parameters→ Transform→ ModClark.
c. The ModClark parameters are undefined; initial estimates need to be entered. The Expression
Calculator is an easy way to quickly populate initial parameter estimates for each subbasin.
Select Calculator... to launch it.
d. The Expression Calculator can access two types of data: Grids (external grids that have been imported
into HEC-HMS via the Grid Data Manager) and Stats (the computed Basin Characteristics values). We
will use the Stats tab to compute initial estimates for Time of Concentration. In the Expression
Calculator, ensure that the Field: Time of Concentration is selected and then select the Stats tab to
access the Basin Characteristics variables.
e. For this tutorial, we are going to use the below equation to estimate Time of Concentration which
was developed for the modeling area of interest. This equation establishes a relationship between
Longest Flowpath Length (L) in miles, Centroidal Flowpath Length (LC) in miles, and 10-85 Slope in
feet per mile. It is important to note that these types of equations are usually developed for use with
specific watersheds and are not always applicable to other regions, land cover types, soil types, etc.
Care should be taken in choosing/developing an equation that applies to the modeling area of
interest. For simplicity, we will assume an R/(TC+R) relationship of 0.5 which means that the Storage
Coefficient (R) will be equal to the Time of Concentration (TC).
i.
ii.
f. In the Time of Concentration equation box, enter the above TC equation by double-clicking variables,
functions, and entering numbers as appropriate; select Calculate when finished. Notice how the
global editor automatically populates with Time of Concentration estimates. If you have problems
entering the TC equation, try copying/pasting the following text into the equation box. 2.2 *
(pow(([Longest Flowpath Length (MI)] * [Centroidal Flowpath Length (MI)]) / (pow([10-85 Flowpath
Slope (FT/MI)],0.5)),0.3))
g. Since we are assuming an R/(TC+R) relationship of 0.5, the initial Storage Coefficient estimate will
equal the Time of Concentration calculation. Within the Expression Calculator window, select
Field: Storage Coefficient, then select Calculate, and then select Close. Within the ModClark Global
Editor window, select Apply to save the initial estimates and then select Close.
h. Notice how the Expression Calculator populates the global editor with three significant digits. This
can be changed by selecting Tools→ Program Settings→ Compute and then adjusting
the Expression calculator precision toggle box. For this tutorial, you can leave the precision to the
default value of 3.
b. The Loss parameters are undefined; initial estimates need to be entered. Select Calculator... to
launch the Expression Calculator.
c. Make sure to set the Field: to Initial Deficit and make sure that the Grids tab is selected.
d. Notice how four different variables are available in the Expression Calculator. This is because four
different loss grids have already been imported for your use via the Grid Data Manager and can be
accessed by the Expression Calculator. More info on Grid Data28 can be found in the HEC-HMS
Manual.
e. Double-click Moisture Deficit: initial_deficit to enter the variable into the Initial Deficit equation box
and select Calculate. Notice how the global editor automatically populates with Initial Deficit
estimates. Repeat this step by selecting the three other variables in the Expression Calculator to
estimate Maximum Storage, Constant Rate, and Percent Impervious. When finished, select Close in
the Expression Calculator window. Within the Deficit and Constant Global Editor window,
select Apply to save the initial estimates and then select Close.
3.3.5 Results
1. Now that the Transform and Loss parameters have been estimated with Basin Characteristics data and Grid
data, you can make a simulation run and view the results. A Sep2018 run has already been setup for you.
28 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsum/latest/shared-component-data/grid-data
a. Start a simulation by selecting the Compute tab, then right-click on Sep2018, and select Compute.
b. Alternatively, you can also select the Sep2018 simulation on the toolbar and click on the compute
raindrop button.
2. Once the compute is complete, close the compute progress bar and navigate to the Results tab. Under
the Simulation Runs folder, review your results by expanding the Sep2018 simulation and clicking through
your model elements. When you are done reviewing your results, click on the Outlet element and
select Graph. The graph below in question 1 should appear showing the combined simulated flow (blue
lines) compared with the observed flow (black dotted lines).
3.3.6 Question
1. How do the computed results compare to the observed data?
Click for Answer...
29 https://www.hec.usace.army.mil/confluence/download/attachments/43811406/Punx_EstimatingParameters_Final.zip?
api=v2&modificationDate=1614738515425&version=2
There are some common trends between the computed and observed data, but improvements should
definitely be made to the initial parameter estimates. The takeaway from this tutorial is that the Expression
Calculator can assist in quickly populating the global editors with initial estimates or guesses, but there is no
substitute for calibration when observed data is available. Model calibration will be covered in a separate
tutorial.
The Mahoning Creek watershed at Punxsutawney, PA is part of the Allegheny River Basin located in western
Pennsylvania. The primary conveyance system of this 158 sq. mi. watershed includes the following streams: Stump
Creek, East Branch Mahoning Creek, and Mahoning Creek. The confluence of Stump Creek and East Branch
Mahoning Creek is located east of the enclave of Big Run. Mahoning Creek is downstream of the confluence. The
rectangular box shows the study area.
Below are three guides that will step you through preparing data and creating Shared Component Data necessary
for performing a hydrologic simulation.
4.1.1 Here is a project to get you started on any of the three below tutorials. It
can be opened in HEC-HMS version 4.8 or later, after unzipping: Shared
Component Data.zip30
30 https://www.hec.usace.army.mil/confluence/download/attachments/42173310/Shared%2BComponent%2BData.zip?
api=v2&modificationDate=1614719510189&version=1
Download the Final project files for Time-Series Data, Paired Data, and Grid Data here - Shared
Component Data Final.zip31
Channel Cross-Sections
Another type of paired data used in reach elements of an HEC-HMS model is cross-section data. These are simplified
representations of channel geometry and are used in routing methods such as Muskingum-Cunge.
Click here to access this tutorial: Creating Cross Section Paired Data Curves(see page 427)
31 https://www.hec.usace.army.mil/confluence/download/attachments/42173310/
Final%20W%204%20Introduction%20to%20Shared%20Component%20Data.zip?api=v2&modificationDate=1614816888638&version=1
Step 3. A new window will appear that allows you to create a new time-series data. The default data type is for
precipitation gages, but when you want to create other types of time-series data, make sure you change the data
type before giving it a name and a description. You have been provided precipitation gage data for three stations
near the Mahoning Creek watershed, and you will create a gage for the station nearest the city of Punxsutawney. It
has a station ID of PNXP which you should enter as the time series data name. A description is optional but always a
good idea. Provide a reasonable description for these data, and then select the Create button.
Step 4. You should notice that a new folder is added in the tree view for the Watershed Explorer on the left side for
time-series data. Expand this folder to see that there is a folder for precipitation gages. HEC-HMS organizes Shared
Data Components by type, so multiple time-series data can have the same name as long as they are not of the same
type. Expand the precipitation gages folder to see the PNXP gage, and then select the PNXP gage so that you can
see its settings.
Step 5. Precipitation data for the PNXP gage has been provided in a DSS file in the project folder. In order to link
this data to the new precipitation gage, select Single Record HEC-DSS from the Data Source option for the time-
series gage.
Step 6. The options in the component editor will change so that you can select the precipitation data in the HEC-
DSS file:
Step 7. The red asterisk indicates that these settings are required. First, browse to the provided HEC-DSS file by
selecting the folder icon to the right of the DSS Filename text field. The DSS file is in the HEC-HMS project folder, in a
subfolder called "data." The file is called "gage_precip.dss." Select this HEC-DSS file and press Select.
Step 8. This will populate the DSS Filename text field, but the DSS Pathname (which specifies which data you want
out of the HEC-DSS file) also needs to be populated. Select the button to the right of the DSS Pathname text field (it
looks like a white rectangle with horizontal lines on it.) This will open a viewer for the HEC-DSS file.
Step 9. You are interested in the precipitation data for the PNXP gage, so select the row with the Part B that says
PNXP. Then, press the Set Pathname button. This will populate the DSS Pathname text field. After specifying the
DSS Filename and DSS Pathname, your component editor should look something like this:
Step 10. Save your project by clicking the save icon in the toolbar, or by pressing Ctrl+S. You should notice that the
Units and Time Interval settings will automatically update to match the data in the HEC-DSS file.
Step 11. Next, you will specify the location of the precipitation gage by entering data in the remaining six text fields.
The PNXP gage is located at 40°56'21"N, 79°00'31"W. HEC-HMS uses the convention that negative values of latitude
are in the southern hemisphere, and negative values of longitude are used for the western hemisphere. Enter the
correct values in the component editor, and then save your project.
Why is the location information important for a precipitation gage? What kinds of applications can you see
that would require this?
Answer
The inverse distance precipitation method in the meteorologic models relies on the gage locations in order to
estimate precipitation depth using an inverse distance weighting scheme.
Step 12. Finally, you will create a time window for the precipitation gage. Time windows are only used to view sub-
sets of the precipitation gage data and do not affect hydrologic simulation. They are optional. The default time
window of 01Jan2000 00:00 - 02Jan2000 00:00 does not contain any of the observations contained in this
precipitation data. The first event contained in this gage data is from 01Apr1994 00:00 - 30Apr1994 24:00. Right-click
on the PNXP precipitation gage in the watershed explorer and select Create Time Window.
Step 13. Enter the correct start and end date and time in the new window, and press Add, and then press Close.
Step 14. You should notice a new time window underneath the PNXP gage in the watershed explorer. Click on this
time window so that you can view the precipitation data as a graph.
Step 16. A new window will appear that allows you to create a new time-series data. The default data type is for
precipitation gages, and you will want to change the data type before giving it a name and a description. You have
been provided stream gage data for the discharge gage on Mahoning Creek at Punxsutawney, and you will create a
discharge gage for this station. It has a station ID of USGS 03034000 which you should enter as the time series data
name. A description is optional but always a good idea. Provide a reasonable description for these data, and then
select the Create button.
Step 17. Make sure the time-series data folder is expanded to see that there is a folder for discharge gages. Expand
the discharge gages folder to see the USGS 03034000 gage, and then select it so that you can see its settings.
Step 18. You will be entering the observed discharge data from a provided Microsoft Excel spreadsheet that
contains the half-hourly flow observations for the April 1994 flood event. The units are in cubic feet per second.
Change the setting for units to the correct units, and then change the time interval to 30 minutes.
Step 19. Next, save your project by selecting the save icon in the toolbar, or by pressing Ctrl+S.
Step 20. Create a time window by right clicking on the USGS 03034000 discharge gage and selecting Create Time
Window.
Step 21. Enter a start date and time of 01Apr1994 00:00 and an end date and time of 01May1994 23:30. Then, press
Add, and then press Close.
Step 22. Delete the default time window (01Jan2000 - 02Jan2000) by right clicking on it and selecting Delete Time
Window.
Step 23. Select the time window you just created beneath the USGS 03034000 gage data. Click on the Table tab in
the component editor. You will see a spreadsheet-like table that has the date-time pre-populated for the time
window in the first column, and a column of blank entries to enter discharge data.
Step 24. In the HEC-HMS project folder, there is a sub-folder called "data" and it contains a spreadsheet called
"USGS03034000_Apr1994.xlsx." Navigate to that folder and open the spreadsheet in Microsoft Excel.
The spreadsheet contains two columns of data. The first contains the same dates as were shown in the time series
editor in HEC-HMS. The second column contains the data that you will want to copy into the empty fields in the
time series editor of HEC-HMS.
Step 25. Select the entire column of discharge data beginning with the first 968 value, and copy it from the
spreadsheet. Then go back to HEC-HMS where the time series editor is open for the discharge gage. Click in the first
cell of the editor and select Paste.
Step 26. The column will then be populated with the values from the spreadsheet. Scroll to the bottom to check to
make sure that the last data value coincides with the last time value.
Step 27. Save your HEC-HMS project. You can view the graph of this discharge gage by switching from the Table to
the Graph tab.
Step 3. A new window will appear prompting you to give a name and description to a new paired data curve. But
first, select Elevation-Discharge Functions for the data type.
Step 4. The data you will be given in order to define this function has been interpolated from field observations
using LOESS on the log-transformed stages and flows. Name the paired data Interpolated Rating Curve.
Descriptions are optional but good practice, so give this paired data curve a reasonable description(e.g. LOESS
Interpolated log-log field measurements). After populating the information, click Create.
Step 5. After clicking Create, you should notice a new folder in the tree view of the Watershed Explorer for Paired
Data. Expand this folder, and you should see a sub-folder labeled Elevation-Discharge Functions. HEC-HMS
organizes each data type into its own distinct folder. Expand this folder, and you should see the rating curve you
just created. Click on it to select it and open its properties in the Component Editor.
Step 6. The default settings for the data source is acceptable. Make sure the units label is set to "FT : CFS". Click on
the Table tab to see the manual data entry view.
You have been provided a comma-separated value file containing the interpolated rating curve called
"mahoning_rating.csv" in the data folder inside the HEC-HMS project:
Step 7. Right-click on this file and select Open with > Excel.
In Excel you will see two columns, one for stage and one for discharge. It is important to note that for paired data,
the values must be strictly ascending - meaning they must increase going downwards, and cannot repeat.
Step 8. Select the entire column of stage values and copy it. Then, return to HEC-HMS. Click in the first cell of the
Elevation (FT) column of the table view. Then, right-click and select Paste.
Step 9. You should see several hundred values fill in. Repeat this procedure for the discharge values by switching
back to Excel, copying the column of discharge values, clicking in the first cell of the Discharge (CFS) column of the
table view, right-clicking, and pasting. After pasting the discharge values, the first few rows should look like this:
Step 10. Save your project by clicking the save icon in the toolbar, or by pressing Ctrl+S. Then, from the Component
Editor, switch from the table view to the graph view by clicking on the Graph tab. You should see a graph like below:
Step 11. Finally, you can verify that this curve is available for estimating stage by opening the basin model. Select
the Sep2018 basin model, and click on the Outlet element in the tree.
Step 12. In the Component Editor, switch to the Options tab. You will see that the menu item for Elevation-
Discharge is expandable, and that you can select the Interpolated Rating Curve paired data. Only elevation-
discharge paired data curves are able to be selected in this menu. If you do not see the interpolated rating curve,
then verify that it was created as an elevation-discharge function.
What potential limitation does using only field observations for a rating curve create in a hydrologic
simulation? What does HEC-HMS not do with data in a paired data curve?
Answer
Field observations only span the range of observed flows. If HEC-HMS simulates a flow larger than observation, it
will not be able to use the paired data to estimate the stage associated with that flow. HEC-HMS does not
extrapolate paired data curves. If you want to ensure that the paired data curve covers the entire range of
simulation, you will need to credibly extrapolate it yourself.
A window will prompt you for a name, description and data type. Note that there are two general types of grids in
this list: those ending in "Gridsets", and those ending in "Grids". Gridsets in general represent a time-series of
grids, and are typically hydrometeorological boundary conditions such as precipitation and temperature grids in time.
Grids are static and generally represent spatially distributed hydrologic parameters such as moisture deficit or
percolation rate.
The following tutorial will demonstrate the difference between these two types of data using a static grid (a
precipitation-frequency grid) and a gridset (a time series of precipitation grids.)
32 https://hdsc.nws.noaa.gov/hdsc/pfds/
Step 2. Begin by creating a new Precipitation-Frequency grid by using the Components menu (see the beginning of
this tutorial.) When the creation window comes up, select the Precipitation-Frequency Grid data type (near the
bottom of the list.)
Step 3. Then, name the grid data 24hr 100yr. Providing a description is optional but generally good practice. After
providing a name and description, select Create and a new grid data component will be created.
Step 4. You will find the new grid data in the Watershed Explorer under a new folder called Grid Data. Underneath
that folder in the tree, will be another folder called Precipitation-Frequency Grids. HEC-HMS organizes all of its
grids by data type, and all data types are distinct grids from one another. Underneath the Precipitation-Frequency
Grids folder will be the grid data you just created. Click on that grid data to select it and open its properties in the
Component Editor:
Step 5. The settings for the grid are used to link the Shared Component Data to data on disk. For most grid types,
the data is stored only in Single Record HEC-DSS files. However, to make precipitation-frequency grids easier to
work with, the ASCII grid option is available for these data. For the Data Source, select ASCII. You will see the other
options change in response.
Step 6. For the Filename, use the file browser option (the folder icon) to navigate to the project folder, and then to
the data folder inside of that. In that folder you should only see one file called "orb100yr24ha.asc". Select this file
and select Open.
Step 7. The Filename text box will be populated. You will also need to specify the Units and Unit Scale Factor.
NOAA Atlas 14 data are stored in 1000ths of inches, so select "in" for Units, and 1000 for Unit Scale Factor. (A unit scale
factor is typically used to store floating point data as integers, with the scale factor selected in order to preserve the
desired level of precision.) Your options should look like the following:
Step 8. Finally, save your project by using the Save button in the toolbar or by pressing Ctrl+S.
the precipitation gridset of interest, for the project domain. It will still need to be linked to a Grid Data inside of HEC-
HMS.
Step 10. You will then see a wizard appear. In the first step, accept the default options (to export subbasin
elements) by selecting "Next >" in the bottom-center. On the second step of the wizard you will see an option to set
the destination file. Click on the folder icon to the right of the Filename (*.shp) text field to open the file browser.
Navigate to the data folder inside your HEC-HMS project. Once inside that folder, specify the filename as
"punx_subbasin" and select Save.
Step 11. This will return you to the second step of the wizard. Complete the wizard by selecting Finish in the
bottom-center. This will create the necessary boundary shapefile in the data folder.
Step 13. You will see the Gridded Data Import Wizard open. Click on the folder icon to browse to the data directory
of your HEC-HMS project.
Step 14. Inside of the data directory is a folder called QPE. Open that folder, and you should see 24 MRMS QPE files,
one for each hour of 01 September 2018. Press Ctrl+A to select all files, and then press Open.
Step 15. The list of files inside the import wizard will be populated with the 24 GRIB files you just selected. Press
Next to go to the next step.
Step 16. You should see the variable selection panes. There is only one variable stored in these files called
"GaugeCorrQPE01H_altitude_above_msl". Select it in the left pane, and press the ">" button to import that
variable. Then, press Next to go to the next step.
Step 17. In this step you will specify the geometric options for the import. The first field is for the clipping
datasource. We will use the shapefile we exported in an earlier step. Click on the folder icon to the right of the
clipping datasource text field to browse to the "punx_subbasin.shp" file in the data folder created earlier. Next,
specify a target wkt by selecting the globe icon to the right of the text field of the target wkt text field. This will open
a list of pre-defined projections. The default option of SHG is good, so select Ok in this window. The target wkt text
field will be populated with the SHG wkt. Next, specify the target cell size by clicking on the grid icon to the right of
the text field for target cell size. The default option of 2000 for the cell size is good, so select Ok in this window. The
target cell size text field will be populated. Finally, leave the default resampling method (Bilinear) - this is
acceptable for precipitation data. The wizard should look like this with everything specified:
Step 18. Press Next to go to the next step. Here you will specify the destination file. Click on the folder icon to the
right of the destination text field and navigate to the data folder inside your project. Once there, enter the name
"punx_qpe.dss" in the file name text field and select Open.
Step 19. Upon returning to the wizard the destination text field will be populated, and a set of DSS path text fields
will appear. For Part B use "Punx" and for part F use "MRMS-QPE". Then press Next. The importer will show an
indefinite progress bar and take several seconds to process. When it is finished it will say "Import complete." You
can then close the wizard.
Step 21. Now, in the Watershed Explorer beneath a folder called Grid Data will be a folder called Precipitation
Gridsets. Each grid type in HEC-HMS is distinct from one another and organized in groups. Expand this folder to find
the MRMS-QPE grid dataset you just created. Click on it to select it and open the Component Editor.
Step 22. The two required settings for the gridset are the DSS Filename and DSS Pathname. First, specify the DSS
Filename by clicking on the folder icon to the right of the DSS Filename text field. Then, navigate to the HEC-HMS
project folder, and the data folder inside of that, where you created the "punx_qpe.dss" file previously using the
grid data importer. Select the "punx_qpe.dss" file by clicking on it, and then pressing Select.
Step 23. Next you will specify the DSS Pathname by clicking on the rectangular icon to the right of the text field for
the DSS Pathname. After clicking on that icon, a DSS viewer will open. Click on the first row of the DSS file and then
click Set Pathname at the bottom.
Step 24. Finally, back in the component editor, look at the DSS Pathname text field. There are two dates in the path
separated by "/" characters. Clear out both dates so that the component editor looks like below, and save by
pressing the save icon in the toolbar, or by pressing Ctrl+S.
Precipitation methods are broadly categorized into two types of methods: Hypothetical and Historical.
• Hypothetical methods are idealized representations of precipitation in time and space that are used for
design and analysis. Examples include frequency storms and hypothetical storms, which are typically
associated with a design return interval precipitation frequency event, or HMR 52 storms which are a
representation of probable maximum precipitation. These precipitation methods do not represent storm
events that have occurred, only events which have properties which are useful for design or analysis.
• Historical methods provide means for representing precipitation as it was observed. These methods can
provide additional capability to describe rainfall in space and time. The most simple method is the specified
hyetograph method, which provides a specific precipitation time series for each subbasin element in the
watershed model. The next more complex methods are gage weights and inverse distance. These methods
use precipitation stations and perform some level of averaging in space and time in order to describe the
variability of precipitation. These methods compute the precipitation time series for each subbasin element
instead of being provided directly. Finally, the gridded precipitation method uses a time series of grids that
represent the variability of precipitation in space, generally on a sub-daily time step. Gridded methods are
especially useful for modeling historical events when the precipitation was observed by radar or satellite, or
modeled using reanalysis.
This tutorial will focus on the two most commonly-used methods for modeling historical precipitation: gage
weights and gridded precipitation. A watershed model for Mahoning Creek above Punxsutawney, PA is provided
below to get you started.
The Mahoning Creek watershed at Punxsutawney, PA is part of the Allegheny River Basin located in western
Pennsylvania. The primary conveyance system of this 158 sq. mi. watershed includes the following streams: Stump
Creek, East Branch Mahoning Creek, and Mahoning Creek. The confluence of Stump Creek and East Branch
Mahoning Creek is located east of the enclave of Big Run. Mahoning Creek is downstream of the confluence. The
rectangular box shows the study area.
Below are two guides that will step you through setting up meteorologic models necessary for performing a
hydrologic simulation of historical events.
5.2.1 Introduction
In this tutorial you will associate precipitation gage data with a gage weights precipitation method in a
meteorologic model in order to model the May 1996 precipitation event in the Mahoning Creek watershed. The gage
data has already been associated time-series data in the HEC-HMS model; for more information on creating
precipitation gage time-series, see Creating Time Series Data(see page 70). The watershed has been represented using
a single sub-basin for the entire drainage area above the Punxsutawney stream gage.
33 https://www.hec.usace.army.mil/confluence/download/attachments/42173304/Gage%2BWeights.zip?
api=v2&modificationDate=1614731425535&version=1
Step 2. Expand the basin models node of the watershed explorer tree and select the May1996 basin model. You will
see that this model contains two elements, a sub-basin and a sink. You should see in the basin model map a display
that looks like the watershed above. Next, expand the time-series data node of the tree and you will see that there
are two types of time-series data in the model. This tutorial is concerned with the precipitation data time series.
Expand that sub-node of the tree and you should see three precipitation gages.
These three gages are located around the watershed and represent different observations of similar storm events.
Due to the inherent variation of precipitation in space and time, these three gages are likely to observe different
values of precipitation despite not being very far apart. The map below shows the location of these three stations
relative to the watershed.
Step 3. Expand the gages in the watershed explorer to see that there is a time window associated with each of
them, corresponding to the May 1996 event. You can click on the time window and then select Graph in the
component editor to see a graph of each of the gages' data for this event.
What is the time step of these precipitation gages? For a watershed of this size, is this an appropriate level of
granularity for modeling a flood event?
Answer
The precipitation gages are hourly. For this size of basin 1-hour precipitation should reasonably capture the peak
rainfall intensity and therefore peak runoff. If the watershed were smaller, a finer timestep might be required.
Step 5. For the name, enter May 1996 Gage Weights. Descriptions are optional but generally good practice. Press
Create to create the new met model.
Step 6. Now there should be a new node under the watershed explorer labeled Meteorologic Models. Expand this
node of the tree to see the met model you just created. Click on it to select it and so that you can edit it in the
component editor.
Step 7. In the component editor, select the Gage Weights method for precipitation.
Step 8. Next, switch to the Basins tab and associate the met model with the May1996 basin model by toggling
Include Subbasins to Yes, and saving your model by pressing the save icon in the toolbar or by pressing Ctrl+S.
Step 9. You will see that two additional items appear beneath the May 1996 Gage Weights met model in the
watershed explorer: an entry called Precipitation Gages, and then the subbasin element MahoningCreek. Expand
the MahoningCreek node to see the option to set the gage weights, and then select it.
Step 10. Switch over to the Selections tab of the component editor. Here you should see the three precipitation
time-series gages DUJP, MFFP and PNXP. Enable all three of them in the met model by toggling the Use Gage option
over to Yes. Then, save your project.
Step 11. Switch to the Weights tab of the component editor. Now you should see a table where the first column has
the three gages you just enabled. The other two columns contain the parameters for this precipitation method, the
depth weight and the time weight.
Depth Weight
The depth weight controls how much each gage contributes to the total estimate for precipitation depth for a storm
event. Generally this is estimated using area-average weighting, for example using Thiessen polygons. The
influence of each gage on the spatial estimate of rainfall can be estimated by constructing Thiessen polygons, and
then estimating the percentage of the watershed area covered by each gage's polygon, as below.
These depth weight values can be estimated using GIS by intersecting the Thiessen polygons with the watershed
boundary. The estimated values serve as a starting point for estimating the contribution of each gage to the basin-
averaged precipitation depth and may be adjusted during the calibration process. When there are many gages, the
weight for each becomes less impactful on the result.
Step 12. For this model, enter the following depth weights that correspond to the Thiessen polygons above on the
Weights tab of the component editor:
Time Weight
The time weight controls how much each gage contributes to the time pattern used for the storm event. In general,
it is best to choose one gage located closest to the center of the watershed and give it a time weight of 1 (and the
other gages 0).
Identify a side effect of averaging together time patterns that might make it harder to model a flood peak.
Answer
Primarily, averaging together the time patterns for multiple stations when the peak rainfall intensity is not
coincident can result in a double-peaked hyetograph that has intensity that is overall too low.
Is it better to use a gage with a short or a long timestep for the time pattern?
Answer
Select a gage that has a fine time step so that the temporal resolution of the precipitation pattern can be well-
defined. A daily gage may be available near the watershed centroid, but this will not represent sub-daily intensities.
A nearby gage that is not ideally located, but has hourly or 15-minute data, will be better.
When calibrating a gage weights model, it can be beneficial to test which of the gages being used gets the time
weight of 1. This can affect timing of rainfall and the shape of the overall hyetograph, and using the best gage
possible for timing can improve calibration.
Step 13. For this model, set the time weight for PNXP (the gage inside the basin boundary) to 1, and the others to 0:
Step 15. From the simulation toolbar, select the May 1996 run from the drop down menu, and press the "exploding
raindrop" (compute all elements) button to run the simulation.
Step 16. You can view the results by choosing the Results tab of the watershed explorer, expanding the Simulation
Runs node of the tree, selecting the May 1996 simulation run, and clicking on the elements below it.
5.3.1 Introduction
In this tutorial, you will associate gridded precipitation data with a gridded precipitation method in an HEC-HMS
meteorological model. Gridded precipitation data comes from a number of sources. NEXRAD radar became
operational in the 1990s across much of the United States and is the primary source of gridded precipitation data
since then. Data available prior to the "radar era" are generally created by interpolating gage observations,
reanalysis using numerical weather forecast or general circulation models, or a combination of both. Satellite
observations began in the late 1980s and have improved in quality since, but are generally considered less reliable
than other sources of precipitation observation. Modern products incorporate gage, radar, and satellite data into
multi-sensor products with high spatial and temporal resolution, but such products have limited periods of record.
You will model the September 2018 event in the Mahoning Creek watershed, and an HEC-HMS model containing the
data you need in order to get started has been provided. The gridded precipitation data has already been imported
and associated with a Grid Data component; for more information on importing grid data to your project see
Creating Grid Data(see page 89). The watershed has been represented using a single sub-basin for the entire drainage
area above the Punxsutawney gage.
34 https://www.hec.usace.army.mil/confluence/download/attachments/42173304/Gage%2BWeights%2BFinal.zip?
api=v2&modificationDate=1614731629434&version=1
Step 2. Expand the basin models node of the watershed explorer tree and select the Sep2018 basin model. You will
see that this model contains two elements, a subbasin and a sink. You should see in the basin model map, a display
that looks like the watershed above. There is a second basin model in this project, for the May 1996 event (which
was used in the Gage Weights Precipitation Method(see page 104) tutorial.) It is set up to work with the gage weights
precipitation method.
What are two differences between the basin model for the May 1996 event, which is set up to work with gage
weights, and the basin model for the September 2018 event, which is set up to work with gridded
precipitation?
Answer
The MahoningCreek subbasin in the Sep2018 basin model contains a structured discretization method, and uses
the ModClark transform method, instead of None and Clark for the May1996 model.
Step 3. Next, expand the grid data node of the tree and you will see that there are two types of grid data included in
this model: precipitation and temperature. HEC-HMS keeps different types of grid data separate so that situations
that call for a specific type of grid will only show that type of grid available. For this tutorial, the gridded
precipitation method will only allow you to select precipitation grids as an input, for example.
35 https://www.hec.usace.army.mil/confluence/download/attachments/42173306/Gridded.zip?
api=v2&modificationDate=1614731961342&version=2
Step 5. For the name enter Sep 2018 Gridded. Descriptions are optional but generally good practice. Press Create
to create the new met model.
Step 6. Expand the meteorologic models node of the tree to see the met model you just created. Click on it to select
it and so that you can edit it in the component editor.
Step 7. In the component editor, select the Gridded Precipitation method for precipitation. Make sure the Unit
System is set to U.S. Customary.
Step 8. Next, switch to the Basins tab and associate the met model with the Sep2018 basin model by toggling
Include Subbasins to Yes, and saving your model by pressing the save icon in the toolbar or by pressing Ctrl+S.
Step 9. You will see that an additional item appears beneath the Sep 2018 Gridded met model in the watershed
explorer, an entry called Gridded Precipitation. Click on the Gridded Precipitation node to select it.
Step 10. After selecting the precipitation method you should see a component editor with two options in it. For Grid
Name, select the MRMS Precipitation grid component. The Time Shift (HR) option is used to transpose the grids in
time so that differences such as time zone misalignment can be corrected. In this case, the observed discharge data
and the precipitation data have been set up in the same time zone so no shift will be needed. Save your project.
Compare the parameterization of the gridded precipitation model and the gage weights model. What comparisons
can you make between the gage weights method and the gridded precipitation method?
Answer
Gage weights takes point observations and uses a simple model in order to distribute them in space and time.
However, it cannot create observations where there are no gages. Gridded data explicitly models the temporal and
spatial variability of rainfall by providing an estimate for rainfall intensity everywhere. The gridded data method
assumes that the gridded data are an accurate representation of the observed rainfall, and there are no parameters
to adjust how the precipitation is applied to the watershed. The gage weights method allows some adjustment of
its parameters to change how the combination of gage data estimates the basin-average precipitation, but relies on
averaging a small number of point precipitation depth estimates.
Select Run2018 from drop down window and Compute All Elements
Optional Step 13. Under the Sep 2018 Gridded met model there will now be an additional node for Gridded Hamon
ET. Select this node so that it can be edited in the component editor.
What grid datasets are available to be selected in this menu? How many grid datasets are there in this HEC-
HMS project?
Answer
Only one dataset is available, the RTMA dataset, which is a temperature grid set. There are two grid data in the HEC-
HMS project, but the other is for precipitation data. It is important when creating grid data to select the correct data
type, because HEC-HMS organizes grids so that the correct types are used in the correct places.
Optional Step 14. Set the Temperature Grid to RTMA, but leave the default value for the Hamon Coefficient. This
value can be adjusted during the model calibration process, and is typically changed in continuous models. Re-
compute the Sep 2018 simulation.
5.4.1 Introduction
In this tutorial, you will create an HEC-HMS meteorological model that uses the inverse distance method to model
the May 1996 precipitation event in the Mahoning Creek watershed. This is an interpolation method that weights
the gage data based on the gage locations. The basic premise is that gages that are closer to a specific area will be
weighted more heavily than gages that are a further distance away. The gage network has already been associated
with time-series data in the HEC-HMS model; for more information on creating precipitation gage time-series,
see Creating Time Series Data(see page 70). The watershed has been represented using a single sub-basin for the
entire drainage area above the Punxsutawney stream gage.
36 https://www.hec.usace.army.mil/confluence/download/attachments/42173306/Gridded%2BFinal.zip?
api=v2&modificationDate=1614731920147&version=1
Step 2. Expand the basin models node, the meteorological models node, and the time series data | precipitation
gages node of the watershed explorer tree. In this tutorial, we will be using the May1996 basin model (which was
also used in the Gage Weights Precipitation Method(see page 104) tutorial). The Sep2018 model was used in the
37 https://www.hec.usace.army.mil/confluence/download/attachments/27232303/Inverse_Distance.zip?
api=v2&modificationDate=1614814796094&version=1
Gridded Precipitation Method(see page 113) tutorial. We will be using the same three precipitation gages that were
setup previously in the Gage Weights Precipitation Method tutorial. With the gage weights method, the user
specifies the depth weights and time weights that HEC-HMS uses to interpolate the gaged rainfall data. With the
inverse distance method, the gages need to have a specified latitude and longitude before the inverse distance
weighting algorithm can be applied.
Latitude Degrees 41 40 40
Latitude Minutes 7 58 56
Latitude Seconds 33 48 21
Longitude Minutes 45 41 0
Longitude Seconds 50 22 31
Step 4. The two meteorologic models that are currently in this project are from previous tutorials. Let's create a
new meteorological model
by using the Components menu, then selecting Create Component > Meteorologic Model...
Step 5. For the name enter May 1996 Inverse Distance. Descriptions are optional but generally good practice;
enter DUJP-MFFP-PNXP inverse distance. Press Create to create the new met model.
Step 6. Now there should be a new node under the watershed explorer labeled Meteorologic Models. Expand this
node of the tree to see the met model you just created. Click on it to select it and so that you can edit it in the
component editor.
Step 7. In the component editor, select the Inverse Distance method for precipitation. Also, make sure to select
U.S. Customary.
Step 8. Next, switch to the Basins tab and associate the met model with the May1996 basin model by toggling
Include Subbasins to Yes, and saving your model by pressing the save icon in the toolbar or by pressing Ctrl+S.
Step 9. You will see that two additional items appear beneath the May 1996 Inverse Distance met model in the
watershed explorer: an entry called Precipitation Gages, and then the subbasin element MahoningCreek. Select
Precipitation Gages and then toggle the Use Gage options to Yes as seen below.
Step 10. Set the search distance. The distance between the subbasin node and a precipitation gage must be less
than the search distance for HEC-HMS to use the precipitation gage when calculating subbasin average
precipitation. Select the "Options" tab and set the search distance to 62 miles as shown below.
Step 11. Add an inverse distance subbasin node by specifying a node location (latitude, longitude). In this example,
we will use the basin centroid location for the node location. Expand the watershed explorer by clicking the plus
sign next to the MahoningCreek subbasin name one level below the inverse distance meteorologic model. Select
the Inverse Distance element to open a Component Editor with three tabs, "Nodes," "Latitudes," and
"Longitudes." Enter a Node Name of Basin Centroid and a Weight of 1.0 (this is a weighting of 100%).
Step 12. Enter the node coordinate using the basin centroid coordinates (Latitude 40o 58' 34" , Longitude -78o 51'
58") as shown in the below figures.
Step 14. From the simulation toolbar, select the May 1996 Inverse Distance run from the drop down menu, and
press the "exploding raindrop" (compute all elements) button to run the simulation.
Step 15. You can view the results by choosing the Results tab of the watershed explorer, expanding the Simulation
Runs node of the tree, selecting the May 1996 Inverse Distance simulation run, and clicking on the elements below
it.
Precipitation values for each gage are weighted based on the distance the gage is from the node location.
38 https://www.hec.usace.army.mil/confluence/download/attachments/27232303/Inverse_Distance_Final.zip?
api=v2&modificationDate=1614814851816&version=1
The flow of water in either saturated or unsaturated soil can almost always be described by Darcy's Law. It is a
basic relationship that states the vertical flow of water is proportional to the potential gradient. When the spatial
coordinate is taken as zero at the soil surface and measured downward, Darcy's Law can be stated in the following
form:
where ν is the flow per unit area, ψ is the matric potential (a negative value), z is the spatial coordinate (measured
positive downward), and K is the hydraulic conductivity. If the soil is saturated, then K is the saturated hydraulic
conductivity and is a function of the soil properties and the water properties. For unsaturated conditions the
conductivity is still a function of soil and water properties, but is additionally a function of the matric potential. As
the water content decreases from saturated toward the residual content, the matric potential becomes increasingly
negative. The relationship between conductivity and matric potential is nonlinear resulting in the magnitude of the
conductivity varying by several orders of magnitude over the possible range of water content. Darcy's Law is part of
the derivation of all physically based models of infiltration and is also used on many conceptual models.
However, not all precipitation losses are due to soil infiltration alone. Interception of precipitation by a vegetation
canopy or surface depressions (ponding) can reduce the amount of precipitation that will either runoff or be
infiltrated. Also, once water is in the soil, a portion that was infiltrated may return to the atmosphere through
evapotranspiration. Users are directed to the HEC-HMS documentation39 for further information regarding the
available canopy and surface methods.
6.1.3 Background
In this tutorial group, the study area is designated as the Mahoning Creek watershed upstream of the town of
Punxsutawney, PA. Mahoning Creek is part of the Allegheny River Basin and is located in western Pennsylvania, as
shown in Figure 2.
39 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsum/latest
The primary conveyance system of this 158 sq. mi. study area includes the following major streams: Stump Creek,
East Branch Mahoning Creek, and Mahoning Creek. The confluence of Stump Creek and East Branch Mahoning
Creek is located east of the enclave of Big Run. Mahoning Creek is downstream of the confluence. Elevations within
the study area range from approximately 1175 ft to 2260 ft as referenced to the North American Vertical Datum of
198840 (NAVD88), as shown in Figure 3.
40 https://geodesy.noaa.gov/datums/vertical/north-american-vertical-datum-1988.shtml
6.1.4 Data
A 17-day time period in September 2018 will be simulated within this tutorial. During this time period, the remnants
of Hurricane Florence41 passed through the study area and brought heavy rainfall and subsequent flooding.
41 https://www.ncdc.noaa.gov/sotc/tropical-cyclones/
201809#:~:text=Significant%20Events&text=Hurricane%20Florence%20was%20a%20long,along%20its%20route%2C%20reaching%20Cat
Precipitation
Multi-Radar/Multi-Sensor System42 (MRMS) gridded precipitation will be used as the boundary condition for all
simulations within this tutorial. The MRMS gridded dataset has a native spatial resolution of 1 kilometer x 1
kilometer and a time step of 1-hour. The raw data files were ingested, projected, and output in DSS format using
the HEC Vortex suite of tools43. The time zone of this data is Coordinated Universal Time (UTC).
Streamflow
Both stage and flow data is recorded by the U.S. Geological Survey (USGS) at Punxsutawney44 and stored within the
National Water Information System (NWIS). The location of this stream gage is shown in Figure 4. The outlet of the
study area within the provided HEC-HMS project is co-located with this gaging station.
42 https://www.nssl.noaa.gov/projects/mrms/
43 https://github.com/HydrologicEngineeringCenter/Vortex
44 https://waterdata.usgs.gov/usa/nwis/uv?03034000
15-minute observed streamflow data was imported from NWIS using HEC-DSSVue45. The data was then shifted +5
hours to align with the previously-mentioned MRMS data (i.e. UTC). Figure 5 shows the basin-average precipitation
for the study area and observed streamflow data for the Mahoning Creek at Punxsutawney gage for the simulation
time period.
45 https://www.hec.usace.army.mil/software/hec-dssvue/downloads/dev/
46 https://viewer.nationalmap.gov/basic/#/
The subbasin element was configured to use a 2000-meter SHG Structured discretization, the Simple canopy
method, the Clark unit hydrograph transform method, and the Recession baseflow method. The previously-
mentioned observed discharge record from the Mahoning Creek at Punxsutawney gage was linked to both the
subbasin and the sink elements.
6.2.1 Overview
In this tutorial you will apply the HEC-HMS Initial and Constant loss method to a modeling application. Initial
parameter estimates will be estimated using GIS information and the model will be calibrated through trial and
error. HEC-HMS version 4.7 was used to created this tutorial. You will need to use HEC-HMS version 4.7, or newer, to
open the project files.
6.2.2 Background
The Initial and Constant loss method uses a hypothetical single soil layer to account for changes in moisture
content. While this method is very simple, it is widely used to model watersheds that lack detailed subsurface
information. Since no means for extracting infiltrated water is included, this method should only be used for
event simulation. The underlying concept of the initial and constant loss method is that the maximum potential
rate of precipitation loss, fc, is constant and unchanging during a simulation. Thus, if the precipitation rate, pt, is
less than fc, no runoff will occur. An initial loss, Ia, can be specified to represent the volume of precipitation that is
expected to be lost prior to the onset of runoff initiation. Until the accumulated precipitation volume, pi, exceeds Ia,
no runoff will occur. After Ia has been satisfied, runoff will only occur if and when pt is in excess of fc. Parameters
that are required to utilize this method within HEC-HMS include the Initial Loss [inches or millimeters]
and Constant Rate [in/hr or mm/hr]. The Directly Connected Impervious Area [percent] is an optional parameter
and can be specified by the user.
Initial Loss
The initial loss defines the volume of water that is required to fill the soil layer at the start of the simulation. This
parameter is typically defined using the product of the soil moisture state at the start of the simulation and an
assumed active layer depth, but it should be calibrated using observed data. In order to estimate an appropriate
initial loss, the antecedent conditions at the beginning of September 2018 must be investigated. Daily average
humidity, daily precipitation accumulation, daily average temperature, and daily average wind speed are shown in
Figure 1.
47 https://www.hec.usace.army.mil/confluence/download/attachments/46007050/Loss_Methods_start.zip?
api=v2&modificationDate=1614784516863&version=1
48 https://www.publications.usace.army.mil/Portals/76/Publications/EngineerManuals/EM_1110-2-1417.pdf?ver=VFC-
A5m2Q18fxZsnv19U8g%3d%3d
49 https://www.hec.usace.army.mil/confluence/hmsdocs/hmstrm
Use Figure 1 and the following information to answer Question 1. Assume that the soil within the study area:
1. Was completely saturated by rainfall on August 22, 2018 and
2. Requires approximately 14 days with no rain to completely dry under normal, summer conditions
Question 1: Estimate an initial loss (in inches) for the study area at the beginning of September 2018.
Answer
A reasonable initial loss is approximately 2 inches. However, this parameter must be calibrated using observed
data.
Constant Rate
The constant rate defines the rate at which precipitation will be infiltrated into the soil layer after the initial loss
volume has been satisfied. Typically, this parameter is equated with the saturated hydraulic conductivity of the soil
which is defined as the rate at which water moves through a unit area of saturated soil in a unit time under a unit
hydraulic gradient. To estimate saturated hydraulic conductivity throughout the study area, Gridded Soil Survey
Geographic50 (gSSURGO) data for Pennsylvania was obtained from the U.S. Department of Agriculture's Geospatial
Data Gateway51. Using GIS tools, surficial soil textures were then extracted from the gSSURGO data and are shown
in Figure 2. Also, the percent of the study area encompassed by each soil texture is shown in Table 1.
50 https://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/home/?cid=nrcs142p2_053628
51 https://datagateway.nrcs.usda.gov/
Sand 0.1
Loam 17.1
Clay 0.9
Bedrock 10.7
These surficial soil textures can be used to estimate initial parameter values for nearly all loss methods within HEC-
HMS, including the Initial and Constant loss method. For instance, Rawls, Brakensiek, and Miller (1983)52 assembled
data from thousands of soil samples located throughout the United States and related soil textures to various
parameters that are useful within hydrologic modeling. The saturated hydraulic conductivity of various soil
textures is shown in Table 2.
Table 2. Soil Textures and Saturated Hydraulic Conductivity, reproduced from Rawls, Brakensiek, and Miller (1983)
Sand 4.6
Loam 0.1
52 https://ascelibrary.org/doi/10.1061/%28ASCE%290733-9429%281983%29109%3A1%2862%29
Clay 0.01
Question 2: What are the top five predominant soil textures throughout the study area?
Answer
Silt loam, loam, bedrock, silty clay loam, and sandy loam are the five most predominant soil textures (by area)
throughout the study area. In fact, these five soil textures cover over 98% of the total study area. The remaining
soil textures account for a very small percentage of the total area and can be disregarded.
Next, use Table 2 to answer the following question. Also, (conservatively) assume that bedrock has a saturated
hydraulic conductivity of 0 in/hr.
Question 3: What is the saturated hydraulic conductivity for the top five predominant soil textures found
when answering Question 2?
Answer
Silt loam = 0.3 in/hr, loam = 0.1 in/hr, bedrock = 0 in/hr, silty clay loam = 0.04 in/hr, and sandy loam = 0.4 in/hr.
Finally, estimate an average (i.e. representative) saturated hydraulic conductivity for the study area using your
answers to the previous two questions. Don't spend too much time being overly precise; instead, quickly estimate
the percentage of the study area that is encompassed by each predominant soil texture and multiply by the
corresponding saturated hydraulic conductivity. Once you've done that for each of the five predominant soil
textures, sum the values.
Question 4: What is an average (i.e. representative) saturated hydraulic conductivity for the study area?
Answer
Silt loam covers approximately 60% of the total area, loam covers approximately 17% of the total area, bedrock
covers approximately 10% of the total area, silty clay loam covers approximately 7% of the total area, and sandy
loam covers approximately 3% of the total area. Thus, (0.6 * 0.3 in/hr) + (0.17 * 0.1 in/hr) + (0.1 * 0.0 in/hr) + (0.07 *
0.04 in/hr) + (0.03 * 0.4 in/hr) = 0.2 in/hr.
6. Create a new simulation run by clicking Compute | Create Compute | Simulation Run.
a. Name the new simulation run "Sept2018_InitialConst".
b. Select the Mahoning Creek basin model.
c. Select the MRMS meteorologic model.
d. Select the September 2018 control specifications.
7. Select the Sept2018_InitialConst simulation run from the Compute toolbar,
To better match the observed runoff, the Initial and Constant loss method parameters must be
c. calibrated.
of 15 minutes, the effective rate is 0.025 inches / 15 minutes (i.e. 0.1 in/hr * 15 / 60).
c. When looking at the result graph and summary table for the Mahoning Creek subbasin element, the
computed hydrograph begins to rise at approximately the same time as the observed hydrograph, as
shown in Figure 6. However, the computed runoff volume (2.88 inches) is still much less than the
2. Continue modifying the loss parameters to approximately match the observed runoff volume.
a. Change the constant loss rate to 0.05 in/hr and rerun the simulation.
b. Notice that the computed runoff volume is now slightly greater than the observed runoff volume (5.0
vs 4.4 inches), as shown in Figure 7. Also, the computed peak flow rate more closely matches the
observed peak flow rate. However, the shape of the computed runoff hydrograph doesn't match the
observed hydrograph shape on September 9th and 10th; in particular, the computed runoff is much
c. To reduce the amount of runoff on September 9th and 10th, increase the initial loss to 1.75 inches
and rerun the simulation.
d. Notice that the computed runoff volume is now approximately equal to the observed runoff volume
(4.49 vs 4.4 inches), as shown in Figure 8. Also, the shape of the computed hydrograph matches the
3. Continue adjusting the Initial and Constant loss method parameters in an attempt to simultaneously best
match the peak flow rate, runoff volume, and hydrograph shape. Record your RMSE Std Dev, Percent Bias,
and Nash-Sutcliffe statistical metrics as you adjust model parameters.
a. At this point, it should be apparent that simultaneously matching peak flow rate, runoff volume, and
hydrograph shape can be difficult.
b. For instance, the "double peak" runoff response within the observed results between September 9th
- 11th is nearly impossible to recreate when using a single subbasin. This type of "signature"
suggests two discrete pulses of runoff. To better match this response, two or more subbasins should
be delineated.
53 https://www.hec.usace.army.mil/confluence/download/attachments/46007050/Initial_and_Constant_finish.zip?
api=v2&modificationDate=1614784771362&version=1
Question 6. Is it possible to reduce the computed runoff response on September 17th - 18th while not
adversely affecting the computed runoff response on September 9th - 12th when using the Initial and
Constant loss method? If not, why?
Answer
No. When using the Initial and Constant loss method, after the initial loss has been satisfied, the constant rate is the
only means by which precipitation can be lost. In this instance, the initial loss is satisfied on September 9th.
Increasing the constant loss rate to reduce the computed runoff on September 17th - 18th will unfavorably affect
the runoff response on September 9th - 12th.
The Initial and Constant loss method does not include a way for infiltrated water to be extracted through
evapotranspiration. In fact, a warning message is written to the messages panel:
WARNING 45231: Transpiration is not computed for Initial and Constant loss method in subbasin "Mahoning
Creek".
If this is important to the specific application at hand, another loss method (e.g. Layered Green and Ampt, Deficit and
Constant, Soil Moisture Accounting, etc) should be used that can simulate these processes.
6.3.1 Overview
In this tutorial you will apply the HEC-HMS Green and Ampt loss method to a modeling application. Initial parameter
estimates will be estimated using GIS information and the model will be calibrated through trial and error. HEC-
HMS version 4.7 was used to created this tutorial. You will need to use HEC-HMS version 4.7, or newer, to open the
project files.
6.3.2 Background
The Green and Ampt loss method was derived using a simplification of the comprehensive Richard's equation (1931)
for unsteady water flow in soil. Within this method, infiltration proceeds with so-called piston displacement, as
shown in Figure 1.
At the start of the simulation, the soil is assumed to be uniformly saturated and infinitely deep (Green & Ampt, 1911).
The method automatically accounts for ponding at the soil surface. Similar to the Initial and Constant method, no
means for extracting infiltrated water is included. Therefore, this method should only be used for event
simulation. A layered version of the Green and Ampt method is also available within HEC-HMS that does include
these processes but will not be used within this tutorial.
Parameters that are required to utilize this method within HEC-HMS include the initial moisture content or deficit
[in/in or mm/mm], wetting front suction head [in or mm], saturated hydraulic conductivity [in/hr or mm/hr], and
directly connected impervious area [percent].
54 https://www.hec.usace.army.mil/confluence/download/attachments/46007050/Loss_Methods_start.zip?
api=v2&modificationDate=1614784516863&version=1
Effective Porosity
To estimate a representative effective porosity throughout the study area, Gridded Soil Survey
Geographic57 (gSSURGO) data for Pennsylvania was obtained from the U.S. Department of Agriculture's Geospatial
Data Gateway58. Using GIS tools, surficial soil textures were then extracted from the gSSURGO data and are shown
in Figure 2. Also, the percent of the study area encompassed by each soil texture is shown in Table 1.
55 https://www.publications.usace.army.mil/Portals/76/Publications/EngineerManuals/EM_1110-2-1417.pdf?ver=VFC-
A5m2Q18fxZsnv19U8g%3d%3d
56 https://www.hec.usace.army.mil/confluence/hmsdocs/hmstrm
57 https://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/home/?cid=nrcs142p2_053628
58 https://datagateway.nrcs.usda.gov/
Sand 0.1
Loam 17.1
Clay 0.9
Bedrock 10.7
These surficial soil textures can be used to estimate initial parameter values for nearly all loss methods within HEC-
HMS, including the Green and Ampt loss method. For instance, Rawls, Brakensiek, and Miller (1983)59 assembled
data from thousands of soil samples located throughout the United States and related soil textures to various
useful parameters. The effective porosity of various soil textures is shown in Table 2.
Table 2. Soil Textures and Effective Porosity, reproduced from Rawls, Brakensiek, and Miller (1983)
Sand 0.42
Loam 0.43
59 https://ascelibrary.org/doi/10.1061/%28ASCE%290733-9429%281983%29109%3A1%2862%29
Clay 0.39
Question 2: What is the effective porosity for the top five predominant soil textures found when answering
Question 1?
Answer
Silt loam = 0.49 in3 / in3, loam = 0.43 in3 / in3, bedrock = 0 in3 / in3, silty clay loam = 0.43 in3 / in3, and sandy loam =
0.41 in3 / in3.
Finally, estimate an average (i.e. representative) effective porosity for the study area using your answers to the
previous two questions. Don't spend too much time being overly precise; instead, quickly estimate the percentage
of the study area that is encompassed by each predominant soil texture and multiply by the corresponding effective
porosity. Once you've done that for each of the five predominant soil textures, sum the values.
Question 3: What is an average (i.e. representative) effective porosity for the study area?
Answer
Silt loam covers approximately 60% of the total area, loam covers approximately 17% of the total area, bedrock
covers approximately 10% of the total area, silty clay loam covers approximately 7% of the total area, and sandy
loam covers approximately 3% of the total area. Thus, (0.6 * 0.49 in3 / in3) + (0.17 * 0.43 in3 / in3) + (0.1 * 0.0 in3 / in3)
+ (0.07 * 0.43 in3 / in3) + (0.03 * 0.41 in3 / in3) = 0.41 in3 / in3.
Soil Moisture
In order to estimate the amount of moisture present within the soil throughout the study area, the antecedent
conditions at the beginning of September 2018 must be investigated. Daily average humidity, daily precipitation
accumulation, daily average temperature, and daily average wind speed are shown in Figure 3.
Use Figure 3 and the following information to answer Question 4. Assume that the soil within the study area:
1. Was completely saturated by rainfall on August 22, 2018 and
2. Requires approximately 14 days with no rain to completely dry under normal, summer conditions
Question 4: Estimate an initial content (in3 / in3) for the study area at the beginning of September 2018.
Answer
It's reasonable to assume that approximately 75% of the soil moisture has been removed by either gravity or
evapotranspiration following the rainfall event on August 22nd. Thus a reasonable estimate of initial content
is 0.1 in3 / in3 (i.e. 0.25 * 0.41 in3 / in3). However, this parameter must be calibrated using observed data.
Table 3. Soil Textures and Wetting Front Suction Head, reproduced from Rawls, Brakensiek, and Miller (1983)
Sand 1.9
Loam 3.5
Clay 12.5
Use Figure 2, Table 1, Table 3, and your answers to the previous questions to answer the following question.
Question 5: What is an average (i.e. representative) wetting front suction head for the study area?
Answer
Silt loam covers approximately 60% of the total area, loam covers approximately 20% of the total area, bedrock
covers approximately 10% of the total area, silty clay loam covers approximately 7% of the total area, and sandy
loam covers approximately 3% of the total area. Thus, (0.6 * 6.6 in) + (0.17 * 3.5 in) + (0.1 * 0.0 in) + (0.07 * 10.7 in) +
(0.03 * 8.6 in) = 5.6 in.
Sand 4.6
Loam 0.1
Clay 0.01
Use Figure 2, Table 1, Table 4, and your answers to the previous questions to answer the following question.
Question 6: What is an average (i.e. representative) saturated hydraulic conductivity for the study area?
Answer
Silt loam covers approximately 60% of the total area, loam covers approximately 17% of the total area, bedrock
covers approximately 10% of the total area, silty clay loam covers approximately 7% of the total area, and sandy
loam covers approximately 3% of the total area. Thus, (0.6 * 0.3 in/hr) + (0.17 * 0.1 in/hr) + (0.1 * 0.0 in/hr) + (0.07 *
0.04 in/hr) + (0.03 * 0.4 in/hr) = 0.2 in/hr.
6. Create a new simulation run by clicking Compute | Create Compute | Simulation Run.
a. Name the new simulation run "Sept2018_GreenAmpt".
b. Select the Mahoning Creek basin model.
c. Select the MRMS meteorologic model.
d. Select the September 2018 control specifications.
7. Select the Sept2018_GreenAmpt simulation run from the Compute toolbar,
To better match the observed runoff, the Green and Ampt loss method parameters must be
c. calibrated.
c. When looking at the result graph and summary table for the Mahoning Creek subbasin element,
the computed hydrograph begins to rise at approximately the same time as the observed
hydrograph, as shown in Figure 7. However, the computed runoff volume (3.5 inches) is still less than
the observed runoff volume (4.4 inches) so further modifications are required.
2. Continue modifying the loss parameters to approximately match the observed runoff volume.
a. Change the wetting front suction head to 7.5 inches and rerun the simulation.
b. As shown in Figure 9, more precipitation is infiltrated than before. However, the asymptotic limit (i.e.
saturated hydraulic conductivity) is still the same. Thus, the initially high rate at which precipitation
is infiltrated decreases FASTER than it did when the wetting front suction head was set to 5.6
inches. This demonstrates the effect that changes in wetting front suction head have on infiltration
when using the Green and Ampt method: as wetting front suction head increases, so does the rate at
which infiltration approaches the saturated hydraulic conductivity (and vice versa).
c. Change the saturated hydraulic conductivity to 0.025 in/hr and rerun the simulation.
d. As shown within Figure 10, the minimum infiltration rate has decreased. But, the amount of
precipitation loss prior to runoff initiation has also decreased. When using the Green and Ampt loss
method, changes to saturated hydraulic conductivity will likely result in the greatest change to the
resultant runoff hydrograph. Ultimately, more runoff is produced than before (computed = 4.88 vs
observed = 4.4 inches). Also, the computed peak flow rate more closely matches the observed peak
flow rate. However, the shape of the computed runoff hydrograph doesn't match the observed
hydrograph shape on September 9th and 10th; in particular, the computed runoff is much greater
than the observed runoff within those two days. Thus, additional modifications are necessary.
e. To reduce the amount of runoff on September 9th and 10th, decrease the initial content to 0.05 in3 /
in3 and rerun the simulation.
f. Notice that the computed runoff volume is now better approximating the observed runoff volume
(4.69 vs 4.4 inches), as shown in Figure 11.
3. Continue adjusting the Green and Ampt loss method parameters in an attempt to simultaneously best match
the peak flow rate, runoff volume, and hydrograph shape. Record your RMSE Std Dev, Percent Bias, and
Nash-Sutcliffe statistical metrics as you adjust model parameters.
a. At this point, it should be apparent that simultaneously matching peak flow rate, runoff volume, and
hydrograph shape can be difficult.
b. For instance, the "double peak" runoff response within the observed results between September 9th
- 11th is nearly impossible to recreate when using a single subbasin. This type of "signature"
suggests two discrete pulses of runoff. To better match this response, two or more subbasins should
be delineated.
Using an initial content of 0.05 in3 / in3, a saturated content of 0.41 in3 / in3, a wetting front suction head of 10 in,
and a saturated hydraulic conductivity of 0.02 in/hr offers an RMSE Std Dev, Percent Bias, and Nash-Sutcliffe
60 https://www.hec.usace.army.mil/confluence/download/attachments/46007062/Green_and_Ampt_finish.zip?
api=v2&modificationDate=1614786762579&version=1
statistical metrics of 0.4, 9.69%, and 0.808, respectively, as shown in Figure 12. However, these are by no means
the only parameter combinations that can result in acceptable model calibration.
Question 8. Is it possible to reduce the computed runoff response on September 17th - 18th while not
adversely affecting the computed runoff response on September 9th - 12th when using the Green and
Ampt loss method? If not, why?
Answer
No. When using the Green and Ampt loss method (as implemented within HEC-HMS), precipitation will always be
infiltrated at an initially high rate but decrease towards the saturated hydraulic conductivity as additional
precipitation is infiltrated. In this instance, increasing the saturated hydraulic conductivity (or wetting front suction
head) to reduce the computed runoff on September 17th - 18th will unfavorably affect the runoff response on
September 9th - 12th.
The Green and Ampt loss method does not include a way for infiltrated water to be extracted through
evapotranspiration. In fact, a warning message is written to the messages panel:
WARNING 45231: Transpiration is not computed for Green and Ampt loss method in subbasin "Mahoning Creek".
If this is important to the specific application at hand, another loss method (e.g. Layered Green and Ampt, Deficit and
Constant, Soil Moisture Accounting, etc) should be used that can simulate these processes.
6.4.1 Overview
In this tutorial you will apply the HEC-HMS Deficit and Constant loss method to a modeling application. Initial
parameter estimates will be estimated using GIS information and the model will be calibrated through trial and
error. HEC-HMS version 4.7 was used to created this tutorial. You will need to use HEC-HMS version 4.7, or newer, to
open the project files.
6.4.2 Background
The Deficit and Constant loss method is very similar to the Initial and Constant loss method in that a hypothetical
single soil layer is used to account for changes in moisture content. However, the deficit and constant method
allows for continuous simulation when used in combination with a canopy method that will extract water
from the soil in response to potential evapotranspiration computed in the meteorologic model. Between
precipitation events, the soil layer will lose moisture as the canopy extracts infiltrated water. Unless a canopy
method is selected, no soil water extraction will occur. This method may also be used in combination with a
surface method that will hold water on the land surface. The water in surface storage can infiltrate into the soil
layer and/or be removed through evapotranspiration. The infiltration rate is determined by the capacity of the soil
layer to accept water. When both a canopy and surface method are used in combination with the deficit constant
loss method, the system can be conceptualized as shown in Figure 1.
If the moisture deficit is greater than zero, water will infiltrate into the soil layer. Until the moisture deficit has been
satisfied, no percolation out of the bottom of the soil layer will occur. After the moisture deficit has been satisfied,
the rate of infiltration into the soil layer is defined by the constant rate. The percolation rate out of the bottom of
the soil layer is also defined by the constant rate while the soil layer remains saturated. Percolation stops as soon
as the soil layer drops below saturation (moisture deficit greater than zero). Moisture deficit increases in response
to the canopy extracting soil water to meet the potential evapotranspiration demand. Parameters that are required
61 https://www.hec.usace.army.mil/confluence/download/attachments/46007050/Loss_Methods_start.zip?
api=v2&modificationDate=1614784516863&version=1
to utilize this method within HEC-HMS include the Initial Deficit [inches or millimeters], Maximum Deficit [inches or
millimeters], and Constant Rate [in/hr or mm/hr]. The Directly Connected Impervious Area [percent] is an optional
parameter and can be specified by the user.
Initial Deficit
The initial deficit defines the volume of water that is be required to fill the soil layer at the start of the simulation
while the maximum deficit specifies the total amount of water the soil layer can hold. This parameter is most often
estimated using the product of the soil moisture state at the start of the simulation and an assumed active layer
depth. However, this parameter must be calibrated using observed data, must be less than or equal to the
maximum deficit, and must be specified as an effective depth (i.e. inches or millimeters).
In order to estimate an appropriate initial deficit, the antecedent conditions at the beginning of September 2018
must be investigated. Daily average humidity, daily precipitation accumulation, daily average temperature, and
daily average wind speed are shown in Figure 2.
62 https://www.publications.usace.army.mil/Portals/76/Publications/EngineerManuals/EM_1110-2-1417.pdf?ver=VFC-
A5m2Q18fxZsnv19U8g%3d%3d
63 https://www.hec.usace.army.mil/confluence/hmsdocs/hmstrm
Use Figure 2 and the following information to answer Question 1. Assume that the soil within the study area:
1. Was completely saturated by rainfall on August 22, 2018 and
2. Requires approximately 14 days with no rain to completely dry under normal, summer conditions
Question 1: Estimate an initial deficit (in inches) for the study area at the beginning of September 2018.
Answer
A reasonable initial deficit is approximately 2 inches. However, this parameter must be calibrated using observed
data.
Maximum Deficit
The maximum deficit is the driest a soil can become under the influence of gravity, evaporation, and transpiration.
This parameter is typically estimated as the difference between the saturation storage of the soil and the wilting
point storage over an assumed active soil layer depth. However, this parameter should be calibrated using
observed data, must be greater than or equal to the initial deficit, and also must be specified as an effective depth
(i.e. inches or millimeters).
Saturation Storage
Saturation storage refers to the amount of moisture that can be held within a soil column when all air bubbles have
been forced out. This parameter is a function of the effective porosity. To estimate a representative effective
porosity throughout the study area, Gridded Soil Survey Geographic64 (gSSURGO) data for Pennsylvania was
obtained from the U.S. Department of Agriculture's Geospatial Data Gateway65. Using GIS tools, surficial soil
textures were then extracted from the gSSURGO data and are shown in Figure 3. Also, the percent of the study area
encompassed by each soil texture is shown in Table 1.
64 https://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/home/?cid=nrcs142p2_053628
65 https://datagateway.nrcs.usda.gov/
Sand 0.1
Loam 17.1
Clay 0.9
Bedrock 10.7
These surficial soil textures can be used to estimate initial parameter values for nearly all loss methods within HEC-
HMS, including the Deficit and Constant loss method. For instance, Rawls, Brakensiek, and Miller
(1983)66 assembled data from thousands of soil samples located throughout the United States and related soil
textures to various useful parameters. The effective porosity of various soil textures is shown in Table 2.
Table 2. Soil Textures, Effective Porosity, and Wilting Point, reproduced from Rawls, Brakensiek, and Miller (1983)
Sand 0.42
66 https://ascelibrary.org/doi/10.1061/%28ASCE%290733-9429%281983%29109%3A1%2862%29
Loam 0.43
Clay 0.39
Question 3: What is the effective porosity for the top five predominant soil textures found when answering
Question 2?
Answer
Silt loam = 0.49 in3 / in3, loam = 0.43 in3 / in3, bedrock = 0 in3 / in3, silty clay loam = 0.43 in3 / in3, and sandy loam =
0.41 in3 / in3.
Estimate an average (i.e. representative) effective porosity for the study area using your answers to the previous
two questions. Don't spend too much time being overly precise; instead, quickly estimate the percentage of the
study area that is encompassed by each predominant soil texture and multiply by the corresponding effective
porosity. Once you've done that for each of the five predominant soil textures, sum the values.
Question 4: What is an average (i.e. representative) effective porosity for the study area?
Answer
Silt loam covers approximately 60% of the total area, loam covers approximately 17% of the total area, bedrock
covers approximately 10% of the total area, silty clay loam covers approximately 7% of the total area, and sandy
loam covers approximately 3% of the total area. Thus, (0.6 * 0.49 in3 / in3) + (0.17 * 0.43 in3 / in3) + (0.1 * 0.0 in3 / in3)
+ (0.07 * 0.43 in3 / in3) + (0.03 * 0.41 in3 / in3) = 0.41 in3 / in3.
Sand 0.03
Loam 0.12
Clay 0.27
Use Figure 3, Table 1, Table 3, and your answers to the previous questions to answer the following question.
Question 5: What is an average (i.e. representative) wilting point storage for the study area?
Answer
Silt loam covers approximately 60% of the total area, loam covers approximately 17% of the total area, bedrock
covers approximately 10% of the total area, silty clay loam covers approximately 7% of the total area, and sandy
loam covers approximately 3% of the total area. Thus, (0.6 * 0.13 in3 / in3) + (0.17 * 0.12 in3 / in3) + (0.1 * 0.0 in3 / in3)
+ (0.07 * 0.21 in3 / in3) + (0.03 * 0.1 in3 / in3) = 0.12 in3 / in3.
Finally, assume an active soil layer depth of 24 inches and use your answers to the previous questions to answer the
following question.
Question 6: What is an average (i.e. representative) maximum deficit for the study area?
Answer
This parameter can be estimated by taking the difference between the saturation storage and the wilting point
storage over an assumed active soil layer depth. Thus, (0.41 in3 / in3 - 0.12 in3 / in3 ) * 24 inches = 7.0 inches.
However, this parameter should be calibrated using observed data.
Constant Rate
The constant rate defines the rate at which precipitation will be infiltrated into the soil layer after the initial deficit
has been satisfied in addition to the rate at which percolation occurs once the soil layer is saturated. Typically, this
parameter is equated with the saturated hydraulic conductivity of the soil which is defined as the rate at which
water moves through a unit area of saturated soil in a unit time under a unit hydraulic gradient. This parameter can
also be estimated using surficial soil textures. The saturated hydraulic conductivity of various soil textures is shown
in Table 4.
Table 4. Soil Textures and Saturated Hydraulic Conductivity, reproduced from Rawls, Brakensiek, and Miller (1983)
Sand 4.6
Loam 0.1
Clay 0.01
Use Figure 3, Table 1, Table 4, and your answers to the previous questions to answer the following question.
Question 7: What is an average (i.e. representative) saturated hydraulic conductivity for the study area?
Answer
Silt loam covers approximately 60% of the total area, loam covers approximately 17% of the total area, bedrock
covers approximately 10% of the total area, silty clay loam covers approximately 7% of the total area, and sandy
loam covers approximately 3% of the total area. Thus, (0.6 * 0.3 in/hr) + (0.17 * 0.1 in/hr) + (0.1 * 0.0 in/hr) + (0.07 *
0.04 in/hr) + (0.03 * 0.4 in/hr) = 0.2 in/hr.
6. Create a new simulation run by clicking Compute | Create Compute | Simulation Run.
a. Name the new simulation run "Sept2018_DefConst".
b. Select the Mahoning Creek basin model.
c. Select the MRMS meteorologic model.
d. Select the September 2018 control specifications.
7. Select the Sept2018_DefConstsimulation run from the Compute toolbar,
To better match the observed runoff, the Deficit and Constant loss method parameters must be
c. calibrated.
precipitation on 17Sep2018.
c. When looking at the result graph and summary table for the Mahoning Creek subbasin element, the
computed hydrograph begins to rise at approximately the same time as the observed hydrograph, as
shown in Figure 7. However, the computed runoff volume (2.40 inches) is still much less than the
2. Continue modifying the loss parameters to approximately match the observed runoff volume.
a. Change the constant loss rate to 0.05 in/hr and rerun the simulation.
b. Notice that the computed runoff volume is now approximately equal to the observed runoff volume
(4.36 vs 4.4 inches), as shown in Figure 8. Also, the computed peak flow rate more closely matches
the observed peak flow rate. However, the shape of the computed runoff hydrograph doesn't match
the observed hydrograph shape on September 9th and 10th; in particular, the computed runoff is
much greater than the observed runoff within those two days.
c. To reduce the amount of runoff on September 9th and 10th, increase the initial deficit to 1.75 inches
and rerun the simulation.
d. Notice that the computed runoff volume is now less than the observed runoff volume (3.84 vs 4.4
inches), but the computed peak flow rate more closely matches the observed peak flow rate and the
statistical metrics are very good, as shown in Figure 9. Also, the shape of the computed hydrograph
matches the observed hydrograph much better on September 9th and 10th.
3. Continue adjusting the Deficit and Constant loss method parameters in an attempt to simultaneously best
match the peak flow rate, runoff volume, and hydrograph shape. Record your RMSE Std Dev, Percent Bias,
and Nash-Sutcliffe statistical metrics as you adjust model parameters.
a. At this point, it should be apparent that simultaneously matching peak flow rate, runoff volume, and
hydrograph shape can be difficult.
b. For instance, the "double peak" runoff response within the observed results between September 9th
- 11th is nearly impossible to recreate when using a single subbasin. This type of "signature"
suggests two discrete pulses of runoff. To better match this response, two or more subbasins should
be delineated.
67 https://www.hec.usace.army.mil/confluence/download/attachments/46007079/Deficit_and_Constant_finish.7z?
api=v2&modificationDate=1609890962926&version=1
Question 9. Is it possible to reduce the computed runoff response on September 17th - 18th while not
adversely affecting the computed runoff response on September 9th - 12th when using the Deficit and
Constant loss method? If not, why?
Answer
YES! When using the Deficit and Constant loss method, the moisture deficit is continuously tracked. Also, when
used in combination with a canopy method that will extract water from the soil in response to potential
evapotranspiration computed in the meteorologic model, the soil layer will lose moisture as the canopy extracts
infiltrated water.
6.4.7 Overview
In this tutorial you will apply the HEC-HMS Deficit and Constant loss method to a modeling application. Initial
parameter estimates will be estimated using GIS information and the model will be calibrated through trial and
error. HEC-HMS version 4.7 was used to created this tutorial. You will need to use HEC-HMS version 4.7, or newer, to
open the project files.
6.4.8 Background
The Deficit and Constant loss method is very similar to the Initial and Constant loss method in that a hypothetical
single soil layer is used to account for changes in moisture content. However, the deficit and constant method
allows for continuous simulation when used in combination with a canopy method that will extract water
from the soil in response to potential evapotranspiration computed in the meteorologic model. Between
precipitation events, the soil layer will lose moisture as the canopy extracts infiltrated water. Unless a canopy
method is selected, no soil water extraction will occur. This method may also be used in combination with a
surface method that will hold water on the land surface. The water in surface storage can infiltrate into the soil
layer and/or be removed through evapotranspiration. The infiltration rate is determined by the capacity of the soil
layer to accept water. When both a canopy and surface method are used in combination with the deficit constant
loss method, the system can be conceptualized as shown in Figure 1.
If the moisture deficit is greater than zero, water will infiltrate into the soil layer. Until the moisture deficit has been
satisfied, no percolation out of the bottom of the soil layer will occur. After the moisture deficit has been satisfied,
the rate of infiltration into the soil layer is defined by the constant rate. The percolation rate out of the bottom of
the soil layer is also defined by the constant rate while the soil layer remains saturated. Percolation stops as soon
as the soil layer drops below saturation (moisture deficit greater than zero). Moisture deficit increases in response
to the canopy extracting soil water to meet the potential evapotranspiration demand. Parameters that are required
to utilize this method within HEC-HMS include the Initial Deficit [inches or millimeters], Maximum Deficit [inches or
millimeters], and Constant Rate [in/hr or mm/hr]. The Directly Connected Impervious Area [percent] is an optional
parameter and can be specified by the user.
68 https://www.hec.usace.army.mil/confluence/download/attachments/46007050/Loss_Methods_start.7z?
api=v2&modificationDate=1609890225151&version=1
Initial Deficit
The initial deficit defines the volume of water that is be required to fill the soil layer at the start of the simulation
while the maximum deficit specifies the total amount of water the soil layer can hold. This parameter is most often
estimated using the product of the soil moisture state at the start of the simulation and an assumed active layer
depth. However, this parameter must be calibrated using observed data, must be less than or equal to the
maximum deficit, and must be specified as an effective depth (i.e. inches or millimeters).
In order to estimate an appropriate initial deficit, the antecedent conditions at the beginning of September 2018
must be investigated. Daily average humidity, daily precipitation accumulation, daily average temperature, and
daily average wind speed are shown in Figure 2.
Use Figure 2 and the following information to answer Question 1. Assume that the soil within the study area:
1. Was completely saturated by rainfall on August 22, 2018 and
2. Requires approximately 14 days with no rain to completely dry under normal, summer conditions
Question 1: Estimate an initial deficit (in inches) for the study area at the beginning of September 2018.
Answer
A reasonable initial deficit is approximately 2 inches. However, this parameter must be calibrated using observed
data.
Maximum Deficit
The maximum deficit is the driest a soil can become under the influence of gravity, evaporation, and transpiration.
This parameter is typically estimated as the difference between the saturation storage of the soil and the wilting
point storage over an assumed active soil layer depth. However, this parameter should be calibrated using
observed data, must be greater than or equal to the initial deficit, and also must be specified as an effective depth
(i.e. inches or millimeters).
Saturation Storage
Saturation storage refers to the amount of moisture that can be held within a soil column when all air bubbles have
been forced out. This parameter is a function of the effective porosity. To estimate a representative effective
porosity throughout the study area, Gridded Soil Survey Geographic69 (gSSURGO) data for Pennsylvania was
obtained from the U.S. Department of Agriculture's Geospatial Data Gateway70. Using GIS tools, surficial soil
textures were then extracted from the gSSURGO data and are shown in Figure 3. Also, the percent of the study area
encompassed by each soil texture is shown in Table 1.
69 https://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/home/?cid=nrcs142p2_053628
70 https://datagateway.nrcs.usda.gov/
Sand 5.2
Loam 17.1
Clay 0.9
These surficial soil textures can be used to estimate initial parameter values for nearly all loss methods within HEC-
HMS, including the Deficit and Constant loss method. For instance, Rawls, Brakensiek, and Miller
(1983)71 assembled data from thousands of soil samples located throughout the United States and related soil
textures to various useful parameters. The effective porosity of various soil textures is shown in Table 2.
Table 2. Soil Textures, Effective Porosity, and Wilting Point, reproduced from Rawls, Brakensiek, and Miller (1983)
Sand 0.42
Loam 0.43
71 https://ascelibrary.org/doi/10.1061/%28ASCE%290733-9429%281983%29109%3A1%2862%29
Clay 0.39
Question 3: What is the effective porosity for the top five predominant soil textures found when answering
Question 2?
Answer
Silt loam = 0.49 in3 / in3, loam = 0.43 in3 / in3, bedrock = 0 in3 / in3, silty clay loam = 0.43 in3 / in3, and sandy loam =
0.41 in3 / in3.
Estimate an average (i.e. representative) effective porosity for the study area using your answers to the previous
two questions. Don't spend too much time being overly precise; instead, quickly estimate the percentage of the
study area that is encompassed by each predominant soil texture and multiply by the corresponding effective
porosity. Once you've done that for each of the five predominant soil textures, sum the values.
Question 4: What is an average (i.e. representative) effective porosity for the study area?
Answer
Silt loam covers approximately 60% of the total area, loam covers approximately 20% of the total area, bedrock
covers approximately 10% of the total area, silty clay loam covers approximately 7% of the total area, and sandy
loam covers approximately 3% of the total area. Thus, (0.6 * 0.49 in3 / in3) + (0.2 * 0.43 in3 / in3) + (0.1 * 0.0 in3 / in3) +
(0.07 * 0.43 in3 / in3) + (0.03 * 0.41 in3 / in3) = 0.41 in3 / in3.
Sand 0.03
Loam 0.12
Clay 0.27
Use Figure 3, Table 1, Table 3, and your answers to the previous questions to answer the following question.
Question 5: What is an average (i.e. representative) wilting point storage for the study area?
Answer
Silt loam covers approximately 60% of the total area, loam covers approximately 20% of the total area, bedrock
covers approximately 10% of the total area, silty clay loam covers approximately 7% of the total area, and sandy
loam covers approximately 3% of the total area. Thus, (0.6 * 0.13 in3 / in3) + (0.2 * 0.12 in3 / in3) + (0.1 * 0.0 in3 / in3) +
(0.07 * 0.21 in3 / in3) + (0.03 * 0.1 in3 / in3) = 0.12 in3 / in3.
Finally, assume an active soil layer depth of 24 inches and use your answers to the previous questions to answer the
following question.
Question 6: What is an average (i.e. representative) maximum deficit for the study area?
Answer
This parameter can be estimated by taking the difference between the saturation storage and the wilting point
storage over an assumed active soil layer depth. Thus, (0.41 in3 / in3 - 0.12 in3 / in3 ) * 24 inches = 7.0 inches.
However, this parameter should be calibrated using observed data.
Constant Rate
The constant rate defines the rate at which precipitation will be infiltrated into the soil layer after the initial deficit
has been satisfied in addition to the rate at which percolation occurs once the soil layer is saturated. Typically, this
parameter is equated with the saturated hydraulic conductivity of the soil which is defined as the rate at which
water moves through a unit area of saturated soil in a unit time under a unit hydraulic gradient. This parameter can
also be estimated using surficial soil textures. The saturated hydraulic conductivity of various soil textures is shown
in Table 4.
Table 4. Soil Textures and Saturated Hydraulic Conductivity, reproduced from Rawls, Brakensiek, and Miller (1983)
Sand 4.6
Loam 0.1
Clay 0.01
Use Figure 3, Table 1, Table 4, and your answers to the previous questions to answer the following question.
Question 7: What is an average (i.e. representative) saturated hydraulic conductivity for the study area?
Answer
Silt loam covers approximately 60% of the total area, loam covers approximately 20% of the total area, bedrock
covers approximately 10% of the total area, silty clay loam covers approximately 7% of the total area, and sandy
loam covers approximately 3% of the total area. Thus, (0.6 * 0.3 in/hr) + (0.2 * 0.1 in/hr) + (0.1 * 0.0 in/hr) + (0.07 *
0.04 in/hr) + (0.03 * 0.4 in/hr) = 0.2 in/hr.
4.
6. Create a new simulation run by clicking Compute | Create Compute | Simulation Run.
a. Name the new simulation run "Sept2018_DefConst".
b. Select the Mahoning Creek basin model.
c. Select the MRMS meteorologic model.
d. Select the September 2018 control specifications.
7. Select the Sept2018_DefConst simulation run from the Compute toolbar,
To better match the observed runoff, the Deficit and Constant loss method parameters must be
c. calibrated.
precipitation on 17Sep2018.
c. When looking at the result graph and summary table for the Mahoning Creek subbasin element, the
computed hydrograph begins to rise at approximately the same time as the observed hydrograph, as
shown in Figure 7. However, the computed runoff volume (2.40 inches) is still much less than the
2. Continue modifying the loss parameters to approximately match the observed runoff volume.
a. Change the constant loss rate to 0.05 in/hr and rerun the simulation.
b. Notice that the computed runoff volume is now approximately equal to the observed runoff volume
(4.36 vs 4.4 inches), as shown in Figure 8. Also, the computed peak flow rate more closely matches
the observed peak flow rate. However, the shape of the computed runoff hydrograph doesn't match
the observed hydrograph shape on September 9th and 10th; in particular, the computed runoff is
much greater than the observed runoff within those two days.
c. To reduce the amount of runoff on September 9th and 10th, increase the initial deficit to 1.75 inches
and rerun the simulation.
d. Notice that the computed runoff volume is now less than the observed runoff volume (3.84 vs 4.4
inches), but the computed peak flow rate more closely matches the observed peak flow rate and the
statistical metrics are very good, as shown in Figure 9. Also, the shape of the computed hydrograph
matches the observed hydrograph much better on September 9th and 10th.
3. Continue adjusting the Deficit and Constant loss method parameters in an attempt to simultaneously best
match the peak flow rate, runoff volume, and hydrograph shape. Record your RMSE Std Dev, Percent Bias,
and Nash-Sutcliffe statistical metrics as you adjust model parameters.
a. At this point, it should be apparent that simultaneously matching peak flow rate, runoff volume, and
hydrograph shape can be difficult.
b. For instance, the "double peak" runoff response within the observed results between September 9th
- 11th is nearly impossible to recreate when using a single subbasin. This type of "signature"
suggests two discrete pulses of runoff. To better match this response, two or more subbasins should
be delineated.
72 https://www.hec.usace.army.mil/confluence/download/attachments/46007079/Deficit_and_Constant_finish.zip?
api=v2&modificationDate=1614786892841&version=1
Question 9. Is it possible to reduce the computed runoff response on September 17th - 18th while not
adversely affecting the computed runoff response on September 9th - 12th when using the Deficit and
Constant loss method? If not, why?
Answer
YES! When using the Deficit and Constant loss method, the moisture deficit is continuously tracked. Also, when
used in combination with a canopy method that will extract water from the soil in response to potential
evapotranspiration computed in the meteorologic model, the soil layer will lose moisture as the canopy extracts
infiltrated water.
7.1.1 Background
The unit hydrograph is a technique for modeling the transformation of excess precipitation to runoff at the
watershed scale. When the watershed is gaged with both observed rainfall and flow data available for a storm
event, the unit hydrograph relationship can be derived directly. When the watershed is ungaged, the unit
hydrograph relationship cannot be derived directly; therefore, many synthetic unit hydrograph methods have been
developed to help with ungaged watersheds (though not limited for use to just ungaged watersheds). There are
different synthetic unit hydrograph methods available in HEC-HMS, such as the Snyder and SCS Unit Hydrograph
Methods, for which the user can choose. In this tutorial, the Clark unit hydrograph (UH) method is applied.
7.1.2 Overview
In this tutorial, you will learn one method for estimating the time of concentration (Tc) and storage coefficient (R)
for Clark's synthetic unit hydrograph(see page 198) and understanding the impacts these parameters have to the
simulated hydrograph(see page 205). We will be estimating the Clark UH parameters for three subbasins in the
Punxsutawney watershed. There are various methods available to estimate Tc and R and some of the quickest
ways are to use empirically derived equations and/or regional regression equations. Regression equations are
developed for regions with specific types of terrain, land use, basin/river slopes, (i.e. rural agricultural land, small
mountains, air fields) and watershed sizes (i.e. areas less than 10 square miles). Care must be taken before applying
these equations to your study. Equations typically rely on extracting basin characteristics to estimate the Clark Unit
Hydrograph parameters. Once the appropriate basin characteristics are extracted, the Tc and R parameters can be
quickly calculated.
7.1.3 Goal
In this workshop you will:
• Estimate Clark's time of concentration (Tc) for three subbasins using regression equation.
• Estimate Clark's storage coefficient (R) for three subbasins.
• Compare the simulation results for the Punxsutawney watershed against observed flow data.
• Change Tc and R and understand their impacts to the simulate hydrograph
7.1.4 Data
You will be using the Punxsutawney Watershed HEC-HMS model to estimate Tc and R values. All model and
parameter components other than the Unit Hydrograph parameters have been configured for you.
7.2.1 Overview
where = time of concentration (hrs); longest flow path (mi); = Centroidal flow path (mi); =
average slope of the flow path represented by 10 to 85 percent of the longest flow path (ft/mi).
73 https://www.hec.usace.army.mil/confluence/download/attachments/42173766/Tc_R_Initial.zip?
api=v2&modificationDate=1616020779207&version=2
Once the time of concentration is estimated, the storage coefficient will be estimated using the following
relationship from nearby watersheds:
The 10-85 Flowpath Slope is given in ft/ft! You will need to convert your units to ft/mile
when computing the time of concentration
2. Extract the Longest Flowpath Length, Centroidal Flowpath Length, and the 10-85 Flowpath Slope for Stump
Creek, Mahoning Creek, and EB Mahoning Creek and
b. In the Component Editor, select the Transform tab and enter your Tc and R values for East Branch
Mahoning Creek. Leave the Method as "Standard" and the Time-Area Method as "Default".
Using the HEC-HMS Technical Reference Manual74, can you find the default time-area relationship
built into HMS?
The time-area relationship can be found here: Clark Unit Hydrograph Model75 under Estimating
Model Parameters
c. Repeat the above two steps for the remaining two subbasins.
4. Once you have your parameters entered, start a simulation by moving to your Compute tab, right click on
Sep2018, and click Compute.
74 https://www.hec.usace.army.mil/confluence/display/HMSTRM
75 https://www.hec.usace.army.mil/confluence/display/HMSTRM/Clark+Unit+Hydrograph+Model
a. Alternatively, you can also select the Sep2018 simulation on the toolbar and click on the compute
raindrop button.
5. Once the compute is complete, close the compute progress bar and navigate to the Results tabs. Under the
Simulation Runs folder, review your results by expanding the Sep2018 simulation and clicking through your
model elements. When you are done reviewing your results, click on the Mahoning Creek at Punx Gage
element and select Graph. The following graph should appear showing the combined simulated flow (blue
lines) compared with the observed flow (black dotted lines)
6. BONUS - re-compute Time of Concentration and Storage Coefficient using Expression Calculator.
a. Try using the Expression Calculator in the Clark Unit Hydrograph Global Editor to compute Tc and R
described here(see page 58).
7.2.4 Questions
1. What values did you receive for your Tc and R?
a. Using the regression equation and basin characteristics from HMS:
Subbasin Tc R
76 https://www.hec.usace.army.mil/confluence/download/attachments/42173766/Tc_R_final.zip?
api=v2&modificationDate=1615926652758&version=1
The simulated hydrograph matches pretty closely to the observed hydrograph! This is likely because other
model parameters have been calibrated for you.
7.3.1 Overview
77 https://www.hec.usace.army.mil/confluence/download/attachments/42174078/Tc_R_impacts_initial.zip?
api=v2&modificationDate=1614810609399&version=1
c. Create three more copies of the Sep2018 basin model and name it Sep2018 - half Tc, Sep2018 -
double R, Sep2018 - half R. You should now have five basin models listed
2. We will need to make sure the Meteorologic Model can work with our four additional basin models.
a. In the Components tab under the Meteorologic Models, select Sep2018 meteorologic model.
b. In the Components Editor below, select the Basins tab. Under Include Subbasins, ensure all Basin
Models are selected Yes
a. Navigate to your Compute tab, expand the Simulation Runs folder by clicking on the + sign, right
click on Sep2018 Simulation Run, and select Create Copy...
b. Name the copied simulation run Sep2018 - double Tc. Select this simulation run and change in the
Component Editor Basin Model to Sep2018 - double Tc
c. Create another copy of the simulation run and name it Sep2018 - half Tc. Repeat the previous step
above and select Sep2018-half Tc in the Basin Model.
d. Repeat for Sep2018 - double R and Sep2018 - half R
b. Click on the View Graphs for Selected Elements button to display the plots in a separate window
a. Navigate back to the Components tab. Select the Sep2018 - double R Basin model and expand the
element tree. For each subbasin, double the storage coefficient while keeping the time of
concentration values the same as Sep2018.
b. Do the same for the Sep2018 - half R Basin model but divide the storage coefficient for each
subbasin by half.
2. In the Compute tab (or using the compute toolbar) run your simulation for Sep2018 - double R and
Sep2018 - half R.
3. In the Results tab, review your simulation results by click through each element.
4. Plot the Outlet element for each simulation together on the same plot
a. Expand the Mahoning Creek Gage at Punx element for each of the Storage Coefficient simulations.
Click and hold the Ctrl button and select Outflow for each simulation to plot the outflow graph
together. You should three outflow graphs plotted. Use the magnify button and zoom into the
hydrograph starting from September 9 - 13.
7.3.5 Bonus:
1. If time allows, try other combinations of Tc and R and see how they change the simulated hydrograph. By
trying various combinations, you are essentially calibrating the model by attempting to match the observed
hydrograph.
2. Try another synthetic unit hydrograph method (SCS, Snyders)
7.3.6 Questions
1. What happens when you change your time of concentration in terms of hydrograph peak, shape, and
timing? In general, what might cause the time of concentration to be shorter or longer for any watershed?
How did the Tc changes compare to the observed hydrograph?
The time of concentration is the time it takes for water to travel from the most remote point in the
watershed to the watershed outlet (pourpoint). By reducing the time of concentration, the hydrograph
arrives at the outlet sooner. By increasing Tc, you see the hydrograph shift to the right meaning water arrives
at the watershed outlet later. The hydrograph peak and shape do change slightly with changes to Tc but the
most apparent changes is the timing of the hydrograph. Looking back at the initial Tc equation, the longest
flow path, centroidal flow path, and slope are basin parameters that could affect the time of concentration.
Other potential factors could be the basin and channel roughness, channel shape, or urbanization. By
increasing Tc, the simulated hydrograph appears to match the observed hydrograph much better.
2. What happens when you changed the storage coefficient in terms of hydrograph peak, shape, and timing?
How does increasing or decreasing R compare to the observed hydrograph?
The storage coefficient is a representation of temporary water storage. By reducing "R", the hydrograph
shape "narrows" (the rising limb and recession limb have a steeper slopel) causing the peak value to
increase while also reducing the time to peak. By increasing R, the hydrograph shape "flattens" or
attenuates which reduces the hydrograph peak and extends the recession limb of the hydrograph later in
time. Reducing R narrows the shape of the simulated hydrograph and increases the peak above the
observed hydrograph. Increasing R attenuates the hydrograph such that the hydrograph shape matches
pretty well but the peak value is much lower compared to the observed hydrograph.
78 https://www.hec.usace.army.mil/confluence/download/attachments/42174078/Tc_R_impacts_final.zip?
api=v2&modificationDate=1614812396367&version=1
Overview
In this tutorial group you will become familiar with different baseflow methods in HEC-HMS, separate examples for
using the constant monthly(see page 221), linear reservoir(see page 226), and recession baseflow(see page 237) methods
are included. There are questions at the end of each tutorial.
Background
In this tutorial group, you will be working with the Mahoning Creek watershed. The model was originally created
using terrain data downloaded from the USGS National Map Viewer. Soils and landuse information were used to
estimate initial model parameters, including loss and Clark unit hydrograph parameters.
The Mahoning Creek watershed at Punxsutawney, PA is part of the Allegheny River Basin located in western
Pennsylvania. The primary conveyance system of this 158 sq. mi. watershed includes the following streams: Stump
Creek, East Branch Mahoning Creek, and Mahoning Creek. The confluence of Stump Creek and East Branch
Mahoning Creek is located east of the enclave of Big Run. Mahoning Creek is downstream of the confluence. The
Data
The simulation time-window for this tutorial group is September 7, 2018 through the end of September 23, 2018.
Multi-RADAR/Multi-Sensor (MRMS) gridded precipitation is used as the boundary condition for the model. The
MRMS gridded dataset has a spatial resolution of 4 square kilometers and a time step of 1-hour. The monthly
average evapotranspiration method is used to compute canopy evapotranspiration to remove soil moisture. 15-
minute observed streamflow data was gathered from the USGS for the Mahoning Creek at Punxsutawney gage. The
data was shifted 5 hours to be on Coordinated Universal Time (UTC), also Greenwich Mean Time (GMT). By default,
MRMS data is saved on UTC. The following figure shows the observed streamflow data for the Mahoning Creek at
Punxsutawney gage for the simulation time period.
Model Configuration
The HEC-HMS project for this tutorial group has three basin models. Each basin model is named based on the
baseflow method used by the Mahoning Creek subbasin element. Each of the basin models is similar, the only
difference will be the baseflow method.
The following figure shows the Basin Model configuration. There is one subbasin element connected to a sink
element. Flow from the Mahoning Creek at Punxsutawney gage was linked to both the subbasin and the sink
elements. The subbasin element was configured to use the 2000-meter SHG structured discretization method, the
simple canopy method, the Deficit and Constant loss method, and the Clark unit hydrograph method.
Simulation runs have already been configured that link the three basin models with the meteorologic model and
control specifications. Initially, results at the Mahoning Creek Gage at Punxsutawney sink element will look like the
following figure. Notice how the computed flow is missing an interflow and baseflow portion of the runoff
hydrograph. The computed flow only includes direct runoff. The remaining tutorials in this group will show you how
to turn on the different baseflow methods, how to parameterize the baseflow model, and how to calibrate the
baseflow model.
8.2.1 Overview
In this tutorial you will apply the HEC-HMS constant monthly baseflow method to a modeling application. Initial
parameter estimates will be determined, and the model will be calibrated through trial and error. HEC-HMS version
4.7 was used to created this tutorial. You will need to use HEC-HMS version 4.7, or newer, to open the project files.
79 https://www.hec.usace.army.mil/confluence/download/attachments/41419158/initial%20baseflow.zip?
api=v2&modificationDate=1614788080441&version=2
4. Open the Component Editor for the Mahoning Creek subbasin element. Change the Baseflow Method from
None to Constant Monthly.
5. Go to the Baseflow tab to enter Constant Monthly parameters. Even though the simulation only simulates a
period in the month of September, you must define a constant baseflow rate for each month in the year.
Enter a constant flow rate of 500 cfs for September, enter a flow rate of 0 cfs for the other 11 months.
6. Re-run the simulation and look at the results plot and summary table. Notice baseflow in the plot and how
the addition of baseflow to direct runoff results in an increase in the simulated flow. The total runoff volume
increased as well. The direct runoff is 2.79 inches and the baseflow is 2.01 inches. A constant baseflow rate of
500 cfs overestimates baseflow at the beginning of the simulation but is a better approximation of baseflow
after the first and second events.
8.2.4 Questions:
1) Can you model both interflow and baseflow with the Constant Monthly baseflow method?
Answer
No, the constant monthly baseflow method can only model the longer responding baseflow portion of the
runoff hydrograph. For those watersheds where interflow is a large portion of the runoff hydrograph,
interflow would have to be included in direct runoff. This would require loss rates and unit hydrograph
parameters to be adjusted accordingly.
2) What type of applications would you use the Constant Monthly baseflow method?
Answer
As shown in the Punxsutawney model application, the constant monthly baseflow method is not able to
model both interflow and baseflow. The constant monthly baseflow method could be used for those
modeling applications where interflow is not a large contributor to total runoff. This baseflow method is not
designed to perform well for individual flood events during a continuous simulation. You can modify model
parameters for one hydrograph peak, but it is not possible to accurately simulate recession for multiple flood
hydrographs in a continuous simulation. Therefore, this method should be limited to event simulation.
8.3.1 Overview
In this tutorial you will apply the HEC-HMS linear reservoir baseflow method to a modeling application. Initial
parameter estimates will be determined, and the model will be calibrated through trial and error. HEC-HMS version
4.7 was used to create this tutorial. You will need to use HEC-HMS version 4.7, or newer, to open the project files.
80 https://www.hec.usace.army.mil/confluence/download/attachments/41419158/final%20baseflow.zip?
api=v2&modificationDate=1606315060356&version=1
81 https://www.hec.usace.army.mil/confluence/download/attachments/41419107/initial%20baseflow.zip?
api=v2&modificationDate=1614790987009&version=4
4. Open the Component Editor for the Mahoning Creek subbasin element. Change the Baseflow Method from
None to Linear Reservoir.
6. Re-run the simulation and look at the results plot and summary table. Notice baseflow in the plot and how
the addition of baseflow to direct runoff results in an increase in the simulated peak flow. The total runoff
volume increased as well. The direct runoff is 2.79 inches and the baseflow is 1.46 inches. Water that
infiltrates due to the constant loss rate now contributes to runoff from the subbasin element.
The next step is to identify the parameters to adjust and understand their impact on computed results.
Since the Linear Reservoir baseflow method conserves mass, the infiltration rate should be adjusted
along with the baseflow parameters when improving model results.
2. Change the Constant Loss Rate to 0.08 in/hr. Rerun the simulation. Notice how the simulated peak flow is
reduced while the baseflow hydrograph increased (less surface runoff and more baseflow when increasing
3. In the figure above, notice how the baseflow hydrograph peaks on September 10 while the total flow
hydrograph peaks on September 11. Baseflow should occur after direct runoff. Increasing the GW 1
Coefficient will shift the time of peak flow for the baseflow hydrograph. Increasing the number of Steps
(number of linear reservoir) has a more pronounced shift in the timing of the baseflow hydrograph. Increase
the GW 1 Steps from 1 to 2. Do not modify the GW 1 Coefficient. Rerun the simulation and look at the results
plot. The timing of the baseflow hydrograph looks better, and the shape of the recession curve matches the
4. The GW 1 and GW 2 Fractions can be used to shift infiltrated water between groundwater layers. Change the
GW 1 Fraction to 0.1 and the GW 2 Fraction to 0.9. Rerun the simulation and look at the results plot. Overall,
the results are not improved with the change. The simulated peak flow is reduced, while there is slight
improvement on the baseflow only portion of the hydrograph. More infiltrated water goes to the slower
responding GW 2 linear reservoir (90 percent to the GW 2 reservoir vs. 10 percent going to the GW 1 linear
reservoir). The GW 1 and 2 Fractions can be used to remove water to a deep, non-contributing aquifer, if
needed. The GW Fractions do not have to sum to a value of 1. If the Fractions do not sum to 1, then the
difference (1 - sum of the GW Fractions) is lost to a deep aquifer and will not contribute to runoff from the
subbasin.
5. Continue adjusting the Constant Loss Rate, and the GW 1 and 2 Fractions, Coefficients, and number of
Steps to calibrate the model. Record your Nash-Sutcliffe score as you adjust model parameters.
8.3.4 Questions:
1) Why should the linear reservoir groundwater coefficients be related to the Clark storage coefficient.
Answer
It is not necessary to link the linear reservoir groundwater coefficients to the Clark storage coefficient.
However, similar processes that impact storage, attenuation, and timing of surface water impacts the
response of interflow and baseflow. Interflow and baseflow happen throughout the watershed. Infiltrated
water moves through the soil and then exfiltrates into a small channel. Once in the channel network, water
flows in the stream network to the subbasin outlet. The path water takes after it exfiltrates into a channel is
similar to the path direct runoff takes. The difference is that groundwater flow / baseflow spends a portion of
the travel time in the soil, which is accounted for by increasing the baseflow coefficient or adding multiple
82 https://www.hec.usace.army.mil/confluence/download/attachments/41419107/final%20baseflow.zip?
api=v2&modificationDate=1606314320892&version=2
linear reservoirs (steps). Another reason to ratio unit hydrograph and baseflow storage coefficients is to
reduce the complexity of parameterizing a basin model with many subbasins. Using ratios is a quick way to
parameterize a model while linking parameters that are conceptually similar.
2) What were your final loss and groundwater parameters? What was the Nash-Sutcliffe score for the calibrated
model?
Answer
The final Constant Loss Rate was set to 0.07 in/hr. The following figure shows the final Linear Reservoir
Baseflow parameters. The Nash-Sutcliffe score was 0.932.
8.4.1 Overview
In this tutorial you will apply the HEC-HMS Recession baseflow method to a modeling application. Initial parameter
estimates will be determined, and the model will be calibrated through trial and error. HEC-HMS version 4.7 was
used to created this tutorial. You will need to use HEC-HMS version 4.7, or newer, to open the project files.
83 https://www.hec.usace.army.mil/confluence/download/attachments/41419175/initial%20baseflow.zip?
api=v2&modificationDate=1614791263281&version=2
3. View the result graph and summary table for the Mahoning Creek subbasin element. Notice the computed
flow does a good job matching the peak observed flow, but the interflow and baseflow portions of the
computed hydrograph do not match observed flow. The summary table shows a large difference in observed
and computed runoff volume (4.40 inches vs. 2.79 inches). Leave the summary table and plot open so you
can see results change as you add baseflow and modify baseflow parameters.
4. Open the Component Editor for the Mahoning Creek subbasin element. Change the Baseflow Method from
None to Recession.
5. Go to the Baseflow tab to enter Recession parameters. Change the Initial Type to Discharge per Area.
Enter an Initial Discharge of 0 CFS/MI2. Enter 0.8 as the Recession Constant. Enter a Ratio to Peak value of
0.2.
6. Re-run the simulation and look at the results plot and summary table. Notice baseflow in the plot and how
the addition of baseflow to direct runoff results in an increase in the simulated flow. The total runoff volume
increased as well. The direct runoff is 2.79 inches and the baseflow is 1.48 inches.
should notice a close match in observed and computed flow at the beginning of the simulation.
2. The Recession Constant parameter impacts the slope of the baseflow hydrograph. Change the Recession
Constant from 0.8 to 1.0 and rerun the simulation. You will see a constant/flat baseflow response, there is no
recession when the Recession Constant is 1.0. Notice the loss and baseflow volumes in the summary table.
The baseflow volume, 4.55 inches, exceeds the loss volume, 4.22 inches.
3. Change the Recession Constant to 0.4 and rerun the simulation. Notice the baseflow response now
simulates a recession curve (there is a decreasing flow rate). A value of 0.4 is too low, the recession is too
fast, but do not change the Recession Constant as we look at the Ratio to Peak in the next step.
4. The Ratio to Peak parameter determines when the baseflow recession curve is added to the total runoff
hydrograph. If the peak flow rate is 100 cfs and the ratio to peak is 0.2, then the recession baseflow curve will
turn on when the hydrograph reaches 20 cfs. In our September 2018 simulation, the peak flow is
approximately 7800 cfs on September 11. Looking at the observed flow hydrograph, it looks like the
inflection point in the recession limb occurs around 3000 cfs (at this point, interflow/baseflow become the
major contributors to total runoff and not direct runoff). Dividing 3000 cfs by 7800 cfs equals approximately
0.38. Change the Ratio to Peak parameter to 0.38 and rerun the simulation. You should see the baseflow
5. Continue adjusting the Recession Constant and Ratio to Peak parameters to calibrate the model. Record
your Nash-Sutcliffe score as you adjust model parameters.
8.4.4 Questions:
1) Can the Recession baseflow method model both interflow and baseflow?
Answer
No, the recession method is limited to modeling either interflow or baseflow because you can only define one
recession constant and ratio to peak parameters. You can choose to model interflow or baseflow, or attempt
to approximate both. The following figure shows results from the recession baseflow method in the example
application. Notice how the recession curve is not steep enough for interflow and is too steep for the slower
responding baseflow. We would need two sets of recession parameters to model interflow from September
11 through the September 13 and then baseflow from the September 13 through the September 17.
84 https://www.hec.usace.army.mil/confluence/download/attachments/41419175/final%20baseflow.zip?
api=v2&modificationDate=1606315848624&version=1
2) Does the Recession baseflow method conserve volume? What information should you always look for when using
the Recession baseflow method.
Answer
No, the recession baseflow method does not conserve volume. It is possible for the model to generate more
runoff volume than precipitation volume when using the recession method. This is why it is important to look
at the subbasin summary plot and make sure the precipitation, loss, direct runoff, baseflow, and total
discharge volumes are reasonable/appropriate. An important consideration is that the same recession
baseflow parameters are not transferable to a range of flood events, from mild to extreme. For example, a
ratio to peak parameter that might be appropriate for a minor flood event will not be appropriate for a larger
event as it will result in too much baseflow volume (0.2 X 1,000 cfs vs. 0.2 X 100,000 cfs is a different recession
curve and baseflow volume).
3) What were your final Recession baseflow parameters? What was the Nash-Sutcliffe score for the calibrated
model?
Answer
The final Recession Constant was 0.6 and the final Ratio to Peak was 0.4. The Nash-Sutcliffe score was 0.908
with the final recession baseflow parameters.
9.1.3 Background
In this tutorial group, the study area is designated as the Mahoning Creek watershed upstream of the town of
Punxsutawney, PA. Mahoning Creek is part of the Allegheny River Basin and is located in western Pennsylvania, as
shown in Figure 1.
The primary conveyance system of this 158 sq. mi. study area includes the following major streams: Stump Creek,
East Branch Mahoning Creek, and Mahoning Creek. The confluence of Stump Creek and East Branch Mahoning
Creek is located east of the enclave of Big Run. Mahoning Creek is downstream of the confluence. Elevations within
the study area range from approximately 1175 ft to 2260 ft as referenced to the North American Vertical Datum of
198885 (NAVD88), as shown in Figure 2.
The portion of Mahoning Creek shown within Figure 3 will be used within all channel routing tutorials. This
approximately 13 mile-long reach is located upstream of the U.S. Geological Survey (USGS) stream gage at
Punxsutawney86 and downstream of the Stump Creek and East Branch Mahoning Creek confluence.
85 https://geodesy.noaa.gov/datums/vertical/north-american-vertical-datum-1988.shtml
86 https://waterdata.usgs.gov/usa/nwis/uv?03034000
9.1.4 Data
A 1.5-day time period in April 1994 will be simulated within this tutorial.
Precipitation
No precipitation will be simulated within these tutorials.
Streamflow
A "known" hydrograph will be used as an upstream boundary condition to the reach shown within Figure 3. All of
the subsequent channel routing tutorials will route this hydrograph using various routing methods and
parameterizations. This "known" hydrograph was also routed using unsteady HEC-RAS computations to develop
an outflow hydrograph from the reach in question. This HEC-RAS outflow hydrograph will be used as an "observed"
hydrograph with which all subsequent results will be compared. These two hydrographs are shown and detailed
within Figure 4.
87 https://viewer.nationalmap.gov/basic/#/
The "Inflow" source element was set to use the previously-described upstream boundary condition (i.e. "known"
inflow hydrograph). The previously-mentioned HEC-RAS outflow hydrograph (i.e. "observed") was linked to the
sink element.
9.2.1 Overview
In this tutorial you will apply the HEC-HMS Lag routing method to a modeling application. Initial parameter
estimates will be estimated using GIS and observed information and the model will be calibrated through trial and
error. HEC-HMS version 4.7 was used to created this tutorial. You will need to use HEC-HMS version 4.7, or newer,
to open the project files.
9.2.2 Background
The lag routing method is the most simplistic hydrologic routing method available for use within HEC-HMS. This
method can only represent the translation of flood waves and does not include any representation of
attenuation or diffusion processes. Inflow to the reach is delayed in time by a user-specified amount of time.
Consequently, it is best suited to short stream segments with a predicable travel time that doesn't vary with
changing conditions. Parameters that are required to utilize this method within HEC-HMS include the initial
conditions and lag time [minutes].
Lag Time
Lag time is the amount of time (i.e. travel time) that the inflow hydrograph will be translated as it moves through
the reach. This parameter can be estimated in multiple ways including:
• Using known hydrograph data.
• Comparing flow length to a flood wave velocity.
• Using regression equations which were developed from observed data in a similar region.
88 https://www.hec.usace.army.mil/confluence/download/attachments/46008016/Reach_Routing_start.zip?
api=v2&modificationDate=1614786968249&version=1
where B = the top width of the water surface and = slope of the flow-stage rating curve; both of these variables
must be estimated for the flow rate in question and at a cross section that is representative of the routing
reach. The USGS maintains a flow-stage rating curve for the stream gage at Punxsutawney89 and is shown in Figure
2.
Use the previously-shown equations and Figure 2 to answer the following questions. Assume the following:
• The length of the reach within this example is 68860 feet.
• The slope of the rating curve shown in Figure 2 at a reference flow of 6500 cfs is approximately 1316 cfs / ft.
• The top width of Mahoning Creek at a reference flow of 6500 cfs is approximately 200 ft.
89 https://waterdata.usgs.gov/usa/nwis/uv?03034000
Question 4: For a reference flow of 6500 cfs, estimate a representative flood wave velocity (in ft / s) using the
Kleitz-Seddon Law.
Answer
Question 5: Using the previously-computed representative flood wave velocity, estimate a representative
travel time (in hours) for this event.
Answer
Question 6: Do you think a single, representative travel time can be estimated for both large and small flood
events? Do you have any reservations about the use of a single value to represent the travel time through a
reach?
Answer
Estimating a single, representative travel time for flood events large and small is difficult (if not impossible) in real
world streams. In general, the speed at which a flood wave moves depends upon the depth of flow. For instance,
within most stream systems, larger floods (with greater depths) propagate faster than smaller floods (which have
shallower depths). Also, the speed at which a flood wave moves through a reach depends upon the geometry of the
stream in question. Top width, flow area, effective roughness, and even the flow length (due to increasing/
decreasing sinuosity) of a stream can change with the magnitude of an event. As such, the travel time through a
reach depends upon the event and reach in question. Channel routing parameters, like travel time, must be
calibrated and validated based upon the model's intended use.
7. Create a new simulation run by clicking Compute | Create Compute | Simulation Run.
a. Name the new simulation run "April1994_Lag".
b. Select the Mahoning Creek basin model.
c. Select the NoRain meteorologic model.
d. Select the April 1994 control specifications.
8. Select the April1994_Lag simulation run from the Compute toolbar,
b. Notice the computed peak discharge, time to peak discharge, and discharge hydrograph shape differ
from the observed data. This is confirmed within the summary table.
c. To better match the observed discharge, the Lag routing method parameters must be calibrated.
90 https://www.hec.usace.army.mil/confluence/download/attachments/46008016/Lag_finish.zip?
api=v2&modificationDate=1614787163279&version=1
Answer
Using a lag time of 165 minutes results in a difference of approximately 1600 cfs, 75 minutes, and 141 ac-ft when
comparing the peak flow rate, time of peak, and volume of the computed and observed discharge hydrographs,
respectively. Also, these parameters result in an RMSE Std Dev, Percent Bias, and Nash-Sutcliffe statistical metrics
of 0.2, 1.26%, and 0.974, respectively, as shown in Figure 5. However, these are by no means the only parameter
combinations that can result in acceptable model calibration.
Question 8. Is it possible to reduce the computed discharge hydrograph peak flow rate when using the
Lag routing reach method? If not, why?
Answer
No. When using the Lag routing reach method, only translation processes are simulated. No attenuation or
diffusion of the inflow hydrograph will be simulated.
9.3.1 Overview
In this tutorial you will apply the HEC-HMS Muskingum channel routing method to a modeling application. Initial
parameter estimates will be estimated using GIS and observed information and the model will be calibrated
through trial and error. HEC-HMS version 4.7 was used to created this tutorial. You will need to use HEC-HMS
version 4.7, or newer, to open the project files.
9.3.2 Background
The Muskingum routing method uses a conservation of mass approach to route an inflow hydrograph. The
Muskingum method can also account for “looped” storage vs. outflow relationships that commonly exist in most
rivers (i.e. hysteresis). As such, this method can simulate the commonly observed increased channel storage during
the rising side and decreased channel storage during the falling side of a passing flood wave. To do so, the total
storage in a reach is conceptualized as the sum of prism (i.e. rectangle) and wedge (i.e. triangle) storage, as shown
in Figure 1. During rising stages on the leading edge of a flood wave, wedge storage is positive and added to the
prism storage. Conversely, during falling stages on the receding side of a flood wave, wedge storage is negative and
subtracted from the prism storage. Through the inclusion of a travel time for the reach and a weighting between
the influence of inflow and outflow, it is possible to approximate attenuation.
Parameters that are required to utilize this method within HEC-HMS include the initial condition, K [hours], X, and
the Number of Subreaches.
K
The Muskingum K parameter is equivalent to the travel time through the reach. This parameter can be estimated in
multiple ways including:
• Using known hydrograph data.
• Comparing flow length to a flood wave velocity.
• Using regression equations which were developed from observed data in a similar region.
91 https://www.hec.usace.army.mil/confluence/display/HMSGUIDES/Applying+the+Lag+Routing+Method?
preview=%2F46008016%2F52758161%2FReach_Routing_start.zip
The inflow hydrograph reaches 6500 cfs at approximately 4/11/1994 17:00. The outflow hydrograph reaches 6500
cfs at approximately 4/11/1994 19:30. The difference in time between these two points is 2.5 hours.
where B = the top width of the water surface and = slope of the flow-stage rating curve; both of these variables
must be estimated for the flow rate in question and at a cross section that is representative of the routing
reach. The USGS maintains a flow-stage rating curve for the stream gage at Punxsutawney92 and is shown in Figure
3.
Use the previously-shown equations and Figure 3 to answer the following questions. Assume the following:
92 https://waterdata.usgs.gov/usa/nwis/uv?03034000
Question 5: Using the previously-computed representative flood wave velocity, estimate a representative
travel time (in hours) for this event.
Answer
Question 6: Do you think a single, representative travel time can be estimated for both large and small flood
events? Do you have any reservations about the use of a single value to represent the travel time through a
reach?
Answer
Estimating a single, representative travel time for flood events large and small is difficult (if not impossible) in real
world streams. In general, the speed at which a flood wave moves depends upon the depth of flow. For instance,
within most stream systems, larger floods (with greater depths) propagate faster than smaller floods (which have
shallower depths). Also, the speed at which a flood wave moves through a reach depends upon the geometry of the
stream in question. Top width, flow area, effective roughness, and even the flow length (due to increasing/
decreasing sinuosity) of a stream can change with the magnitude of an event. As such, the travel time through a
reach depends upon the event and reach in question. Channel routing parameters, like travel time, must be
calibrated and validated based upon the model's intended use.
X
The Muskingum X parameter is a dimensionless coefficient that lacks a strong physical meaning. This parameter
must range between 0.0 (maximum attenuation) and 0.5 (no attenuation). When this parameter is set to a value of
0, storage within the reach is computed solely as a function of outflow. This is equivalent to level pool routing and
results in the maximum possible amount of attenuation. When this parameter is set to a value of 0.5, equal weight
is given to both inflow and outflow when determining storage within the reach. This results in no attenuation to the
inflow hydrograph as it progresses through the reach. For most applications, an initial estimate of 0.25 is further
refined through model calibration.
Number of Subreaches
The number of subreaches parameter affects attenuation. One subreach results in the maximum amount of
attenuation and increasing the number of subreaches approaches zero attenuation. An initial estimate of this
parameter can be obtained by dividing the Muskingum K parameter by the simulation time step, :
For natural channels that vary in cross-section dimension, slope, and storage, the number of subreaches can be
treated as a calibration parameter. The number of subreaches may be used to introduce numerical attenuation
which can be used to better represent the movement of flood waves through the natural system.
Question 7: Assuming a simulation time step of 15 minutes, use your answers to the previous questions in
addition to the above equation to estimate the number of routing reaches for the event in question.
Answer
7. Create a new simulation run by clicking Compute | Create Compute | Simulation Run.
a. Name the new simulation run "April1994_Muskingum".
c. To better match the observed discharge, the Muskingum routing method parameters must be
calibrated.
b. When looking at the result graph and summary table for the Mahoning Creek reach element, the
computed hydrograph begins to rise at approximately the same time as the observed hydrograph, as
shown in Figure 6. However, the computed peak discharge (12610 cfs) is still greater than the
observed peak discharge (11293 cfs) so further modifications are required.
2. Continue modifying the routing parameters to approximately match the peak discharge of the observed
hydrograph.
a. Change the Muskingum X parameter to 0.1, change the number of subreaches to 6, and rerun the
simulation.
b. Notice that the computed peak discharge is now closer to the observed peak discharge (12217 vs
11293 cfs), as shown in Figure 7. However, the computed peak discharge is still greater and occurs
sooner than the observed discharge hydrograph.
3. Continue adjusting the Muskingum routing method parameters in an attempt to simultaneously best match
the peak discharge flow rate, time of peak discharge, hydrograph shape, and discharge volume. Record the
various differences between the computed and observed discharge hydrographs in addition to the RMSE Std
Dev, Percent Bias, and Nash-Sutcliffe statistical metrics as you adjust model parameters.
93 https://www.hec.usace.army.mil/confluence/download/attachments/46008030/Muskingum_finish.zip?
api=v2&modificationDate=1614787201027&version=1
Question 8. Based on your parameter modification of the Muskingum K and X parameters to match the results
from HEC-RAS, how much confidence do you have in the Muskingum method in ungaged areas?
Answer
In ungaged areas, the Muskingum coefficients are somewhat uncertain. This is because both of the parameters K
and X are difficult to estimate accurately using only physical characteristics of the channel. Also, the physical
characteristics that are used are taken as averages for an entire reach. However, even excluding this effect, there
are still other problems with the Muskingum method which are discussed in the next question.
Question 9. As you can imagine, the travel time estimated for the Muskingum method will be sensitive to
channel characteristics such as roughness, slope, length, and cross section. If you consider routing a range of
flows from small to large, how accurate do you think it is to use a single set of K and X coefficients for the
Muskingum method?
Answer
The Muskingum linear storage routing coefficient, K, can cause problems for channel reaches that have cross
sections with significant overbanks. K represents a linear relation between storage and outflow. A single flow value
is normally assumed when estimating a flood wave velocity, which is then used in the estimation of K. In general,
velocities in channels will vary with water depth. This is particularly so in a stream that has significant overbank
areas. As the flow goes out of the channel and into a wide overbank, the depth-area relationship changes
significantly. Because the Muskingum method reflects a linear storage-outflow relationship, it cannot accurately
represent flows that range from in-channel to floodplain conditions. Generally, the larger the floodplain, the
greater the difference between K for in-channel and K for flood flow conditions.
Additionally, the Muskingum X coefficient will also vary with different flows. One flood that remains within the
banks of the channel may require a relatively high Muskingum X coefficient (for example 0.4). Meanwhile, a larger
flood that goes out of bank may require a lower Muskingum X coefficient (0.015 for example).
9.4.1 Overview
In this tutorial you will apply the HEC-HMS Muskingum-Cunge channel routing method to a modeling application.
Initial parameter estimates will be estimated using GIS and observed information and the model will be calibrated
through trial and error. HEC-HMS version 4.7 was used to created this tutorial. You will need to use HEC-HMS
version 4.7, or newer, to open the project files.
9.4.2 Background
The Muskingum-Cunge routing method builds upon concepts within the Muskingum method(see page 261). This
method uses a combination of the continuity equation and a simplified form of the momentum equation. Within
the previously mentioned Muskingum method, the X parameter is a dimensionless coefficient that lacks a strong
physical meaning. Cunge (1969) evaluated the numerical diffusion that is produced through the use of the
Muskingum routing equation and set this equal to the physical diffusion represented by the convective diffusion
equation (Miller & Cunge, 1975). The Muskingum-Cunge method is sometimes referred to as a “variable coefficient”
method since the routing parameters are recalculated every time step based upon channel properties and the flow
depth.
Parameters that are required to utilize this method within HEC-HMS include the initial condition, the reach length [ft
or m], the friction slope [ft/ft or m/m], Manning’s n roughness coefficient, a space-time interval method and value, an
index method and value, and a cross-section shape and parameters/dimensions. An optional invert can also be
specified.
Reach Length
The reach length should be set as the total length of the reach element. This parameter is typically estimated using
GIS information and field survey data, when available. The total length of the reach in question is 68861 ft and is
shown in Figure 1.
94 https://www.hec.usace.army.mil/confluence/download/attachments/46008051/MuskingumCunge_start.zip?
api=v2&modificationDate=1614788980328&version=1
Friction Slope
The friction slope should be set as the average friction slope for the entire reach. If the friction slope varies
significantly throughout the stream represented by the reach, it may be necessary to use multiple reaches with
different slopes. If no information is available to estimate the friction slope, the bed slope can be used as an
approximation. This parameter is typically estimated using GIS information and field survey data, when
available. In this tutorial, invert elevations were extracted from five locations along the reach of interest and are
shown in Figure 2. The profile data for this information is detailed within Table 1.
Feet Miles
5 Punxsutawney 0 0 1200.8
Use Figure 1 and Table 2 to answer the following questions.
Question 1: Estimate a single, representative slope (in ft/ft) for the reach in question.
Answer
Question 2: Divide the reach into three sub-reaches and estimates a slope (in ft/ft) for each.
Answer
Subreach 1: Point 1 (Conf. with Stump Creek) to Point 2 (Robertsville): slope = 0.001 ft/ft
Subreach 2: Point 2 (Robertsville) to Point 4 (Riker): slope = 0.0022 ft/ft
Subreach 3: Point 4 (Riker) to Point 5 (Punxsutawney): slope = 0.0007 ft/ft
Space-Time Interval
The choice of space and time steps (Δx and Δt, respectively) are critical to ensure accuracy and stability. Three
options for specifying a space-time method are provided within HEC-HMS:
• Auto DX Auto DT
• Specified DX Auto DT
• Specified DX Specified DT
When the Auto DX Auto DT method is selected, space and time intervals will automatically be selected that attempt
to maintain numerical stability. Δt is selected as the minimum of either the user-specified time step or the travel
time through the reach (rounded to the nearest multiple or divisor of the user-specified time step). Once Δt is
computed, Δx is computed as:
where c = flood wave celerity. Upon completion of a simulation, the minimum and maximum celerity of the routed
hydrograph will be displayed as notes. Also, a reference space step, Δxref, will be computed using methodology
presented in Engineer Manual 1110-2-1417 Flood-Runoff Analysis95 (U.S. Army Corps of Engineers, 1994). If Δxref is
not equal to the Δx that was used during the routing computations, a note will be issued. For most applications,
the Auto DX Auto DT method is adequate.
Index Method
The index method is used in conjunction with the physical properties of the channel and the previously mentioned
Δx and Δt interval selection to discretize the routing reach in both space and time. Two options for specifying the
index method are provided within HEC-HMS:
• Celerity
• Flow
When the Celerity index method is used, a reference celerity [ft/s or m/s] must be specified. When the Flow index
method is used, a reference flow [cfs or cms] must be specified and cross-section properties are then used to infer a
celerity. Appropriate reference flows and flood wave celerities are dependent upon the physical properties of the
95 https://www.publications.usace.army.mil/LinkClick.aspx?fileticket=VMW6T8G3GDo%3d&tabid=16439&portalid=76&mid=43544
channel as well as the flood event(s) in question. Experience has shown that a reference flow (or celerity) based
upon average values of the hydrograph in question (i.e. midway between the base flow and the peak flow) is, in
general, the most suitable choice. Reference flows (or celerities) based on peak values tend to numerically
accelerate the flood wave more than would occur in nature while the converse is true if a low reference flow (or
celerity) is used (Ponce, 1983).
For most applications, using the Celerity index method and a celerity of 5 ft/s is adequate.
Cross Section
Six options are provided for specifying the cross-section shape: circle, eight-point, rectangle, tabular, trapezoid, and
triangle. The circle shape is not meant to be used for pressure flow or pipe networks, but is suitable for
representing a free surface inside a pipe. Depending upon the chosen shape, additional information will have to be
entered to describe the size of the cross-section shape. This information may include a diameter (circle) [ft or m],
bottom width (deep, rectangle, and/or trapezoid) [ft or m], or side slope (trapezoid and triangle) [ft/ft or m/m]. In
all cases, cross-section shapes must be defined in such a way that all possible flow depths that will be simulated
will be completely confined within the defined shape. The tabular shape option allows for the use of user-defined
elevation vs. discharge, elevation vs. area, and elevation vs. width relationships (which are all paired data objects).
This option is typically used when relationships derived from hydraulic simulations are available. When the Tabular
shape is selected, no Manning's n roughness coefficients need to be entered. When using the eight-point shape, a
simplified cross-section (which is a paired data object) with eight station vs. elevation values must be selected. The
cross-section is typically configured to represent the main channel plus left and right overbank areas. As such,
separate Manning's n values are required for each overbank.
Cross section shapes and properties are typically estimated using GIS and field survey data, when available. In this
tutorial, cross sections were extracted at three locations along the reach of interest, as shown in Figure 3. The
station-elevation data for these cross sections are detailed within Figure 3, Figure 4, and Figure 5 in addition
to Table 2, Table 3, and Table 4.
0 35.1
7.5 26.2
14.1 19.7
20 15.1
39.4 15.1
61.4 13.1
75.5 14.4
106.6 14.8
120.1 15.1
124.3 7.9
128.9 4.3
134.8 1
144 0.7
160.8 1.3
189 0.3
210 0
213.6 1
219.2 6.6
223.1 13.1
224.1 15.1
232.9 16.1
242.8 15.1
255.9 13.8
305.1 15.1
319.9 15.4
321.9 19.7
324.8 26.2
339.9 35.1
0 35.1
7.5 26.2
14.1 19.7
20 15.1
39.4 14.4
75.8 13.1
121.4 14.4
156.5 13.8
220.1 14.1
227.7 9.8
229.7 4.3
234.9 1
255.9 0.7
269 0
285.8 1.3
297.6 1
315 2
321.5 6.6
328.1 13.1
340.6 15.1
360.9 14.8
393.7 14.1
426.5 16.1
479 14.4
499.3 16.7
502 21.7
541.3 26.2
549.9 35.1
0 35.1
7.5 26.2
14.1 19.7
20 15.4
46.3 15.4
69.6 13.1
88.3 14.1
106.6 16.1
122.7 15.1
124.3 7.9
128.9 4.3
127.6 1
134.8 0
158.1 1
178.8 1.6
201.4 2
211 3.6
219.2 6.6
223.1 13.1
225.1 15.1
244.8 15.4
255.9 15.1
273.3 14.4
305.1 15.1
315.9 16.4
321.9 19.7
338.9 26.2
345.1 35.1
6. Create a new simulation run by clicking Compute | Create Compute | Simulation Run.
a. Name the new simulation run "April1994_MuskCunge".
b. Select the Mahoning Creek basin model.
c. Select the NoRain meteorologic model.
d. Select the April 1994 control specifications.
7. Select the April1994_MuskCunge simulation run from the Compute toolbar,
estimates from GIS and field surveys. This is confirmed within the summary table.
b. However, an improved estimate can be made by subdividing the routing reach into three separate
reaches. This has already been done for you within the Mahoning Creek Multi Reaches basin
model. Further modifications will be made using that basin model.
j. The Muskingum-Cunge Component Editor for the Upper reach element should resemble Figure 8.
j. The Muskingum-Cunge Component Editor for the Middle reach element should resemble Figure 9.
j. The Muskingum-Cunge Component Editor for the Lower reach element should resemble Figure 10.
5. Create a new simulation run by clicking Compute | Create Compute | Simulation Run.
a. Name the new simulation run "April1994_MuskCunge_multi".
b. Select the Mahoning Creek Multi Reaches basin model.
c. Select the NoRain meteorologic model.
d. Select the April 1994 control specifications.
6. Select the April1994_MuskCunge_multi simulation run from the Compute toolbar,
summary table.
9. Continue adjusting the Muskingum routing method parameters in an attempt to simultaneously best match
the peak discharge flow rate, time of peak discharge, hydrograph shape, and discharge volume. Record the
various differences between the computed and observed discharge hydrographs in addition to the RMSE Std
Dev, Percent Bias, and Nash-Sutcliffe statistical metrics as you adjust model parameters.
96 https://www.hec.usace.army.mil/confluence/download/attachments/46008051/MuskingumCunge_finish.zip?
api=v2&modificationDate=1614788797241&version=1
Question 4. Do you consider the differences in the routed hydrographs to be significant? How would you
quantify the significance from a study perspective?
Answer
There is an observable difference between the results obtained through the use of one reach when compared
against to those obtained through the use of three reaches. However, the significance of the differences are
dependent upon the study objective and the study's sensitivity. In some cases, such as one that analyzes possible
levee configurations, the significance could be evaluated by examining the corresponding differences in peak stage
given a fixed stage-flow rating curve. In other cases, such as storage facility analysis, the differences in discharge
volume are more critical to the study's success.
Question 5. If the answer to the previous question is yes, you might be inclined to subdivide the larger
routing reach into smaller subreaches. How would you know when to stop subdividing the reach in
subreaches?
Answer
You would only need to subdivide until the simulated hydrograph stopped changing significantly. Again, the
sensitivity of the particular study objective will determine the point at which the changes in results cease to be
important.
10.1.1 Background
Calibrating a hydrologic model is the process of modifying parameters within acceptable ranges to obtain
simulated results that replicate known conditions. Model calibration is a necessary component to model
development in order to have some level of confidence the simulated results adequately represents the modeled
system. Since models can be used to predict future outcomes or potential changes to the watershed, the model
results must show it works. In order to calibrate a model, observed data, typically collected in the field, is needed
to compare the simulated results. Observed data typically used for calibration are flow discharges and stages.
10.1.2 Overview
In this tutorial you will learn how to perform model calibration as an iterative process of entering and adjusting
parameters and evaluating goodness-of-fit between the computed and observed hydrograph. In the tutorial
Estimating Clark Unit Hydrograph Parameters(see page 197), you got a taste of what model calibration looks like by
increasing and decreasing your unit hydrograph parameters. In other tutorials, you learned about different loss(see
page 129) and baseflow(see page 216) methods and how to initialize those parameters. We will incorporate lessons and
techniques learned from those tutorials to calibrate our basin model. The process of calibration will be iterative,
and include manually changing model parameters and checking how those adjustments impact the model's ability
to match observed data (flow hydrograph). There are other methods of calibration such as using optimization tools,
however, those will not be covered in this tutorial.
10.1.3 Goal
In this tutorial you will
• Enter initial model parameters in HEC-HMS
• Adjust model parameters and compare results
• Calibrate the model to another storm event
10.1.4 Data
You will be using the Punxsutawney Watershed HEC-HMS for this calibration effort. The meteorological model
contains gridded rainfall from MRMS for 2018 and has already been configured for you. The observed hydrograph
has also been downloaded from the USGS and prepared for you to compared your results.
10.2.1 Overview
You will be inputting initial values to each model parameter. These initial values have already been calculated and
computed for you. You can review how these initial values were estimated from previous tutorials.
97 https://www.hec.usace.army.mil/confluence/download/attachments/42176475/Punx_Parameters.xlsx?
api=v2&modificationDate=1610498648833&version=2
98 https://www.hec.usace.army.mil/confluence/download/attachments/42176475/Calibration%20Initial.zip?
api=v2&modificationDate=1616037967744&version=2
6. Add observed hydrograph to Mahoning Creek Gage at Punx. The observed dataset has already been
prepared for you under Time-Series Data → Discharge Gages → Punxsutawney
a. In the Basin Model Sep2018, select the Mahoning Creek Gage at Punx element
b. In the Component Editor, navigate to the Options tab and select Punxsutawney for Observed Flow.
7. Create a simulation run.
a. Navigate to the Compute tab
b. Click on Compute → Create Compute → Simulation Run in the File Menu
c. Name your simulation Sep2018. Select your Basin Model, Meteorological Model, and Control
Specifications in the Create a Simulation Run Wizard
d. Right click Sep2018 → Compute
Download the final project file here: Calibration - Initial Param Populated.zip99
10.2.3 Questions
1. How were each of the initial parameters calculated?
Review excel spreadsheet on Unit Hydrograph and Linear Reservoir Baseflow parameters.
2. How do the results look compared to the simulated results?
Simulated hydrograph has a larger volume than the observed hydrograph. Runoff initiates sooner and peak
timing occurs 6 hours earlier.
10.3.1 Overview
We will start adjusting our parameters to better match the observed hydrograph. The tutorial below will walk you
through one way of systematically calibrating your model to match the observed hydrograph. We will work through
adjusting the losses, then move onto adjusting your unit hydrograph parameters, then lastly your baseflow
parameters. Other hydrologists and engineers may have their own calibration techniques.
Keep in mind that calibration is an iterative process! Just because one parameter looks good after your first
adjustment doesn't mean you won't need to go back and refine the parameter value again; however, calibration
should be more than just adjusting parameters to best match a curve. It is also important to have a good
understanding of the watershed hydrology and know what are the reasonable ranges for model parameters. If your
watershed consists of mainly sandy, well draining soils, it wouldn't make much sense to set your loss rates to 0
inches per hour just so you can match your hydrograph volume. Have a reason to why you are making your
adjustment. Parameter values should make sense to the watershed and the types of events you are modeling. This
will allow you to have confidence in your model results and defend your modeling results when you're able to
explain the why behind the values. When you start needing to adjust your model beyond reasonable limits, this
may suggest issues with your model boundary conditions, your model set up, or something more fundamental
about your basin that isn't being captured.
Download the Initial project files here - Calibration - Initial Param Populated.zip100
99https://www.hec.usace.army.mil/confluence/download/attachments/42176475/Calibration%20-%20Initial%20Param%20Populated.zip?
api=v2&modificationDate=1614722674003&version=1
0 https://www.hec.usace.army.mil/confluence/download/attachments/42176478/Calibration%20-%20Initial%20Param%20Populated.zip?
10
api=v2&modificationDate=1614721869211&version=1
c. By visually inspecting the hydrograph, you can easily observe the simulated flow has a higher volume
and peak value than the simulated hydrograph. You can confirm the high volume by checking the
differences in the Summary Results table (Compute Volume to the Observed Volume). We'll start the
calibration by adjusting the Deficit and Constant Loss rate.
d. Open the global editor for the Deficit and Constant Method by selecting on the File Menu
Parameters → Loss → Deficit and Constant.
The Deficit and Constant global editor should appear. Make sure you have your basin model
Sep2018 selected before opening the editor.
We currently have initial loss set at "0". This means there is no initial deficit of water that needs to be filled
prior to runoff initiation other than the constant loss. It is unlikely for the ground storage to be fully
saturated given the little rainfall from the prior 8 days. Adjust the initial deficit to the point where the
observed hydrograph begins. In the figure below, initial loss was changed to 1.5 inches for all subbasins.
Re-run your simulation and compare the results.
a. Select the File Menu Parameters → Transform → Clark Unit Hydrograph. The Clark Unit Hydrograph
global editor should appear. Make sure you have your basin model Sep2018 selected before opening
the editor.
b. Compare your simulated hydrograph peak to the observed flow. The simulated peak occurs earlier
than the observed peak. Which parameter do you think you should adjust?
Time of Concentration! Try adjusting your time of concentration until you get a good match to the
timing. Since you have three subbasins that contribute to this outlet, consider adjusting all the Tc
values by the same factor. Adjusting models by the same factor can be a useful technique when you
do not have additional information to suggest otherwise. If you are fortunate to have another flow
gage or highwater mark on an upstream tributary, this information should be used to make unique
subbasin parameter adjustments. In the figure below, the Tc was increased by a factor of 2.2 for all
subbasins.
c. The storage coefficient has a large effect on the shape of the hydrograph. The simulate hydrograph
appears to reasonably match the shape of the observed hydrograph but could be adjusted slightly.
Change the storage coefficient by a factor of 1.2 for all three subbasins.
a. Select the File Menu Parameters → Baseflow → Linear Reservoir Baseflow. The Linear Reservoir
Baseflow global editor should appear. Make sure you have your basin model Sep2018 selected
before opening the editor.
b. Compare your simulated hydrograph recession limb to the observed recession.
It appears that the baseflow could be extended out further past the inflection point. What parameter
within the Linear Baseflow Method would you adjust?
GW 1 and GW 2 Coefficients! These coefficients govern the travel time of your interflow (GW 1) and
deeper baseflow (GW 2). Triple both your GW 1 and GW 2 coefficients for all subbasins and review
results. Iterate as necessary.
to revisit our loss rate parameters and see if reducing our constant loss rate might help. Unless there is a
compelling reason, avoid making drastic adjustments to one parameter to compensate for an adjustment to
another. Continue to keep in mind reasonable parameter ranges as you make your adjustments and that the
parameters you are selecting make sense for your basin.
If you haven't already noticed, your Summary Results provides useful statistical metrics to assist in calibration. A
useful metric commonly used is the Nash-Sutcliffe Efficiency (NSE), which measures the predictive skill of a
hydrologic model. A value of "1" means the simulated values can perfectly predict the observed where there is no
estimation error variance, while a value of "0" means the estimation error variance is equal to the error variance in
the observed data. As you were making adjustments to your model, you may have noticed the NSE increasing,
suggesting that your model predictions are improving. Moriasi, et. al, 2007 developed a performance rating table
based on commonly used statistical measures to evaluate model performance.
This can be one of several methods to assess model performance. The NSE and the other metrics provided are
useful indicators on how your model prediction is performing; however, similar to the issue with curve fitting, the
modeler should not focus solely on maximizing the NSE or other single metric. The calibration process should also
take into account the needs of the project (i.e. concerned with peak flow, volume, or both) and you're
understanding of the watershed through research and analysis.
1. Refine parameters
a. Navigate back to your Deficit and Constant loss rate editor. The Summary Results shows simulated
volumes are low. What parameter could we consider adjusting?
Consider reducing your constant loss rate some more. We made changes to the initial deficit initially
but didn't make any changes to the constant rate. Reduce your initial Deficit to 1.8 inches and your
constant loss rate to 0.05 inches per hour.
b. Reducing the constant loss rate helped to better match the hydrograph peak flow. Review your
Summary Results. How do the simulated results compare to the observe results?
Simulated results have improved significantly compared to the initial results. The peak timing, peak
flow, simulate volume, and NSE values are within a reasonable range.
c. Feel free to adjust other parameters and see if you could further improve the simulated results
10.4.1 Overview
Now that you've calibrated one basin model, use your newfound talents to calibrate another event. It is good
practice to calibrate your hydrologic model to more than 1 storm event. Model parameters are not static values
that exist without variability. On the contrary, given the heterogeneity that exists within the watershed basin with
differing soil types, changes in land use and vegetation as well as differing antecedent soil conditions, model
parameters and their associated watershed response are expected to change for different storm events. This is
another important reason to understand how your model predictions will be used and to tailor your calibration
events to reflect what will be predicted. For example, if you are interested in estimating the 1% annual chance
exceedance event, you'll want to look at storm events that produced high flows and not storms that produce
average flows.
Another large storm event within the Mahoning watershed occurred in April 1994. In this tutorial, you'll use what
you learned in the last lesson to calibrate Mahoning Creek watershed. A meteorological model has already been set
up for you and you will not need to make any adjustments. You'll notice that the Meteorological method uses Gage
Weights method rather than Gridded Precipitation. In many cases, gridded precipitation will be harder to come by
as you go further back in time. In such cases, point rain gages may be the only rainfall boundary available for
calibration.
101 https://www.hec.usace.army.mil/confluence/download/attachments/42176478/Calibration%20-%20solution.zip?
api=v2&modificationDate=1614720646438&version=1
102https://www.hec.usace.army.mil/confluence/download/attachments/43811373/Calibration%20-%202nd%20event.zip?
api=v2&modificationDate=1614723272990&version=1
ii. Increased GW 1 Initial to 2 CFS/MI2 and GW 2 Initial to 3 CFS/MI2. This improved the initial
discharge at the beginning of the simulation
ii. Reduced Time of Concentration for all subbasins by a factor of 0.4. This improved the timing
of the simulated hydrograph
iii. The hydrograph shape matches reasonably well. Adjust the Clark storage coefficient by
multiplying the storage coefficient by a factor of 1.1 for all subbasins
i. Went back to Deficit and Constant loss method and increased constant loss to 0.075 in/hr.
Download the final project files here -Calibration - 2nd event - Solution.zip103
10.4.3 Questions
1. Which parameters had the largest impact to the calibration?
The initial baseflow and initial deficit appear to be very sensitive parameter. As mentioned, this shows the
variability that exists with your antecedent conditions and how it can affect watershed response.
2. How do your April 1994 calibrated parameters compare to the Sept 2018? Is one better than the other?
Event Subbasin Deficit and Constant Clark UH Linear
Reservoir
As noted in the first question, the largest changes were seen in the initial baseflow and initial deficit. The
April 1994 event hydrograph had a much lower time of concentration showing a quicker runoff response
than the September 2018 hydrograph, likely due to the differences in antecedent conditions. The watershed
basin appears to have been nearly saturated prior to the April 1994 event given the higher baseflow and low
initial deficit (more water in the channel will result in higher velocities and shorter time of concentration
values). If you have confidence in your model boundary conditions, set up, and model results, the model
parameters are of equal importance and best reflect the state of the watershed for that storm event. If you
are not confident in your boundary conditions or calibration results, you may want to consider calibrating to
another storm event.
3. How did your calibration effort go? What NSE values did you receive?
3 https://www.hec.usace.army.mil/confluence/download/attachments/43811373/Calibration%20-%202nd%20event%20-%20Solution.zip?
10
api=v2&modificationDate=1614724534160&version=1
11.1 Introduction
This guide is designed to be the final workshop in the Basic HEC-HMS class. The goal is to develop a simple
hydrologic model and calibrate the model to observed streamflow using gridded precipitation. Lectures and
workshops from the Basic HEC-HMS class touch on components needed to create a hydrologic model, this guide
ties everything together and gives you an opportunity to complete all the basic steps for a single application. This
guide should take you approximately four hours to complete if you are new to HEC-HMS and hydrologic modeling.
This guide is broken up into the following six steps.
1. Download Streamflow Data
2. Delineate Modeling Elements Using HEC-HMS GIS Tools
3. Parameterize the Subbasin Element
4. Add Streamflow and Precipitation Data to the Project
5. Calibrate the Model
6. Document Final Model Parameters
11.2 Background
The project watershed is located upstream of the USGS stream gage(11132500) at Salsipuedes Creek near
Lompoc, CA stream gage (11132500)106. You will create an HEC-HMS model to simulate a flood event during
February 2017. Multi-Radar Multi-Sensor (MRMS)107 gridded precipitation has already been gathered for you. HEC-
Vortex108 was used to convert the MRMS data to HEC-DSS format records. The gridded precipitation data is in the
MRMS_Precipitation.dss file in the Precipitation folder.
The GIS datasets needed to complete the HEC-HMS project were gathered, processed, and are available in the GIS
folder. If you have time during the project, try to gather the GIS data yourself. Familiarize yourself with the GIS data
available for the project. There is an ArcMap project in the GIS directory. There are shapefiles that include the
stream gage location, the NHD streams and watershed, and a buffered watershed (buffer length was 1 mile). The
terrain dataset is included as well. Once downloaded from the USGS National Map Viewer109, the terrain data was
projected, the vertical units converted to feet, and the extents were clipped to the buffered watershed extents. All
GIS data were projected to the standard USACE MMC projection which is summarized below:
Projection: Albers
False_Easting: 0.0
False_Northing: 0.0
Central_Meridian: -96.0
104 https://www.hec.usace.army.mil/confluence/download/attachments/42177015/GIS.zip?api=v2&modificationDate=1607610800050&version=1
105 https://www.hec.usace.army.mil/confluence/download/attachments/42177015/Precipitation.zip?
api=v2&modificationDate=1607610838204&version=1
106 https://waterdata.usgs.gov/nwis/uv?site_no=11132500
107 https://www.nssl.noaa.gov/projects/mrms/
108 https://github.com/HydrologicEngineeringCenter/Vortex
109 https://viewer.nationalmap.gov/basic/#/
Standard_Parallel_1: 29.5
Standard_Parallel_2: 45.5
Latitude_Of_Origin: 23.0
Linear Unit: Foot_US (0.3048006096012192)
Geographic Coordinate System: GCS_North_American_1983
Angular Unit: Degree (0.0174532925199433)
Prime Meridian: Greenwich (0.0)
Datum: D_North_American_1983
Spheroid: GRS_1980
Semimajor Axis: 6378137.0
Semiminor Axis: 6356752.314140356
Inverse Flattening: 298.257222101
The Salsipuedes Creek watershed is located approximately 50 miles west of Santa Barbara, CA, as shown in the
following figures.
If Interested: Create the stream gage layer, download and process the terrain data from the USGS, and
download and process the NHD110 stream and watershed data from The National Map Viewer111. You will need
GIS software to do this bonus work.
110 https://www.usgs.gov/core-science-systems/ngp/national-hydrography
111 https://viewer.nationalmap.gov/basic/#/
5. Enter the USGS Station ID for the specific station of interest within the "USGS Station ID's" column,
"11132500"
6. Enter the River Name within the "Basin Name (A Part)" column, "Salsipuedes Creek"
7. Enter the Gage Name within the "Location (B Part)" column, "Lompoc"
8. Enter "USGS" within the "Other Qualifier (F Part)" column
9. Click the checkbox within the Import Data column
a. Once the data has been imported, change the Data Type to "Annual Peak Data"
b. Click the Import to DSS button to import the annual maximum series
12. Import Instantaneous Data
a. Once the data has been imported, change the Data Type to "Instantaneous (15-min, hourly)"
b. Enter a start and end date/time that encompasses the flood event in question, 14Feb2017:0000 to
25Feb2017:0000
c. Click the Import to DSS button to import instantaneous data
d. Close the USGS Download editor
13. Convert Irregular Data to Regular Data
a. To be used if the E Part uses a tilde character, as shown below.
d. The data should resemble the figure below, note that the default line styles were changed within this
figure
5. Select the terrain data within the basin model's Component Editor.
7. Compute flow directions and flow accumulations using the GIS | Preprocess Drainage tool
8. Compute stream locations using the GIS | Identify Streams tool
a. Use a stream threshold that is slightly less than the reported drainage area of the gage of interest, 45
sq mi for Salsipuedes Creek (reported drainage area = 47 sq mi)
b. This will minimize the amount of subbasin splitting/merging that will need to be performed later
9. Add a breakpoint at the stream gage location using the Breakpoint tool:
a. Zoom into the stream gage location and make sure you add the break point on top of a cell within the
identified streams layer (you might need to modify the draw properties to better see the identified
streams grid cells)
b. Name the breakpoint "Outlet"
10. Delineate Elements using the GIS | Delineate Elements tool
a. Use a Subbasin Prefix of "S_"
b. Use a Reach Prefix of "R_"
c. Set Insert Junctions to Yes
d. Use a Junction Prefix of "J_"
e. Set Convert Break Points to Yes
11. Click View | Map Layers to open the Map Layers Dialog
a. Turn off all layers except Icons, Subbasins, Discretization, Terrain, and the three shapefiles
12. Rename the subbasin element
a. An appropriate subbasin names is "SalsipuedesCk_S10"
13. Rename the sink element to "Outlet"
14. Ensure the subbasin drainage area is approximately equivalent to the reported drainage area
15. Set subbasin modeling processes
a. Structured discretization method
b. Deficit and Constant loss method
c. Clark Unit Hydrograph transform method
d. Linear Reservoir baseflow method
16. Compute Grid Cells for the subbasin element(s)
a. Ensure the Projection is set to SHG and the Cell Size is set to 2000 on the Discretization tab
b. Click GIS | Compute | Grid Cells
c. The grid cells (i.e. discretization) should be shown in the map panel
2. Use the parameter calculator to estimate the Clark time of concentration using Tc = 2.2 \[(L * Lca) / sqrt(S)
\]^0.3
Transform Parameters
Baseflow Parameters
GW1 Steps 1
GW2 Steps 1
iii. The Math Functions editor should resemble the figure below
iv. Save the new time series (Save As option) using an F-part of "USGS_GMT"
3. Within HEC-HMS, create a new Time Series Discharge Gage
a. Click Components | Time Series Data Manager
b. Select Discharge Gage
c. Click New
d. Name the new gage after the location of interest, "SalsipuedesCk_nr_Lompoc"
e. Select the newly created Time Series | Discharge Gage
f. Set the gage to Single Record HEC-DSS
g. Select the DSS file and Pathname that corresponds to the observed streamflow data in GMT, the
time series that uses an F-part of "USGS_GMT"
4. Set observed data within the basin model
a. Select the junction element that represents the gage location
b. On the Options tab:
i. Enter the annual peak discharge that was downloaded from the USGS within the Ref Flow
entry field
ii. Enter an appropriate text string within the Ref Label entry field
iii. Select the newly created time series discharge gage within the Observed Flow drop down
menu
6. Modify Baseflow parameters (GW 2 Initial) to match the observed discharge at the beginning of the
simulation
7. Iteratively modify Loss, Transform, and Baseflow parameters to achieve a Satisfactory, Good, or Very
Good (preferred) NSE, RSR (RMSE Std Dev), and Percent Bias rating
a. Modify Loss parameters (Initial Deficit and Constant Rate) and Baseflow parameters (GW 1
Fraction and GW 2 Fraction) to approximately match the initiation of runoff and observed discharge
volume
b. Modify Transform parameters (Time of Concentration and Storage Coefficient) to approximately
match the peak discharge, time of peak discharge, and overall shape of the observed discharge
hydrograph
Performance Rating NSE RSR PBIAS
Transform Parameters
Baseflow Parameters
GW1 Steps 1
GW2 Steps 1
2. Answer the following questions and share your thoughts with others
a. Describe the process you followed to calibrate the model.
b. What was the most challenging part of model development and calibration?
c. How much confidence do you have in the initial parameter estimates?
112 https://www.hec.usace.army.mil/confluence/download/attachments/42177015/Salsipuedes_Creek.zip?
api=v2&modificationDate=1607611564756&version=1
This tutorial is designed to help new users of HEC-HMS learn how to georeference existing elements in an HEC-HMS
Basin Model using spatial information in subbasin and reach shapefiles.
12.1.1 Overview
In this tutorial, you will georeference existing elements in an HEC-HMS Basin Model with subbasin and reach
shapefiles.
12.1.2 Background
The Punxsutawney Watershed (400 km2) is part of the Allegheny River Basin located in western Pennsylvania, USA.
Primary conveyance streams include: Stump Creek, East Branch Mahoning Creek, and Mahoning Creek. The
confluence of Stump Creek and East Branch Mahoning Creek is located east of the enclave of Big Run. Mahoning
Creek is downstream of the confluence. A map of the watershed is shown in Figure 1.
1. Double click the HEC-HMS icon to start the program. The main program window will appear.
2. To open an existing project, click File and select Open….
a. The Open an Existing Project window will open; click Browse to open the Select Project File
window.
b. Navigate to the project directory and select the "Georeference.hms" file.
c. Click the Select button and the existing project will open. You will notice the basin model is not
georeferenced because the elements are displayed as icons.
d. The subbasin delineation is shown in Figure 3. Subbasin and reach shapefiles have been created that
contains an attribute with names that match the element names in the HEC-HMS basin model.
3. Go to the GIS menu and choose the Georeference Existing Elements menu option.
a. Choose the Subbasins element type and click the Next button.
b. Navigate to the … \Georeference\maps directory and select the Subbasins.shp file. Click Select.
Click Next.
c. Choose the "Name" attribute field to set the subbasin element names.
d. Click the Finish button.
e. On the Basin Model Coordinate System dialogue box, Click Select….
f. Click Browse, Select the Subbasins.prj file. Click Select Click Set. The coordinate system
information should populate as shown in Figure 4.
4. Follow the previous steps and import the reach feature in the Reach.shp file. Figure 5 shows the three
subbasin elements and the one reach element that were imported from the two shapefiles.
5. Right click on the MahoningatPunx basin model name in the Watershed Explorer and choose the Resort
Elements Hydrologically menu option. The elements in the Watershed Explorer should be ordered from
6. Go to the View menu and choose the Map Layers… menu option. Uncheck the Icons. The elements in the
Basin Model are shown without icons so that the users only interact with polygons and lines, as shown in
Figure 6.
This tutorial is designed to help new users of HEC-HMS learn how to create an HEC-HMS Basin Model by importing
existing shapefiles.
12.2.1 Overview
In this tutorial, you will create a basin model from existing shapefiles.
12.2.2 Background
The Punxsutawney Watershed (400 km2) is part of the Allegheny River Basin located in western Pennsylvania, USA.
Primary conveyance streams include: Stump Creek, East Branch Mahoning Creek, and Mahoning Creek. The
confluence of Stump Creek and East Branch Mahoning Creek is located east of the enclave of Big Run. Mahoning
Creek is downstream of the confluence. A map of the watershed is shown in Figure 1.
to define HEC-HMS subbasin and reach elements. The subbasin delineation is shown in Figure 2.
5. Import GIS element by selecting the GIS Menu and choose Import Georeferenced Elements menu options.
a. Choose the Subbasins element type and click the Next button.
b. Navigate to the … \HMS_Example\maps directory and select the Subbasins.shp file. Click Select.
Click Next.
c. Choose the "Name" attribute field to set the subbasin element names. Click the Finish button.
d. You will see three subbasin elements in the basin model map. You can interact with the elements by
clicking on the subbasin polygons or element icons.
6. Follow the previous steps and import the reach feature in the Reach.shp file. Figure 4 shows the three
subbasin elements and the one reach element that were imported from the two shapefiles. These elements
are not connected to one another.
7. Add a junction element to the basin model map. Click on the Junction Creation Tool toolbar button
and then click in the map around the upstream end of the Mahoning Creek reach element. Name the
junction "Big Run". Click on Create.
8. Connect both Stump Creek and the EB Mahoning Creek subbasin elements to the Big Run junction
element.
a. Then, connect the Big Run junction element to the Mahoning Creek reach element.
b. Elements can be connected by right clicking on the element icon in the map, selecting the Connect
Downstream menu option, and then clicking on the downstream element.
A simple connection should be shown in the map when elements are connected to one another. The
exception is when elements are connected to a reach created by GIS features.
You will not see the simple connection from the Big Run junction to the Mahoning Creek reach
element, but the Big Junction Component Editor will show the connection.
9. Click on the Sink Creation Tool toolbar button and add a sink element and name it Punxsutawney.
a. Connect the Mahoning Creek Local subbasin element and the Mahoning Creek reach element to the
new sink element.
b. Now, the elements should be linked to one another in the basin model map.
c. Right click on the MahoningatPunx basin model name in the Watershed Explorer and choose the
Resort Elements Hydrologically menu option.
d. The elements in the Watershed Explorer should be ordered from upstream to downstream, as
shown in Figure 5.
Sink and Drainage Preprocessing creates the necessary layers to start the process, and have no user inputs. Identify
Streams depends on having the preprocessing complete, and requires a user input. Break points are created with a
new tool in the components toolbar. The break points manager is placed where it is so that the user is clued in that
they need to create a break point at that step. The manager is used to allow to the user to see all of the break points
they have created in a basin model.
The final step in creating a simple delineation is the Delineate Elements step, which will create a delineation as long
as there is at least one break point, defining the watershed outlet.
Split and Merge Elements are available once a delineation is created, and allow the user to make their delineation
finer or coarser by dividing up or combining sub-basin and reach elements.
Mouse controls in the map window have been updated so that the behavior of the regular mouse cursor performs
select, zoom and pan without needing to switch tools. Selection is performed with single clicks on map objects.
Zoom is performed using the mouse scroll wheel. Panning is performed by holding down the mouse button and
dragging. When the Break Point Creation Tool is selected, panning is disabled.
to a standard projection before importing it into a project. Typical projections include state plan, UTM, and Albers.
HEC-HMS does not contain tools for re-projecting GIS data sets.
3. Save. You will be prompted to set the basin model coordinate system. Press Select...
a. You can choose a predefined coordinate system by selecting the Predefined button. Here you can set
the basin model coordinate system to Standard Hydrologic Grid (SHG) or Universal Transverse
Mercator (UTM). This is the preferred option.
b. Alternatively, you can use the projection information from a data file by selecting it using the Browse
button.
i. If you browse to a file that has projection information (such as a TIF with an associated TFW,
or a file that uses a PRJ file, etc.) and select it, HMS will produce the WKT projection
information for it.
c. Finally, you can copy-and-paste a spatial projection WKT into the text field.
4. Press Set. The terrain data will display in the basin model map window.
If you press Skip instead of Select..., the basin model coordinate system will be set automatically to the same
projection as the terrain data set. Another option is to set the basin model coordinate system before selecting the
terrain data set. Select the GIS > Coordinate System menu options to open the Coordinate System dialog and then
use the available options to define the coordinate system. After the coordinate system has been set, select the
basin model's terrain data set. HEC-HMS will re-project the terrain data set to the same coordinate system as the
basin model during the import process.
i. If the layer has a meaningful numeric value (such as the Terrain), when the mouse hovers over
that layer in the Basin Model map window, the value underneath the cursor will be displayed
as a tool tip.
c. Select the Draw Properties… button to modify how a layer is displayed.
4. Select the icon in the menu ribbon that is a crosshair with a red dot in the center, . This will allow you to
place break points for delineation.
a. Place at least one break point at the outlet of the watershed, as close as possible to an identified
stream.
i. It is a good idea to zoom in as far as you can on the terrain data to place the break points as
close to the Identified Streams layer as possible. If Identified Streams is selected, the mouse
hover tip will show a number when hovering over the Identified Streams layer that
corresponds to the drainage area at that point, in km2. If no value comes up, you are not
hovering over the Identified Streams layer.
ii. Choose an appropriate name for your break point (e.g. "Outlet".)
iii. Place any additional break points where you desire a break between sub-basins and/or
reaches. It is suggested that break points be added at all locations where a subbasin outlet is
desired, like at stream gage locations. As mentioned, a break point must fall on top of the
identified streams layer in order for a sub-basin to be delineated. The Identify Streams tool
can be re-run and a smaller drainage threshold defined if the stream network is too coarse.
b. After creating at least an outlet break point, choose from the top menu GIS > Delineate Elements. A
prompt that allows you to set some delineation options will appear. These include options for how
HMS will automatically name elements, whether or not to create junctions automatically, or whether
5. The result will be a network of connected sub-basin and reach elements, with a sink element at the outlet of
the watershed.
The Delineate Elements menu option can be re-run at any point; however, any merge or split operations will be
wiped away and a new delineation will be performed. The new delineation will be based on the identified streams
and the break points.
4. Run the Create Grid Cell File tool from the GIS menu. This menu option will only be active if the Preprocess
Drainage tool was successfully completed.
5. When running through the Create a Grid Region wizard, select the file created in step 4.
This tutorial is designed to help new users of HEC-HMS learn how to update changes to existing subbasin and reach
delineation. Instead of creating an new basin model when there are changes to subbasin and stream delineation for
a particular area, users can import the changes to the delineation with updated subbasin and reach shapefiles.
12.4.1 Overview
In this tutorial, you will update an existing subbasin and reach delineation with more detailed subbasin and reach
delineation contained in shapefiles.
12.4.2 Background
The Punxsutawney watershed (400 km2) is part of the Allegheny River Basin located in western Pennsylvania, USA.
Primary conveyance streams include: Stump Creek, East Branch Mahoning Creek, and Mahoning Creek. The
confluence of Stump Creek and East Branch Mahoning Creek is located east of the enclave of Big Run. Mahoning
Creek is downstream of the confluence.
1. Double-click the HEC-HMS icon to start the program. The main program window will appear.
2. To open an existing project, click File and select Open….
a. The Open an Existing Project window will open; click Browse to open the Select Project File
window.
b. Navigate to the project directory and select the "Subbasin_Update.hms" file.
c. Click the Select button and the existing project will open as shown in Figure 3.
3. To update changes in subbasin and reach delineation for the Stump Creek subbasin, you should first delete
the existing Stump Creek subbasin element.
a. Using the Arrow Tool toolbar button, , select the Stump Creek subbasin from the Watershed
Explorer or basin map as shown in Figure 4.
b. Right-click on the Stump Creek subbasin and select Delete Element. Click OK.
The basin model with deleted Stump Creek subbasin is shown in Figure 5.
A separate HEC-HMS project was used to refine, or re-delineate the Stump Creek watershed. In the
separate project, a terrain model was added and tools on the GIS menu were used to delineation
additional subbasin elements. Figure 6 shows the Stump Creek watershed delineated as one
subbasin.
The updated subbasin and stream delineation of the Stump Creek watershed is shown in Figure 7.
There is a menu option on the GIS menu to export subbasin and reach elements as shapefiles. The
refined subbasin features shown in Figure 7 were saved in the Stump-UpdatedSubbasins.shp
file. Updated reach features were saved in the Stump-UpdatedReach.shp file.
4. This step illustrates how to import GIS elements into a basin model with existing elements. Go to the GIS
menu and choose the Import Georeference Elements menu option.
a. Choose the Subbasins element type and click the Next button.
b. Navigate to the … \Subbasin_Update \maps directory and select the updated subbasin shapefile,
the Stump-UpdatedSubbasins.shp file. Click Select. Click Next.
c. Choose the "Name" attribute field to set the subbasin element names.
d. Click the Finish button.
5. Follow the previous steps and import the updated reach features in the Stump-
UpdatedReach.shp shapefile. Figure 8 shows three subbasins (Stump10, Stump20, and Stump30) and two
reach (R-StumpCreek10 and R-StumpCreek20) elements were imported into the basin model. These new
6. Add two junction elements to the basin model map. Click on the Junction Creation Tool toolbar button,
, and then click in the map near the R-StumpCreek10 reach element.
a. Name the junction "J-Stump10". Click on Create.
b. Name the junction "J-Stump20". Click on Create.
7. Elements can be connected by right clicking on the element icon in the map, selecting the Connect
Downstream menu option, and then clicking on the downstream element.
a. Connect subbasin Stump10 to junction J-Stump10.
b. Connect junction J-Stump10 to the upstream end of reach R-StumpCreek10.
c. Connect subbasin Stump20 and the downstream end of reach R-StumpCreek10 to junction J-
Stump20.
d. Connect junction J-Stump20 to the upstream end of reach R-StumpCreek20.
e. Connect subbasin Stump30 and the downstream end of reach R-StumpCreek20 to junction Big
Run.
The connected elements for the Stump Creek watershed are shown in Figure 9.
The entire basin model with updated elements for the Stump Creek watershed is shown in Figure 10.
12.5.1 Introduction
HEC-HMS 4.7 Beta was used to create this tutorial. The purpose of this tutorial is to show how the new parameter
expression calculator can be used to estimate model parameters using either physical characteristics of subbasins
or readily available GIS datasets. The new parameter expression calculator option is available from a few of the
global parameter editors, including the Clark and S-Graph transform editors and the Deficit and Constant and
Green and Ampt loss editors. The parameter expression calculator will be added to other global editors for future
software releases. The parameter expression calculator will compute subbasin average parameter values from
raster datasets (parameter grids). The program uses the zonal statics GIS function to compute subbasin average
parameter values when raster datasets are used.
The following figure shows the Feature to Raster tool in ArcGIS. This tool will create a new raster dataset using
values in the select field (attribute). In the example shown, the saturated content attribute was selected and a
raster with a cell size of 200 meters will be created. The output cell size does not have to match the cell size used
when discretizing HEC-HMS subbasins.
HEC-HMS accepts both ASCII and GeoTIFF file formats for parameter grids. The ArcGIS Raster to ASCII tool, see
below, can be used to convert a raster from ESRI Grid format to ASCII format. Once the ASCII file is created, copy it
(along with the projection file) to the HEC-HMS project directory. HEC-HMS will use relative pathnames for external
files when they are located within the project directory (which makes it easy to transfer the project from one
computer to another).
The following figures show the landuse and wetting front suction GIS datasets for the Coyote Creek watershed. The
percent impervious area was estimated for each landuse type using the Reclassify tool in ArcGIS. Since the
watershed is mostly shrub and mixed forest land, the percent impervious area is very small throughout the
watershed. The SSURGO dataset was used to create raster layers for the saturated hydraulic conductivity and
wetting front suction. The moisture deficit was estimate as a fraction of the soil's porosity. The moisture deficit
reflects the soil moisture at the beginning of a model simulation.
++
As shown below, HEC-HMS 4.7 has an option to choose an ASCII or GeoTIFF Data Source for parameter grids, in
addition to the HEC-DSS format. In order to use the parameter grids in the parameter expression calculator, the
parameter grids must be added to the project in either the ASCII or GeoTIFF formats. For this tutorial, a percolation
rate, impervious area, water content, and water potential grids have already been added to the project in the ASCII
format. The ASCII files are located in the ...\ExampleProject\gis\ParameterGrids\Albersft directory. A copy of the
ESRI grids used to create these ASCII files are located in the ...\ExampleProject\gis\ArcDatasets directory.
The Clark time of concentration and storage coefficient parameters can be estimated using the following equations
(the coefficient and exponent should be determined/calibrated regionally). The equation for the time of
concentration relates the longest flow path, centroidal flow path, and 10-85 flow path slope to the time of
concentration. HEC-HMS 4.7 will compute these physical characteristics for subbasin elements. In order for the
program to compute subbasin characteristics, the subbasin elements must be delineated using the GIS delineation
tools in HEC-HMS. Subbasin characteristics can still be computed for those existing projects that were completed
before the GIS delineation tools in HEC-HMS were made available. The subbasin elements must be georeferenced
using a shapefile and then terrain data must be added to the basin model and the GIS→Preprocess Drainage and
GIS→Identify Streams steps completed.
The following figure shows the subbasin characteristics for the Coyote Creek watershed. You can see this
information from the Parameters→Characteristics→Subbasin menu option.
2. Select the Parameters→Transform→Clark menu option to open the Clark Unit Hydrograph global editor
as shown below.
3. Click the Calculator... button at the bottom of the global editor. The following figure shows the Expression
Calculator. Select the Time of Concentration in the upper left. Then make sure the Stats tab is selected in
the Variables panel. The stats tab is linked to the subbasin characteristics. The equation panel shows the
equation entered to estimate the time of concentration. After recreating this equation, click the Calculate
button and the the program will add the computed values to the global editor.
4. Assuming an R/(R+TC) relationship of 0.5 means R is equal to TC. Within the global editor, copy values in the
time of concentration column and paste them into the storage coefficient column. Click the Apply and Close
12.5.4 Using the Parameter Estimation Calculator to Estimate Green and Ampt
Parameters
1. Select the Parameters→Loss→Green and Ampt menu option to open the Green and Ampt global editor as
shown below. All parameters except for the wetting front suction have already been estimated.
2. Click the Calculator... button at the bottom of the global editor. The following figure shows the Expression
Calculator. Select the Suction parameter in the upper left. Select the Grids tab in the Variables panel. The
grids tab in linked to all parameter grids that have been added to the project (all grids in ASCII and GeoTIFF
format). As shown below, the equation panel shows the equation entered to estimate the subbasin average
wetting front suction. The water potential grid is referenced; therefore, the program will use zonal statistics
to compute the average value for each subbasin in the basin model when the Calculate button is pressed.
3. The following figure shows the estimated Green and Ampt parameters. Click the Apply and Close buttons.
Select the 1982_subbasins simulation run and compute the simulation. View the results at the FS77 sink element
(this is the most downstream point in the model). Notice the computed hydrograph matches the observed
hydrograph very well for this event. The other model parameters were modified (initial parameter estimates had
been modified to calibrate the model). The saturated hydraulic conductivity was reduced from initial estimates, the
R/(R+TC) ratio was modified, and the baseflow parameters were determined through trial and error.
This tutorial was updated on Sept 25, 2020 using HEC-HMS 4.7 Beta. The structured discretization method can be
used instead of the *.mod file that was created by HEC-GeoHMS and HEC-HMS versions 4.4 - 4.6.1.
For creating a gridded boundary condition see Creating Gridded Boundary Conditions for HEC-HMS(see page 600).
There are existing HEC-HMS models that can be updated and improved to utilize gridded precipitation and
hydrologic methods. Many of these existing models are not georeferenced with coordinate system or projection
information. These models often do not have background maps of the subbasins and reaches. When background
maps of the subbasins and reaches are available the coordinate system information is often missing. These older
models often use the lump approach for representing subbasins and hydrologic methods rather than the more
useful gridded models, which allows the watershed to be represented in greater detail spatially.
In order to use gridded precipitation in the meteorological model and hydrologic methods in the basin model, the
user must first georeference the basin model. When converting a lumped basin into a gridded basin, the user needs
to process the terrain data for flow directions, and then generate either a structured or unstructured grid.
The structured or unstructured grid specifies which grid cells are in each subbasin and the properties of each cell
including location, area within the subbasin, distance to the subbasin outlet, and contain geospatial information
about the grid.
This tutorial is designed to help the user of HEC-HMS learn how to apply a gridded precipitation dataset to a non-
georeferenced project.
13.1 Overview
In this tutorial, you will perform three tasks to georeference existing subbasin and reach elements before
generating a structured grid: 1) georeference the project, 2) process terrain information to create a flow direction
grid, and 3) create a structured (SHG) grid for each subbasin.
13.2 Background
The Punxsutawney Watershed (400 km2) is part of the Allegheny River Basin located in western Pennsylvania, USA.
Primary conveyance streams include: Stump Creek; East Branch Mahoning Creek; and Mahoning Creek. The
confluence of Stump Creek and East Branch Mahoning Creek is located east of the enclave of Big Run. Mahoning
Creek is downstream of the confluence.
A map of the watershed is shown below.
1. Double click the HEC-HMS icon to start the program. The main program window will appear.
2. To open an existing project, click File and select Open….
a. The Open an Existing Project window will open; click Browse to open the Select Project File
window.
b. Navigate to the project directory and select the Gridded Basin.hms file.
c. Click the Select button and the existing project will open as shown below. Notice the basin model is
not georeferenced because the elements are displayed as icons.
d. The subbasin delineation is shown below. Subbasin and reach shapefiles have been created that
contains an attribute with names that match the element names in the HEC-HMS basin model.
3. Go to the GIS menu and choose the Georeference Existing Elements menu option.
a. Choose the Subbasins element type and click the Next button.
b. Navigate to the … \Gridded_Basin\maps directory and select the Basin Punx.shp file. Click Select.
Click Next.
c. Choose the name attribute field to set the subbasin element names. Click the Finish button.
4. Follow the previous steps and import the reach feature in the Reach Punx.shp file. The figure below shows
the three subbasin elements and the one reach element that were imported from the two shapefiles.
5. Right click on the Gridded Punx basin model name in the Watershed Explorer and choose the Resort
Elements Hydrologically menu option. The elements in the Watershed Explorer should be ordered from
upstream to downstream, as shown below.
13.4 Task 2: Load Terrain Dataset and Run the Preprocessing Steps
1. Go to the Components menu and choose the Create Component menu option as shown below.
a. Select Terrain Data, enter a Name of "Terrain Punx". Click Next as shown below.
c. After the terrain data has been added to the project, it can be associated to a basin model. Open the
Gridded Punx basin model's Component Editor and then select Terrain Punx for the Terrain Data as
shown in the two figures below. The terrain's coordinate system is read automatically from the
Terrain Punx.tfw file.
2. The purpose of the next few steps is to perform drainage analysis on the terrain and produce a flow direction
grid that will be needed later to create the grid cell file. Terrain data is a gridded representation of the
watershed surface with elevation for each grid cell. This Preprocess Sinks step fills the sinks or depressions
in the terrain so that the drainage analysis is possible. Go to the GIS Menu and choose the Preprocess Sinks
menu option as shown below. The result of this step is a terrain without depressions, and a Sink Fill raster,
as shown below.
3. Using the Sink Fill data from the previous step, the Preprocess Drainage step performs drainage analysis
and produces the flow direction and flow accumulation data. Go to the GIS Menu and choose
4. This optional step shows how to turn off the visibility of map layers to speed up the map display. Go to the
View Menu and choose Map Layers in the menu option. Uncheck the visibility for the Flow Accumulation,
Flow Direction, Sink Locations, and Sink Fill layers as shown below.
discretization method replaces the grid cell approach (*.mod file) that was available in previous versions of
HEC-HMS. The structured discretization method lets the modeler choose between a grid in either the
Standard Hydrologic Grid (SHG) or UTM grid projections, and the modeler can choose the grid cell size. Once
the modeler chooses the grid type, SHG or UTM, and then the grid size, HEC-HMS automatically creates the
grid and intersects the grid with the subbasin boundary. The figure below show the Structured discretization
method selected for the EB Mahoning Creek subbasin. Choose the Structured discretization method for
all three subbasin elements.
2. Go to the Discretization tab in the Subbasin's Component Editor and make sure the SHG Projection is
selected and that the Grid size is set to 2000 meters, as shown below. Do this for all three subbasins.
3. Go to the GIS menu and select Compute→ Grid Cells as shown below.
4. After the grid cells have been computed, open the Map Layers window and check the box next to
Discretization. The 2000 meter SHG should be displayed in the basin model map as shown below.
The spatial information is saved in an *.sqlite file in the project directory. The *.sqlite file will be named using the
basin model name. You can visualize the tabular data in this file using a tool like DB Browser. The geospatial
information can be displayed in QGIS. Now that a structured grid has been developed for this basin model, gridded
precipitation can be applied.
There are existing HEC-HMS models that can be updated and improved to utilize gridded precipitation and
hydrologic methods. Many of these existing models are not georeferenced with coordinate system or projection
information. These models often do not have background maps of the subbasins and reaches. When background
maps of the subbasins and reaches are available the coordinate system information is often missing. These older
models often use the lump approach for representing subbasins and hydrologic methods rather than the more
useful gridded models, which allows the watershed to be represented in greater detail spatially.
In order to use gridded precipitation in the meteorological model and hydrologic methods in the basin model, the
user must first georeference the basin model. When converting a lumped basin into a gridded basin, the user needs
to process the terrain data for flow directions, and then generate a grid cell file. The grid cell file specifies which grid
cells are in each subbasin and the properties of each cell including location, area within the subbasin, and distance
to the subbasin outlet. The user can then refer to the grid cell file and create a grid region as a model component.
The user then must specify the default grid region to be used in the basin model.
This tutorial is designed to help the user of HEC-HMS learn how to apply a gridded precipitation dataset to a non-
georeferenced project.
14.1 Overview
In this tutorial, you will perform five tasks to georeference existing subbasin and reach elements before generating
a grid cell file and creating a grid region: 1) georeference the project, 2) process terrain information to create a flow
direction grid, 3) generate grid cell file; 4) create grid region; and 5) create links in an HEC-HMS basin model.
14.2 Background
The Punxsutawney Watershed (400 km2) is part of the Allegheny River Basin located in western Pennsylvania, USA.
Primary conveyance streams include: Stump Creek; East Branch Mahoning Creek; and Mahoning Creek. The
confluence of Stump Creek and East Branch Mahoning Creek is located east of the enclave of Big Run. Mahoning
c. Click the Select button and the existing project will open as shown below. Notice the basin model is
not georeferenced because the elements are displayed as icons.
d. The subbasin delineation is shown below. Subbasin and reach shapefiles have been created that
contains an attribute with names that match the element names in the HEC-HMS basin model.
3. Go to the GIS menu and choose the Georeference Existing Elements menu option.
a. Choose the Subbasins element type and click the Next button.
b. Navigate to the … \Gridded_Basin\maps directory and select the Basin Punx.shp file. Click Select.
Click Next.
c. Choose the Name attribute field to set the subbasin element names. Click the Finish button.
d. On the Coordinate System dialogue box, Click Select….
e. Click Browse, Select the Basin Punx.prj file. Click Select Click Set. The coordinate system
information should populate as shown below.
4. Follow the previous steps and import the reach feature in the Reach Punx.shp file. The figure below shows
the three subbasin elements and the one reach element that were imported from the two shapefiles.
5. Right click on the Gridded Punx basin model name in the Watershed Explorer and choose the Resort
Elements Hydrologically menu option. The elements in the Watershed Explorer should be ordered from
upstream to downstream, as shown below.
14.4 Task 2: Load Terrain Dataset and Run the Preprocessing Steps
1. Go to the Components menu and choose the Create Component menu option as shown below.
a. Select Terrain Data, enter a Name of "Terrain Punx". Click Next as shown below.
c. After the terrain data has been added to the project, it can be associated to a basin model. Open the
Gridded Punx basin model's Component Editor and then select Terrain Punx for the Terrain Data as
shown in the two figures below. The terrain's coordinate system is read automatically from the
Terrain Punx.tfw file.
2. The purpose of the next few steps is to perform drainage analysis on the terrain and produce a flow direction
grid that will be needed later to create the grid cell file. Terrain data is a gridded representation of the
watershed surface with elevation for each grid cell. This Preprocess Sinks step fills the sinks or depressions
in the terrain so that the drainage analysis is possible. Go to the GIS Menu and choose the Preprocess Sinks
menu option as shown below. The result of this step is a terrain without depressions, and a Sink Fill raster,
as shown in the second figure below.
3. Using the Sink Fill data from the previous step, the Preprocess Drainage step performs drainage analysis
and produces the flow direction and flow accumulation data. Go to the GIS Menu and choose
4. This optional step shows how to turn off the visibility of map layers to speed up the map display. Go to the
View Menu and choose Map Layers in the menu option. Uncheck the visibility for the Flow Accumulation,
Flow Direction, Sink Locations, and Sink Fill layers as shown below.
data as an input to create the grid cell file. Go to the GIS Menu and choose Create Grid Cell File in the menu
option as shown below.
a. On the Create a Grid Cell File dialogue box, select a Projection of SHG and Cell Size (M) of 2000
(meters). Click Next as shown below.
b. Select a file location and name. In this example, a grid cell file will be created in the directory ...
\Gridded_Basin\regions and it will be named Gridded_Punx.mod. Click Finish as shown below.
c. Navigate to the \Gridded_Basin\regions\... directory and open the grid cell file
Gridded_Punx.mod with a text editor as shown below. On the first line of the grid cell file, the
Parameter Order will show the order of the data for each grid cell in a subbasin. The parameter
Xcoord and Ycoord are the X and Y coordinate of the cell in the SHG grid coordinate system. The
parameter TravelLength (km) is the average travel length distance from the grid cell to the subbasin
outlet. The parameter Area (sq.km.) is the area of the cell within the subbasin. An area of 4.0
sq.km113. is computed when the whole cell is within the subbasin. An area less than 4.0 sq.km114. is
computed when the cell is on the subbasin boundary.
113 http://sq.km
114 http://sq.km
d. HEC-Vortex has many utilities and can be downloaded from link https://github.com/
HydrologicEngineeringCenter/Vortex/releases. The utility named travel-length-grid-cells-exporter
is used to convert the grid cell text file into a shapefile showing the grid cells. By providing the grid
cell text file (Gridded_Punx.mod), projection file (Basin Punx.prj), and cell size (2000), as input into
the utility, a grid cell shapefile (…\Gridded_Basin\maps\Grid_Cell.shp), is produced as an output.
The grid cell shapefile is added as a map overlay showing the gridded representation of the subbasin
as shown below.
a. On the Create a Grid Region dialogue box, enter a Name of Grid Region PunxSHG. Click Next as
shown below.
b. On the Create a Grid Region dialogue box, navigate to the ...\Gridded_Basin\regions directory and
select the Gridded_Punx.mod file. Click Next as shown below. The Grid Region will be added to the
Watershed Explorer as shown in the second figure.
In case of changes to the subbasin name or subbasin delineation, the user must rerun the grid cell file step and
override the previously created grid cell text file. Also, if the gridded precipitation is in a different grid
projection, or the cell size does not match the cell size in the grid cell file, then a new grid cell file would need
to be created, or the gridded precipitation data would need to be modified.
15.1 Overview
In this tutorial, you will perform two tasks to develop a portable HEC-HMS project: 1) Setup a Portable HEC-HMS
Model and 2) Save as a Portable HEC-HMS Model.
For this task, you will move the external data folder inside the project folder.
1. Open File Explorer and go the HEC-HMS project folder as shown in Figure 2.
2. Move the data folder inside the HEC-HMS project folder, Intro_Workshop, as shown in Figure 3.
4. Go to Time-Series Data and click on the Precipitation Gages folder, select the new location of DSS
Filename as …\Task1\Initial\Intro_Workshop\data\observe.dss for DUJP, MFFP, and PNXP gage as
shown in Figure 4. Even though HEC-HMS shows the complete pathname to the DSS file in the
Component Editor, a relative pathname is used when the program saves the pathname.
For this task, you will use the Save As option to create a portable copy of the HEC-HMS project.
7. HEC-HMS will create a new portable HEC-HMS model with a data folder for external data as shown in Figure
7, 8, and 9. A separate DSS file will be created for each time-series gage, as well as each paired data curve
and grid.
myProject = Project.open('C:/Projects/castro/castro.hms')
myProject.computeRun('Current')
myProject.close()
Hms.shutdownEngine()
set "CLASSPATH=C:/Projects/hms/HEC-HMS/build/distributions/HEC-HMS-4.4-beta.2/hms.jar;C:/
Projects/hms/HEC-HMS/build/distributions/HEC-HMS-4.4-beta.2/lib/*"
C:/jython2.7.1/bin/jython -Djava.library.path="C:/Programs/HEC-HMS-4.4-beta.2/bin;C:/
Programs/HEC-HMS-4.4-beta.2/bin/gdal" compute_current.py
Line 1: The path to GDAL binaries prepends the the PATH environment variable. This is probably not necessary
unless you are working with the HMR52 storm in the met model. Paths to TauDEM and MPI are not included because
we will not be accessing these via scripting.
115 https://www.hec.usace.army.mil/software/hec-hms/downloads.aspx
116 https://www.jython.org/download
Lines 2-4: Other GDAL environment variables are configured. These are probably not necessary for most scripting
applications.
Line 5: Set the java classpath to include the path to hms.jar and dependent libraries. All of the dependent libraries
are located in the distribution lib directory. In this case a wildcard * indicates that all libraries in the lib directory
should be added to the classpath.
Line 6: Use the jython executable to execute the script. -Djava.library.path sets the path to native libraries. Paths
should include the distribution bin directory, where javaHeclib.dll and WindowsEnvironment.dll reside, as well as the
bin/gdal directory where gdal201.dll resides.
In Project Structure, add hms.jar as a library. You can optionally add the rest of the dependent library jars here.
In the run configuration, set environment variables. See notes in the Running a script via batch section. In this case
the CLASSPATH environment variable is set to the distribution lib directory. Alternatively you can add each jar
explicitly in Project Structure | Libraries.
In the run configuration, set interpreter options to include -Djava.library.path. See notes in the Running a script via
batch section.
For the HEC-HMS API reference, see Scripting117 in the HEC-HMS User's Manual.
Happy scripting!
117 https://www.hec.usace.army.mil/confluence/display/HMSUM/.Scripting+v4.7
17.1 Overview
In this tutorial, you will compute two simulation runs and compare results from the two meteorologic models. Both
simulation runs use the same basin model and control specifications. The first simulation run uses the Gage
Weights precipitation method and the second simulation run uses the Inverse Distance precipitation
method. Components of both simulation runs are shown below.
Simulation Run name: Specified Wt
Basin Model: Punxsutawney
Meteorologic Model: Specified Gage Wt
Control Specifications: 1 May 1996
5. Click on the Results tab and expand the results for the Punxsutawney subbasin as shown in Figure 1.
6. Click on the subbasin (Punxsutawney) Graph from the Specified Wt simulation run and move the plot to the
side.
7. Click on the subbasin (Punxsutawney) Graph from the Inverse Dist simulation run and move the plot to the
side and under the first plot.
9. Click on the Summary Table of the Specified Wt simulation run and move the table to the side.
10. Click on the Summary Table of the Inverse Dist simulation run and move the table to the side and under
the first table.
11. The summary tables of both simulation runs are shown simultaneously in Figure 3.
12. Click on the Precipitation result for the Puxsutwaney subbasin in the Specified Wt simulation run.
13. Press Control + Click on the Precipitation result for the Punxsutawney subbasin in the Inverse Dist
simulation run.
14. Basin average precipitation hyetographs from the Specified Wt (solid line) and Inverse Dist (dashed
line) simulation runs are shown simultaneously in Figure 4. The differences in the computed basin average
precipitation is due to the different methods for computing basin average precipitation. The gage weights
approach uses user defined gage weights, and the inverse distance method automatically computes the
gage weights using the distance from the gage to a user defined node (subbasin centroid). The example
shows a way to compare time-series results from two separate simulation runs. Additional time-series can
be added to the plot by clicking on the time-series result in the Watershed Explorer and dragging/dropping
18.3 Background
The goal of this tutorial is to become familiar with entering cross section information in the paired data editor. The
Muskingum-Cunge routing method has an option to define the cross section shape using an eight-point cross
section. An eight-point cross section is defined using station and elevation information. If a detailed cross section is
available, the cross section data points must be distilled to eight points that represent the cross section shape.
18.4 Overview
In this workshop you will identify the eight points that define the cross section shape and use the paired data editor
to create a new cross section curve. The cross section can be used in one or many basin models in the HEC-HMS
project. Once the cross section paired data curve is defined, you will select the curve within the Muskingum-Cunge
routing Component Editor.
1. Review the field data shown in Figure 1. A detailed cross section was gathered from a field survey. The
detailed cross section contains more than eight points. Figure 1 also shows eight points were extracted from
the detailed cross section information. The eight points preserve the cross section shape
118 https://www.hec.usace.army.mil/confluence/download/attachments/30802167/Initial_Paired_Data.zip?
api=v2&modificationDate=1594420307547&version=2
119 https://www.hec.usace.army.mil/confluence/download/attachments/30802167/Final_Paired_Data.zip?
api=v2&modificationDate=1594420322855&version=2
2. Open HEC-HMS and select the File menu. Choose the Open… menu option and select the Routing
Workshop project.
3. Open the Paired Data Manager from the Components menu. Select the Cross Sections paired data type
and click the New… button. Enter a Name of “upper” for the paired data curve and click the Create button.
4. Navigate to the new paired data curve in the Watershed Explorer. The paired data curve is located in the
Cross Sections folder within the Paired Data folder. Click on the upper curve to open the Component
Editor.
5. Within the Component Editor, set the Data Source to Manual Entry and the units to FT:FT.
6. Enter the eight station-elevation pairs from Figure 1 into the table provided on the Table tab. Figure 2 shows
the Table and Graph tabs for the new cross section paired data curve.
7. Open the basin model named Musk-Cunge, shown in Figure 3. There are three reaches in the basin model,
named Upper, Middle, and Lower. The reaches are connected to one another with water flowing from the
Inflow Hydrograph source element to reach Upper, reach Middle, reach Lower, and then the Punxsutawney
Local sink element. All Muskingum-Cunge parameters, except for the 8-point cross sections have already
been defined for the three reaches.
8. Click on the Upper reach element to open the Component Editor. Select the Routing tab to see the
Muskingum-Cunge parameters defined for the reach. The cross section shape has been set to Eight Point. As
shown in Figure 4, click the Cross Section drop down list and select the “upper” paired data curve that was
created above.
19.1 Download the Initial project files here -Initial Percentage Curve.zip120
19.2 Download the Final project files here - Final Percentage Curve.zip121
If you are prompted to convert the project to a later version, select "Convert Project" as in the figure below:
This tutorial is designed to show you how to create a storm time pattern, defined as a percentage curve, which
can be used within the Hypothetical Storm precipitation method.
19.3 Overview
In this tutorial, you will perform two tasks to create a time pattern in HEC-HMS and select the time pattern within
the Hypothetical Storm Component Editor:
1) Create a Percentage Curve Paired Data Record.
2) Select the Percentage Curve Time Pattern within the Hypothetical Storm Component Editor.
19.4 Background
The basin model has been created for you by combining the three subbasins (Stump Creek, East Branch Mahoning
Creek, and Punxsutawney Local) as one subbasin labeled Punxsutawney. The study area is 158 square miles and
located in Jefferson County, PA. This tutorial is configured to include a frequency storm precipitation model along
with two hypothetical storm precipitation models. This tutorial will focus on creating the temporal pattern, there is
another tutorial that provides more in-depth description of the hypothetical storm precipitation method and
compares results to the frequency storm precipitation method.
120https://www.hec.usace.army.mil/confluence/download/attachments/30804957/Initial%20Percentage%20Curve.zip?
api=v2&modificationDate=1595788695465&version=1
121https://www.hec.usace.army.mil/confluence/download/attachments/30804957/Final%20Percentage%20Curve.zip?
api=v2&modificationDate=1595788713343&version=1
precipitation were computed as shown in the highlighted columns. The percent time column was computed by
dividing the duration by the total storm duration, the total storm duration was 24 hours in this example. When
defining a temporal pattern percentage curve in HEC-HMS, both independent (percent storm duration) and
dependent (percent total depth) variables must start at 0 percent and end at 100 percent.
The excel spreadsheet (Time Pattern Percent Curve.xls) shows how the incremental precipitation for this tutorial
was converted to percentage curves for application in HEC-HMS.
Date and Time Precipitation Time (%) Normalized Cumulative Precipitation (%)
1. Open the HEC-HMS project named Hypothetical Storm in the Time Pattern folder.
2. Select the Components | Paired Data Manager menu option to open the Paired Data Manager. Choose the
Percentage Curves option in the Data Type drop down list. Click the New… button as shown in Figure 1.
3. Enter a Name of Precip Pattern - June1972 for the new Percentage Curve and click the Create button.
4. Within the Watershed Explorer, navigate to the Paired Data folder. Then expand the Percentage Curves
folder to see the new percentage curve. Click on the Precip Pattern - June1972 node in the Watershed
Explorer to access the Component Editor.
5. Make sure the Date Source is set to Manual Entry as shown in Figure 2.
6. Go to the Table tab and copy/paste the information in the Time Pattern Percent Curve.xls spreadsheet as
shown in Figure 3.
7. Go to the Graph tab to see a plot of the time pattern Percentage Curve, as shown in Figure 4.
19.6 Task 2: Select the Percentage Curve Time Pattern within the
Hypothetical Storm Component Editor
1. Select and expand the Hypothetical Storm 1972P Meteorologic Model in the Watershed Explorer and the
click on the Hypothetical Storm node in the tree, as shown in Figure 5.
2. Click the drop-down list for the Storm Pattern and choose the Precip Pattern – June1972 curve that was
created in Task 1.
20.1 Overview
In this tutorial group, you will focus on filling out the information needed for two precipitation methods used in
hypothetical storm modeling within the Meteorologic Model. If you're interested, the User's Manual and Technical
Reference Manual provide greater details on all available methods.
This tutorial group includes 2 components: 1) How to create a Frequency Storm precipitation model and 2) How to
create a Hypothetical Storm precipitation model.
This tutorial is designed to help new users of HEC-HMS learn how to extract depth-duration information and create
meteorologic model in HEC-HMS with frequency based precipitation.
20.2.1 Overview
In this tutorial, you will perform three tasks while learning to apply the frequency storm precipitation method: 1)
extract depth-duration information from NOAA's ATLAS 14; 2) create a frequency storm in HEC-HMS; and 3) evaluate
results.
20.2.2 Background
The study area for this tutorial is the 158 square mile Punxsutawney watershed in Pennsylvania. A basin model has
been created for you in HEC-HMS. A map of where the watershed is located is shown in Figure 1 and a screenshot of
the basin model is shown in Figure 2.
1. Open the HEC-HMS project named Precip_Frequency_Storm in the Precip Frequency Storm folder.
2. Create a new meteorologic model. Select Components menu and choose the Create Component menu
option. Select Meteorologic Model.
3. In the description box enter Freq Storm 1% and click Create.
4. In the Watershed Explorer, expand the Meteorologic Models folder and select Freq Storm 1%
5. In the Components Editor:
a. Select Meteorology Model tab
b. Select Unit System: US Customary.
c. Change the Precipitation option from Specified Hyetograph to Frequency Storm.
d. Select the Basins tab.
e. Change the Include Subbasins option from No to Yes for the Base 1% basin model.
6. Go back to the Watershed Explorer and click on Frequency Storm under your Freq Storm 1%
meteorologic model.
a. Select the Annual-Partial Conversion option: None.
b. Select the Storm Duration option: 1 Day.
c. Select the Intensity Duration option: 15 Minutes.
d. Select the Intensity Position option: 50 Percent.
e. Select the Area Reduction: TP40
f. Enter Storm Area (MI2): 158.
g. Select the Curve option : Uniform for All Subbasin
7. Enter the precipitation depths straight from ATLAS 14 as shown in Figure 5.
It is important to note that the correct “area-reduced” result under the Peak Discharge (CFS) column
is only applicable to the Punxutawney outlet because a storm area of 158 square miles (which
corresponds to the drainage area at this point) was defined in the meteorologic model. When the
drainage area of 158 square mile is used to compute results at other locations, like the Big Run
junction, it would not result in the correct “area-reduced” 1% flow. To compute the correct 1% flow
at the Big Run junction, a storm area of 70 square miles would need to be defined in the meteorologic
model. For computing “area-reduced” flows at multiple analysis points, HEC-HMS includes a Depth-
Area Analysis simulation where the program runs multiple variations of the frequency storm
meteorologic model with different storm areas.
b. Open summary tables and plots for the Punxsutawney junction element by opening the
Punxsutawney tab and selecting Summary Table as shown in Figure 7.
c. Select Graph and view the plot of the hydrograph for the simulation at Punxsutawney. The plot
should resemble the graph below in Figure 8.
This tutorial is designed to show you how to create and apply the Hypothetical Storm precipitation method within
the HEC-HMS Meteorologic Model.
20.3.1 Overview
In this tutorial, you will perform two tasks while learning to apply the hypothetical storm method: 1) Create a
Meteorologic Model with the Hypothetical Storm Precipitation Method and 2) Run Simulations and Compare
Hypothetical Storm Results from Different Temporal Patterns.
20.3.2 Background
The basin model has been created for you by combining the three subbasins (Stump Creek, East Branch Mahoning
Creek, and Punxsutawney Local) as one subbasin labeled Punxsutawney. The study area is 158 square miles and
located in Jefferson County, PA. This tutorial is configured to include a frequency storm precipitation model along
with two hypothetical storm precipitation models. Simulation runs have already been created as well.
20.3.3 Task 1: Create a Meteorologic Model with the Hypothetical Storm Method
1. Open the HEC-HMS project named Hypothetical Storm in the Initial_Hypothetical_Storm folder.
2. Open the Component Editor for the Hypothetical Storm 1972P meteorologic model. Select the Hypothetical
Storm precipitation method as shown in Figure 1.
3. Expand the Watershed Explorer and click on the Hypothetical Storm node to open the Component Editor.
Enter the following Hypothetical Storm parameters (see Figure 2). Click this link to see another tutorial
showing how to create a user-specified temporal pattern - Creating a Hypothetical Storm Time Pattern
Percentage Curve(see page 430).
HEC-HMS will use the user-specified time pattern (percentage curve), the storm duration, and the storm
depth to create a hyetograph for each subbasin in the basin model. It is important that the storm duration
match the duration of the historical event used to create the percentage curve. If area-reduction is being
used, then the program will reduce the point depth first before creating the subbasin hyetograph.
20.3.4 Task 2: Run Simulations and Compare Hypothetical Storm Results from
Different Temporal Patterns
Three simulations have already been created for you.
• Freq Storm 1% = This run uses the Frequency Storm Precipitation Method (not Hypothetical)
• Hypo Storm 1% Pattern = This run uses the Hypothetical Storm Precipitation Method. The temporal
pattern from the Freq Storm 1% simulation was used for this simulation for illustration only. The temporal
pattern is not based on historic precipitation, it is based on the balanced hyetograph approach.
• Hypo Storm June 1972 Pattern = This run uses the Hypothetical Storm Precipitation Method with the June
1972 temporal pattern.
1. Run all three simulations.
2. Go to the Results tab in the Watershed Explorer and expand the tree for both Hypothetical Storm
simulations. Make sure you expand the tree for the Punxsutawney subbasin element in both simulations as
well.
3. Figure 3 shows the hyetograph and hydrograph plots for both simulations. Notice the simulation that uses
the June 1972 pattern hyetograph has a lower peak precipitation intensity and flow. The idealized
precipitation pattern used in the Hypo Storm 1% Pattern simulation comes from the frequency storm
precipitation method which creates a balanced hyetograph. The balanced hyetograph includes the 1-
percent 1, 2, 3, 6, 12, and 24-hour precipitation depths that are nested within one storm. The temporal
pattern from the June 1972 event has only been scaled so that the 24-hour duration contains the 1-percent
precipitation depth.
4. Click the Summary Table node in the Watershed Explorer to open a window with summary information for
both simulations. Notice the peak discharge from the simulation using the June 1972 pattern is 20,818 cfs
(discharge volume = 4.9 inches) and the peak discharge from the simulation using the idealized temporal
pattern is 25,906 cfs (discharge volume = 4.9 inches). The total precipitation and runoff volume are the same
in both simulations; however, the precipitation intensity is different, which results in different runoff
responses.
This tutorial is designed to help users of HEC-HMS learn how to perform depth-area analysis and automate the
process of developing a separate frequency storm for each analysis point in the basin model. Each depth-area
analysis is based on a basin model and a meteorologic model. The analysis also includes the definition of the
simulation time window, time step, and the selection of analysis points within the basin model. The depth-area
analysis is designed to work with both the frequency storm and hypothetical storm meteorologic models. Each
analysis point is its own separate simulation within the overall depth-area analysis simulation. The drainage area at
each analysis point is extracted and used to update the precipitation depths based on the user selected area-
reduction curve. In the case of the hypothetical storm, both the area-reduction and temporal pattern can be based
on the drainage/storm area.
21.1 Overview
In this tutorial, you will perform two tasks to create and perform depth area analysis: 1) create a depth-area analysis
simulation with a frequency storm and 2) create a depth-area analysis simulation with a hypothetical storm.
21.2 Background
The study area for this tutorial is the 158 square mile Punxsutawney watershed in Pennsylvania. A basin model has
been created for you in HEC-HMS. A map of where the watershed is located is shown in Figure 1 and a screenshot of
the basin model is shown in Figure 2.
The Depth-Area Analysis simulation includes a built in capability where the program runs multiple variations
of the frequency storm meteorologic model; each variation has a different storm area. HEC-HMS allows the
modeler to choose multiple points, analysis points, within the basin model when creating the depth-area
analysis. During the depth-area analysis simulation, the program will extract the drainage area at each of the
selected analysis point locations, interpolate the appropriate area-reduction factor for the analysis point's
drainage area, and create appropriate precipitation hyetographs for the drainage area upstream of each
analysis point. Think of the depth-area analysis simulation as multiple simulations within one. Each analysis
point is its own simulation where the flow at the analysis point is appropriately simulated given the drainage
area, and area-reduction information defined in the frequency storm meteorologic model.
1. Select Compute menu. Select Create Compute and Depth-Area Analysis….as shown in Figure 3.
2. Create a new depth-area analysis named Depth-Area Freq Storm 1%. Click Next.
3. Select basin model Base 1%. Click Next.
4. Select meteorologic model Freq Storm 1%. Click Finish.
5. After the depth-area analysis has been created, go to the Compute tab and select the depth-area analysis to
open the Component Editor.
6. Enter a start date of 01Jan1999, a start time of 01:00, an end date of 11Jan1999, and an end time of 00:00.
Set the Time Interval to 15 minutes as shown in Figure 4.
7. Select the Analysis Points tab. Select the Big Run junction as Point 1 and Punxsutawney junction as Point
2 as shown in Figure 5.
8. Run the depth-area analysis. There are really two simulations that happen during the compute as shown on
Figure 6. For point 1, HEC-HMS will apply area-reduction factors for a storm area of 70 square miles and run
the basin model to the Big Run junction. For point 2, HEC-HMS will apply area-reduction factors for a storm
area of 158 square miles and run the entire basin model.
9. Results for the depth-area analysis should be viewed from the Results tab of the Watershed Explorer, as
shown in Figure 7. The Peak Flow Summary table contains the correct "area-reduced" results for each
analysis point. Below the summary table is a separate folder with results for each of the analysis points.
Notice the folder only contains results for the elements upstream of the analysis point in Figure 8.
2. Click New… and create a new depth-area analysis named Depth-Area Hypothetical. Click Next.
3. Select basin model Base 1%. Click Next.
4. Select meteorologic model Hypothetical Storm Freq 1%. Click Finish.
5. After the depth-area analysis has been created, go to the Compute tab and select the depth-area analysis to
open the Component Editor.
6. Enter a start date of 01Jan1999, a start time of 01:00, an end date of 11Jan1999, and an end time of 00:00.
Set the Time Interval to 15 minutes as shown in Figure 10.
7. Select the Analysis Points tab. Select the Big Run junction as Point 1 and Punxsutawney junction as Point
2 as shown in Figure 11.
8. Run the depth-area analysis. There are really two simulations that happen during the compute as shown on
Figure 12. For point 1, HEC-HMS will apply area-reduction factor for a storm area of 70 square miles and run
the basin model to the Big Run junction. For point 2, HEC-HMS will apply area-reduction factor for a storm
area of 158 square miles and run the entire basin model.
9. Results for the depth-area analysis should be viewed from the Results tab of the Watershed Explorer, as
shown in Figure 13. The Peak Flow Summary table contains the correct "area-reduced" results for each
analysis point. Below the summary table is a separate folder with results for each of the analysis points.
Notice the folder only contains results for the elements upstream of the analysis point in Figure 14.
22.1 Introduction
HEC-HMS version 4.6 has new features122 that support precipitation-frequency based rainfall-runoff modeling:
• A new "Precipitation-Frequency Grid" data type
• Capability to import ASCII-raster format grids to populate the Precipitation-Frequency Grid
• The Hypothetical Storm precipitation method of the meteorologic model can use a Precipitation-Frequency
Grid as the source for precipitation depths
• The Hypothetical Storm precipitation method with a Precipitation-Frequency Grid can be used in the Depth-
Area Analysis simulation type
This guide will provide information useful for creating an HEC-HMS model that allows you to perform hydrologic
simulations with precipitation boundary conditions created by using precipitation-frequency grids. For most of the
United States, the primary source of precipitation depth-duration-frequency data is NOAA Atlas 14123, with the
exception of the northwestern US. Depth-duration-frequency information for Washington, Oregon, Idaho, Montana,
and Wyoming with similar coverage to Atlas 14 is a combination of data from Arkell and Richards (1986)124, NOAA
Atlas 2125, and TP-49126. Other studies covering this region have been performed, such as Schaefer and Barker (2005)
for Oregon127, or Schaefer et al., (2006) for Washington128 (not an exhaustive list.) Atlas 14 contains frequency
estimates for precipitation durations ranging from 5-minute to 60-day totals, and for frequencies from 1/2 AEP to
1/1,000 AEP. Additionally, Atlas 14 produces grid data that shows the depth at any point in the study area for a
selected duration and frequency. NOAA Atlas 2 produces similar grids, except only for the 1/2 and 1/100 AEP
frequencies and 6-hour and 24-hour durations.
Products such as Atlas 14 are point precipitation frequency products, meaning the grid is a representation of the
depth-duration-frequency relationship for a single point in space. The pattern in a grid of point precipitation
frequency values is not realistic for an actual storm event because it assumes the same frequency event (e.g. a 1%
AEP 24-hour storm) is occurring at all locations simultaneously. Typically, depth-area reduction is used to adjust
point values to area-averaged values. A depth-area reduction curve expresses the amount by which the area-
averaged precipitation decreases from the point (or 10 mi2) maximum value for storm events. They may be derived
from a single storm, or by generalizing several storm events. The area component of a depth-area reduction refers
to the overall area, or footprint, of the storm. For hydrologic simulations, a frequently-made simplifying assumption
is to set the storm area equal to the drainage area of the watershed above the location being analyzed. Analyses
that consider multiple locations then have a varying drainage area and therefore storm area, which changes the
area reduction factor for computing area-averaged precipitation. The Depth-Area Analysis in HEC-HMS129
automatically handles this operation.
The HEC-HMS model you are using must be georeferenced, either by creating the basin model using the HEC-HMS
delineation tools(see page 349), by importing georeferenced elements(see page 345), or by georeferencing an existing
model(see page 340).
Consider reviewing the Applying the Depth Area Analysis Simulation Type(see page 449) tutorial for additional details
on using the Depth-Area analysis.
122 https://www.hec.usace.army.mil/confluence/display/HMSUM/.v.4.6.0+Release+Notes+v4.6
123 https://hdsc.nws.noaa.gov/hdsc/pfds/
124 https://www.nws.noaa.gov/oh/hdsc/PF_documents/Arkell_Richards.pdf
125 https://www.nws.noaa.gov/ohd/hdsc/noaaatlas2.html
126 https://www.nws.noaa.gov/oh/hdsc/PF_documents/TechnicalPaper_No49.pdf
127 http://www.mgsengr.com/damsafetyfiles/NRC2013/OregonPrecipitationStudy_Report_Final.pdf
128 https://www.wsdot.wa.gov/research/reports/fullreports/640-1.pdf
129 https://www.hec.usace.army.mil/confluence/display/HMSUM/.Depth-Area+Analyses+v4.4
130 https://hdsc.nws.noaa.gov/hdsc/pfds/pfds_gis.html
This will launch a dialog box that allows you to name and set the data type for the new Grid Data. A convention is to
use the standard filename from Atlas 14 or Atlas 2 (if those are the data source) for the grid name, as this encodes
the study area, AEP, duration, and type of analysis that create the dataset. Choose the "Precipitation-Frequency
Grids" data type (Figure 3).
131 https://hdsc.nws.noaa.gov/hdsc/pfds/pfds_temporal.html
132 https://www.hec.usace.army.mil/confluence/display/HMSUM/.Precipitation+v4.6
interest by selecting the Basins tab and setting "Include Subbasins" to "Yes" for the basin model. Open the
Hypothetical Storm editor by selecting it in the Watershed Explorer beneath the Meteorologic Model. Set the
Method to User-Specified Pattern, and the Precipitation Method to Precipitation-Frequency Grid (Figure 12.)
Figure 12. Hypothetical Storm options that enable using the Precipitation-Frequency Grid selected.
Next, set the Storm Pattern in the drop-down to the pattern entered previously. If the pattern entered previously is
not available, check to make sure it was entered as a Percentage Curve.
Then, set the storm duration equal to the duration specified in the Precipitation-Frequency Grid.
Next, set the Grid to the Precipitation-Frequency Grid created previously.
Finally, set the Area Reduction method. The TP-40 area reduction curve is available, but if you choose to use it,
should only be used with 24-hour duration storms. If you choose the User-Specified option, you will need to
supply an area-reduction function (created as a Paired Data component with the Area-Reduction Function type.)
Once an area reduction method is set, the option to set a Computation Point becomes available. This point will be
used to compute the storm area but will be overridden by analysis points in the Depth-Area Analysis. If this
Hypothetical Storm is used in a regular Simulation Run, it will use the storm area corresponding to the drainage
area above this computation point. A fully-parameterized Precipitation-Frequency Grid Hypothetical Storm is
shown in Figure 13.
23.1 Background
An optimization trial is a simulation type that can identify model parameters which improve the model results
compared to observations. Each trial is composed of a basin model, meteorologic model, and information about
the simulation time window and time step. The trial also includes selections for the objective function, search
method, and adjustable parameters. Optimization involves automated parameter adjustments so that the
simulated results are similar to observations. The objective functions measure the goodness of fit between the
simulated results and observed data. In this tutorial, you will learn how HEC-HMS can be used to optimize a model
by adjusting unit hydrograph and loss rate parameters, given measured discharge and precipitation data.
23.2 Overview
In this workshop you will set up and compute an optimization trial and review results of the trial and make
additional trials as necessary.
1. Open HEC-HMS and select the File menu. Choose the Open… menu option and select the Optimization
Workshop project.
2. Create a discharge gage in HEC-HMS following information in this tutorial - Creating Time Series Data(see
page 70). The observed flow gage will be used by the optimization trial, results from the simulation will be
compared to the observed flow gage and an objective function value will be computed for each iteration
during the optimization.
3. Open the Punxsutawney basin model by clicking on it in the Watershed Explorer. Click on the subbasin
Mahoning Cr in the Watershed Explorer or in the basin map to make it the selected element; its
Component Editor will automatically load.
4. Click on the Options tab in the Component Editor. Select the PNXP gage, created in step 2, for the
Observed Flow as shown in Figure 1.
5. Create three new simulation runs and compute them to make sure they function correctly. All three will have
the same basin model, Punxsutawney, and the same meteorologic model, Specified Gage Wt. Each run will
use a different control specifications: Apr 94, May 95, and May 96 (use descriptive names for the simulation
runs, Run Apr 94, Run May 95, and Run May 96) as shown in Figure 2.
6. Create the first optimization trial using the wizard on the Compute menu as shown in Figure 3.
7. The first trial will be based on the simulation run that uses the Apr 94 control specifications. Choose a name
for the trial (Opt Apr 94), and then select the Punxsutawney basin model, subbasin with observed flow
(Mahoning Cr), and Specified Gage Wt meteorologic model. Select the new optimization trial in the
Watershed Explorer (go to the Optimization Trials folder on the Compute tab) and move to the
Component Editor. Verify that the correct Basin Model and Meteorologic Model are selected. Enter a start
and end date that match the control specifications for the Apr 94 simulation and select a Time Interval of 1-
hour as shown in Figure 4.
8. On the Search tab, select the Simplex method, set a tolerance of 0.01, and a maximum number of iterations
of 50 as shown in Figure 5.
9. Switch to the Objective tab. Verify that the objective function will be evaluated at the Mahoning Cr subbasin
element. Select Minimization as the goal, and set the statistic to Peak-Weighted RMSE. Make sure the
objective function will be evaluated from the beginning to the end of the control specifications as shown in
Figure 6.
10. Right-click on the optimization trial node in the Watershed Explorer and select the Add Parameter
command as shown in Figure 7. A new parameter node will be added under the objective function node. Add
three more parameters the same way so there are a total of four parameters.
11. Click on the first parameter node and then move to the Component Editor. First select the Mahoning Cr
subbasin element. Select the Clark Unit Hydrograph - Storage Coefficient for the parameter as shown in
Figure 8.
12. Click on the other parameter nodes and select the Clark Unit Hydrograph - Time of Concentration, Initial
and Constant - Constant Rate, and Initial and Constant - Initial Loss parameters as shown in Figure
9, Figure 10, and Figure 11. In practice, edit the minimum and maximum values to reflect a reasonable
parameter range for the watershed.
13. Use the Compute menu to select and compute the optimization trial as shown in Figure 12.
14. A Compute screen will open to show progress. When the compute is completed, close it as shown in Figure
13.
15. Switch to the Results tab of the Watershed Explorer and examine the various results for the optimization
trial.
16. Open the Objective Function Summary table to compare the volume, peak flow, and time of peak between
the simulated and observed hydrographs as shown in Figure 14.
17. Open the Optimized Parameters table to examine initial values and optimized values as shown in Figure 15.
18. You can also assess the model's goodness of fit by simply looking at the time-series graph as shown in Figure
16. Choose the Observed Data graph. The first tab is the time-series graph and is shown below in Figure 16.
Inspection of the figure shows that the initial period of baseflow recession, before the event begins, matches
somewhat poorly. The rising limb of the hydrograph is a pretty good match, however. Other graphs can also
help us to see where the match is good and where it is poor.
19. The Compare graph (second tab) shown in Figure 17 and the Residuals graph (third tab) shown in Figure 18
can help. The points in the flow comparison graph should be on a 45 degree line if there is an exact match.
You can see in Figure 17 that they only fall around the 45 degree line but in general are higher or lower than
it. While there may be a linear relationship between the computed and observed flow after the peak, the
points do not fall on a 45 degree line. The residuals in Figure 18 are computed by subtracting the observed
flow from the computed flow. Negative residuals indicate the computed flow is too low, while positive
residuals indicate the computed flow is too high. Ideally, the residuals would have a mean of zero and small
variance. You can see in Figure 18 that the residuals swing back and forth about zero, which is consistent
with the spread around the 45 degree line on the comparison graph (zero residual would fall directly on that
45 degree line).
20. The Objective Function graph shows the objective function value at each step in the search process. Figure
19 shows the objective function graph for the first optimization trial. For a good search evolution, the value
of the objective function should decrease. The search results are not valid if the objective function fails to
decrease or fluctuates wildly. In Figure 19, it is not possible to tell if the search method actually converged
to a good solution. It would be better to make the search tolerance smaller and increase the number of
maximum iterations allowed and then recompute the optimization trial. It should also be noted that there
are several warnings that the compute did not converge within the specified maximum number of iterations,
so it is likely that finding the minimum will require additional iteration.
21. Parameter 1 graph shows the Clark Unit Hydrograph - Storage Coefficient values with each iteration as
shown Figure 20.
24.1 Introduction
Rain-on-snow events are a common occurrence on mountain slopes within the snow transition zone in the
mountainous West. This is especially true for the rivers in the Sierra Nevada Mountains which typically experience
their highest flow rates when warm winter storms interact with extensive snow cover (Kattelmann, 1997). The
effects of rainfall on snowmelt can significantly increase flooding and associated damages. The U.S. Army Corps of
Engineers (USACE) found that the major source of energy for snowmelt during rainfall is the convective transfer of
sensible and latent heats from the atmosphere to the snow (USACE, 1956).
Within version 4.4, the Temperature Index method was enhanced to allow for subbasin-level parameterization in
addition to new wet and dry meltrate methods.
Given the significance of rain-on-snow floods and the forthcoming snowmelt methods within HEC-HMS, a
comprehensive evaluation of the current state-of-the-practice methods and techniques for simulating rain-on-snow
processes was proposed. This evaluation is attempting to answer the following questions:
• Given recent advancements in snow hydrology, what is the best method to use?
• Do more complex snow modeling approaches offer improved accuracy for short/long time scales, small/
large areas, etc?
Additionally, this evaluation aims to provide guidance to users regarding where to find and how to effectively use
data within snowmelt applications. Finally, the amount of effort required for data preparation as well as
computational time will be tracked and reported for each method.
approximately 105 miles north and east into Nevada. Major population centers within the Truckee River watershed
include the cities of Truckee, CA, Reno, NV, and Sparks, NV with a total population greater than 350,000 (as counted
during the 2010 census). While the Truckee River provides a major source of water for its nearby inhabitants, the
river is also prone to significant flooding. Substantial damage has occurred during large scale and widespread rain-
on-snow flood events in 1950, 1955, 1996-1997, 2005-2006, and 2016-2017, as shown in Figure 3.
Numerous state and federal water supply and flood risk management projects have been constructed within this
watershed (USACE, 1985). Several major federally-constructed projects include Martis Creek, Prosser, Stampede,
and Boca dams, as shown in Figure 4.
The Truckee River watershed offers an excellent opportunity to investigate snowmelt modeling techniques within a
real-world watershed that contains an extremely large amount of meteorological, snow, river, and reservoir/lake
observation stations. Approximately 50 active stream, reservoir, and/or lake gages are maintained by the U.S.
Geological Survey (USGS) while approximately 20 Snow Telemetry (SNOTEL) stations are maintained by the Natural
Resources Conservation Service (NRCS), as shown in Figure 5 and Figure 6.
HEC-HMS at each SNOTEL station was compared against observed data, as shown in Figure 8. Temperature Index
parameters were varied in order to achieve adequate calibration using several statistical metrics.
The calibrated Temperature Index parameters achieved during this process were found to vary throughout the
modeling domain. For instance, the calibrated PX and Base temperatures varied by approximately 5 degrees, as
shown in Figure 9.
Similar to the point calibration effort, the distributed Temperature Index parameters achieved during this process
were found to vary throughout the modeling domain. For instance, the calibrated PX and Base temperatures varied
by approximately 7 degrees, as shown in Figure 11.
methods. The results of these simulations will allow for a direct comparison of the benefits and drawbacks unique
to each snowmelt/accumulation method. Model-derived outputs will be compared and notable differences will be
identified, investigated, and described. Finally, recommendations for selecting a snowmelt/accumulation method
will be made given typical study conditions, varying levels of observed data availability, accuracy, uncertainty
considerations, schedule requirements, and available budget through the use of the aforementioned comparisons.
Training materials will be developed for use within USACE training courses geared towards snowmelt/accumulation
modeling. Additionally, HEC-HMS supporting materials (user’s manual, technical reference manual, and
applications guide) will be developed and included within future software releases.
The Temperature Index - Distributed Calibration HEC-HMS project can be downloaded here:
24.7 References
Anderson, E. (2006). Snow Accumulation and Ablation Model - SNOW-17. Natural Resources Conservation Service.
Follum, M. L., Downer, C. W., Niemann, J. D., Roylance, S. M., & Vuyovich, C. M. (2015). A Radiation-Derived
Temperature-Index Snow Routine for the GSSHA Hydrologic Model. Journal of Hydrology, 723-736.
Kattelmann, R. (1997). Very Warm Storms and Sierra Nevada Snowpacks. Proceedings of the Western Snow
Conference 65, (pp. 125-129).
National Operational Hydrologic Remote Sensing Center. (2018, October). Snow Data Assimilation System (SNODAS)
Data Products at NSIDC, Version 1. Retrieved from https://doi.org/10.7265/N5TB14TC.
Tarboton, D. G., & Luce, C. H. (1996). Utah Energy Balance Snow Accumulation and Melt Model (UEB). Logan, Utah:
Utah State University.
U.S. Army Corps of Engineers. (1956). Summary Report of the Snow Investigations, Snow Hydrology. Portland, OR:
North Pacific Division.
U.S. Army Corps of Engineers. (1985). Truckee River Basin Reservoirs, Water Control Manual. Sacramento, CA:
Sacramento District.
U.S. Army Corps of Engineers. (1998). Runoff from Snowmelt, Engineer Manual 1110-2-1406. Washington, D.C.: USACE.
U.S. Army Corps of Engineers. (2015). Corps Water Management System, Final Report for the Truckee River Watershed.
Sacramento, CA: Sacramento District.
24.8 Introduction
Rain-on-snow events are a common occurrence on mountain slopes within the snow transition zone in the
mountainous West. This is especially true for the rivers in the Sierra Nevada Mountains which typically experience
their highest flow rates when warm winter storms interact with extensive snow cover (Kattelmann, 1997). The
effects of rainfall on snowmelt can significantly increase flooding and associated damages. The U.S. Army Corps of
Engineers (USACE) found that the major source of energy for snowmelt during rainfall is the convective transfer of
sensible and latent heats from the atmosphere to the snow (USACE, 1956).
temperature that is either used or generated by the Temperature Index method within HEC-HMS are shown in
Figure 1.
Within version 4.4, the Temperature Index method was enhanced to allow for subbasin-level parameterization in
addition to new wet and dry meltrate methods.
However, the Energy Balance method suffers from the need for a large amount of highly accurate meteorologic
inputs that are not readily available in most watersheds.
Given the significance of rain-on-snow floods and the forthcoming snowmelt methods within HEC-HMS, a
comprehensive evaluation of the current state-of-the-practice methods and techniques for simulating rain-on-snow
processes was proposed. This evaluation is attempting to answer the following questions:
• Given recent advancements in snow hydrology, what is the best method to use?
• Do more complex snow modeling approaches offer improved accuracy for short/long time scales, small/
large areas, etc?
Additionally, this evaluation aims to provide guidance to users regarding where to find and how to effectively use
data within snowmelt applications. Finally, the amount of effort required for data preparation as well as
computational time will be tracked and reported for each method.
Numerous state and federal water supply and flood risk management projects have been constructed within this
watershed (USACE, 1985). Several major federally-constructed projects include Martis Creek, Prosser, Stampede,
and Boca dams, as shown in Figure 4.
The Truckee River watershed offers an excellent opportunity to investigate snowmelt modeling techniques within a
real-world watershed that contains an extremely large amount of meteorological, snow, river, and reservoir/lake
observation stations. Approximately 50 active stream, reservoir, and/or lake gages are maintained by the U.S.
Geological Survey (USGS) while approximately 20 Snow Telemetry (SNOTEL) stations are maintained by the Natural
Resources Conservation Service (NRCS), as shown in Figure 5 and Figure 6.
HEC-HMS at each SNOTEL station was compared against observed data, as shown in Figure 8. Temperature Index
parameters were varied in order to achieve adequate calibration using several statistical metrics.
The calibrated Temperature Index parameters achieved during this process were found to vary throughout the
modeling domain. For instance, the calibrated PX and Base temperatures varied by approximately 5 degrees, as
shown in Figure 9.
Similar to the point calibration effort, the distributed Temperature Index parameters achieved during this process
were found to vary throughout the modeling domain. For instance, the calibrated PX and Base temperatures varied
by approximately 7 degrees, as shown in Figure 11.
methods. The results of these simulations will allow for a direct comparison of the benefits and drawbacks unique
to each snowmelt/accumulation method. Model-derived outputs will be compared and notable differences will be
identified, investigated, and described. Finally, recommendations for selecting a snowmelt/accumulation method
will be made given typical study conditions, varying levels of observed data availability, accuracy, uncertainty
considerations, schedule requirements, and available budget through the use of the aforementioned comparisons.
Training materials will be developed for use within USACE training courses geared towards snowmelt/accumulation
modeling. Additionally, HEC-HMS supporting materials (user’s manual, technical reference manual, and
applications guide) will be developed and included within future software releases.
The Temperature Index - Distributed Calibration HEC-HMS project can be downloaded here:
24.14 References
Anderson, E. (2006). Snow Accumulation and Ablation Model - SNOW-17. Natural Resources Conservation Service.
Follum, M. L., Downer, C. W., Niemann, J. D., Roylance, S. M., & Vuyovich, C. M. (2015). A Radiation-Derived
Temperature-Index Snow Routine for the GSSHA Hydrologic Model. Journal of Hydrology, 723-736.
Kattelmann, R. (1997). Very Warm Storms and Sierra Nevada Snowpacks. Proceedings of the Western Snow
Conference 65, (pp. 125-129).
National Operational Hydrologic Remote Sensing Center. (2018, October). Snow Data Assimilation System (SNODAS)
Data Products at NSIDC, Version 1. Retrieved from https://doi.org/10.7265/N5TB14TC.
Tarboton, D. G., & Luce, C. H. (1996). Utah Energy Balance Snow Accumulation and Melt Model (UEB). Logan, Utah:
Utah State University.
U.S. Army Corps of Engineers. (1956). Summary Report of the Snow Investigations, Snow Hydrology. Portland, OR:
North Pacific Division.
U.S. Army Corps of Engineers. (1985). Truckee River Basin Reservoirs, Water Control Manual. Sacramento, CA:
Sacramento District.
U.S. Army Corps of Engineers. (1998). Runoff from Snowmelt, Engineer Manual 1110-2-1406. Washington, D.C.: USACE.
U.S. Army Corps of Engineers. (2015). Corps Water Management System, Final Report for the Truckee River Watershed.
Sacramento, CA: Sacramento District.
25.1 Introduction
HEC-HMS version 4.7 includes a new 2D Diffusion Wave transform method. It is common knowledge that unit
hydrograph theory transforms excess precipitation to runoff hydrographs without detailed consideration of
complicated internal processes (Bartles, 2014; Szilagyi, 2007; Harrison, 1999). As such, the empirical equations and
parameters that go into many commonly employed synthetic unit hydrograph methods tend to have limited
physical significance. To “ground truth” these methods and parameters, parameters are adjusted through a model
calibration process where model outputs are compared to observed data in order to achieve an adequate “fit”.
Normally, calibrated modeling parameters are dependent upon the magnitude of the event being investigated. For
instance, according to Sherman, who originally proposed the unit hydrograph concept, the unit hydrograph of a
watershed is “…the basin outflow resulting from one unit of direct runoff generated uniformly over the drainage
area at a uniform rainfall rate during a specified period of rainfall duration” (Sherman, 1932). This implies that
ordinates of any hydrograph resulting from a quantity of runoff-producing rainfall of unit duration would be equal
to corresponding ordinates of a unit hydrograph for the same areal distribution of rainfall, multiplied by the ratio of
rainfall excess values. For example, 2 units of excess precipitation would result in a runoff hydrograph that is two
times larger than the unit hydrograph. The unit hydrograph is only scaled by a factor of 2; there is no adjustment to
the shape of the unit hydrograph. However, due to differences in areal distributions of rainfall and hydraulic
reactions between large and small precipitation events, the corresponding unit hydrographs have not been found
to be equal, as implied by unit hydrograph theory (Minshall, 1960). Unlike unit hydrograph theory, the 2D Diffusion
Wave transform method explicitly routes excess precipitation throughout a subbasin element using a combination
of the continuity and momentum equations. As such, this new transform method can be used to simulate the non-
linear movement of water throughout a subbasin when exposed to large amounts of excess precipitation.
This guide provides step-by-step instructions that can be used to create a new HEC-HMS project, delineate a
subbasin element, and compute a runoff hydrograph using the new 2D Diffusion Wave transform method.
Streamflow
Streamflow data for the previously mentioned gage was downloaded from the USGS National Water Information
System (NWIS) database using plugins available within HEC-DSSVue. These plugins import data from the USGS
NWIS database and create new time series records within a DSS file. Specifically, the instantaneous peak
streamflow, daily average flow, and 15-min flow time series for early April 2018 were downloaded and are shown in
Figure 2. This time period will be used to calibrate the HEC-HMS model.
Terrain
A 1/9th arc-second digital elevation model (DEM) was downloaded from the USGS National Map Viewer: https://
viewer.nationalmap.gov/basic/. Within the area of interest, 1/9th arc-second equates to a horizontal resolution of
approximately 3 meters. The DEM was first clipped using a buffered polygon that surrounded the area of interest.
Then, the DEM was projected to the Albers Equal Area Conic coordinate system and the vertical units were
converted to feet. The resultant DEM is shown in Figure 3.
Hydrography
Stream centerlines from the National Hydrography Data Plus (NHDPlus) data set were downloaded: https://
www.usgs.gov/core-science-systems/ngp/national-hydrography/nhdplus-high-resolution. The stream centerlines
were clipped using a buffered polygon that surrounded the area of interest. These stream centerlines are shown in
Figure 4.
Land Use
Land use classifications were downloaded from the Multi-Resolution Land Characteristics Consortium: https://
www.mrlc.gov/data?f%5B0%5D=category%3Aland%20cover. The National Land Cover Database (NLCD) 2016
version represents the most up-to-date nationwide source for land cover classifications. Within the area of interest,
land uses range from developed spaces to emergent herbaceous wetlands. As was previously stated, the
predominant land uses within the area of interest are deciduous and evergreen forests. The NLCD 2016 land cover
classifications are shown in Figure 5.
Precipitation
Hourly Multi-Radar/Multi-Sensor (MRMS) precipitation for April 2018 was downloaded from the Iowa State
University Iowa Environmental Mesonet: https://mesonet.agron.iastate.edu/. MRMS precipitation data is created
by combining data from multiple sources including radar stations, surface and upper air observations, lightning
detection systems, and satellite and forecast models. In general, this data represents the spatial and temporal
distribution of precipitation at an hourly time step very well. However, precipitation accumulations can be
underpredicted within mountainous areas when compared against runoff observations.
In order to correct the hourly MRMS precipitation accumulations, Parameter-elevation Regressions on Independent
Slopes Model (PRISM) precipitation for April 2018 was also downloaded: https://prism.oregonstate.edu/. In general,
this data better represents the spatial and temporal distribution of precipitation at a daily time step better than
MRMS.
Both the MRMS and PRISM data was projected to the Standard Hydrologic Grid (SHG), clipped to the area of
interest, and converted to DSS format using Vortex: https://github.com/HydrologicEngineeringCenter/Vortex.
Then, Vortex tools were used to "correct" the raw MRMS data using the PRISM data. In essence, each grid cell within
the MRMS data set was adjusted to to equal the corresponding grid cell within the PRISM data set. Basin-average
hyetographs were extracted for the raw MRMS data, PRISM data, and the corrected MRMS data and are shown
in Figure 6. Note that the basin average hyetographs are for visualization only; the HEC-HMS model will make use of
the gridded data.
133 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsguides/creating-an-hec-hms-model-from-scratch
134 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsguides/geo-referencing-existing-basin-model-elements-using-shapefile-
information
135 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsguides/updating-subbasin-and-reach-delineations-using-hec-hms-gis-tools
The "RedwoodCreek" subbasin was then exported as a shapefile using the GIS | Export Georeferenced Elements
tool.
Max Iterations 20
Number of Cores 4
Number of Layers 2
GW 1 Fraction 0.5
GW 1 Coefficient (hr) 5
GW 1 Steps 1
GW 2 Fraction 0.5
GW 2 Coefficient (hr) 30
GW 2 Steps 1
In order to better simulate the movement of water throughout the Redwood Creek watershed, a spatially
distributed Manning's n Layer was created within RAS Mapper. To allow for faster movement of water through
defined stream channels, a buffered polygon layer surrounding the previously described NHD streams was created
and imported to RAS Mapper. Then, a new Manning's n layer was created by combining the previously mentioned
NLCD 2016 land use classifications and the buffered streams polygon layer; the buffered streams polygon layer was
placed at the highest hierarchy (i.e. overrides the NLCD 2016 land use classifications). Initial estimates for
Manning's n were assigned using the following table:
Streams 0.06
Grassland/Herbaceous 0.1
Shrub/Scrub 0.1
A new Unsteady Flow file was then created and two boundary conditions were parameterized: 1) 0.1 inch of
spatially and temporally uniform precipitation applied to the 2D mesh and 2) a friction slope of 0.006 ft/ft was
applied to the "Outlet" boundary condition line; the precipitation depth is meaningless as it is only used to allow for
a successful simulation. Next, a new Unsteady Plan was created and the previously described geometry and
unsteady flow files were selected. A valid simulation time window and computation settings were set to allow for a
successful simulation. Finally, the unsteady flow simulation was computed, which generated an Unsteady Plan
HDF file, which has an extension of ".p##.hdf" where "p##" corresponds to the specific plan of interest. This file
contains the computational mesh and boundary condition line information which are currently needed by HEC-
HMS.
On the second panel, the "Apr2018" basin model was selected, as shown in Figure 11.
On the third panel, the "RedwoodCreek" subbasin was selected, as shown in Figure 12.
On the fourth panel, the "RedwoodCreek" 2D Mesh was selected, as shown in Figure 13.
Upon clicking Finish, the 2D mesh was imported, the "RedwoodCreek" subbasin's Discretization was changed to
Unstructured, and a 2D Connection node (called "Outlet") was added. Also, the 2D Mesh and 2D Connection are
shown within the Map Panel, as shown in Figure 14. Note: you may need to click View | Map Layers and turn on the
Discretization and 2D Connection map layers.
Finally, the "Outlet" 2D Connection was connected to the "MuirBeach" sink, the Type was set to Normal Depth, a
Friction Slope of 0.006 ft/ft was entered, and a Ratio of Subbasin Baseflow of 1.0 was set, as shown in Figure 15. The
Ratio of Subbasin Baseflow is defined as the ratio of baseflow generated within the "RedwoodCreek" subbasin that
will reach the particular 2D Connection.
Also, time series for the "Outlet" 2D Connection were plotted and are shown in Figure 17.
Finally, spatial variables like hydraulic depth (computed as the volume of water within a cell, divided by the wetted
area of the cell) were plotted and are shown in Figure 18.
At this point, the modeler can make changes to the loss, baseflow, and 2D Connection methods/parameterizations
to provide a better response throughout the watershed. However, HEC-HMS does not currently allow the user to
modify Manning's n values within a 2D mesh. This must be accomplished within HEC-RAS by generating a new
Unsteady Plan HDF file and importing to HEC-HMS. In the future, users will be able to create and modify Manning's
n Layers entirely within HEC-HMS in addition to many new features related to 2D flow.
25.4 References
Bartles, M. (2014). Hydrologic Analysis of Prompton Dam Using a Physically-Based Rainfall Runoff Model. Villanova,
PA: Villanova University.
Harrison, J. (1999). PMF and the Unit Hydrograph - Piquiing More Than Interest. (pp. 373-388). Lexington, KY:
Association of State Dam Safety Officials.
Minshall, N. E. (1960). Predicting Storm Runoff on Small Experimental Watersheds. ASCE Journal of the Hydraulics
Division, HY8, 17-37.
Sherman, L. K. (1932). Streamflow from Rainfall by the Unit Graph Method. Eng. News Rec.
Szilagyi, J. (2007). Analysis of the Nonlinearity in the Hillslope Runoff Response to Precipitation Through Numerical
Modeling. Journal of Hydrology, 391-401.
NHD: Buto, S.G., and Anderson, R.D., 2020, NHDPlus High Resolution (NHDPlus HR)---A hydrography framework for
the Nation: U.S. Geological Survey Fact Sheet 2020-3033, 2 p., https://doi.org/10.3133/fs20203033.
MRLC: https://www.mrlc.gov/data?f%5B0%5D=category%3Aland%20cover
3DEP: https://www.usgs.gov/core-science-systems/ngp/3dep/
USGS NWIS: https://waterdata.usgs.gov/nwis
MRMS: https://www.nssl.noaa.gov/projects/mrms/
PRISM: PRISM Climate Group, Oregon State University, http://prism.oregonstate.edu136, accessed August 2020.
136 http://prism.oregonstate.edu/
26.1 Introduction
The new 2D Diffusion Wave transform method within HEC-HMS version 4.7 explicitly routes excess precipitation
throughout a subbasin element using a combination of the continuity and momentum equations. As such, this new
transform method can be used to simulate the non-linear movement of water throughout a subbasin when exposed
to large amounts of excess precipitation. A tutorial detailing this use case is contained here: Creating a Simple 2D
Flow Model within HEC-HMS(see page 513).
In addition to routing excess precipitation, the new 2D Diffusion Wave transform method can be used to accurately
simulate the movement of water in both defined stream channels as well as overbank areas. This guide provides
step-by-step instructions that can be used to modify an existing HEC-HMS project to accurately simulate the
movement of water in a complex area using the new 2D Diffusion Wave transform method.
The Truckee River upstream of the United States Geological Survey (USGS) gage at Reno, NV (10348000; https://
waterdata.usgs.gov/nv/nwis/inventory/?site_no=10348000&agency_cd=USGS) drains over 1000 sq mi of rugged
terrain within the Sierra Nevada mountains. Upstream of this gage, numerous peaks in excess of 10,000 ft are
present along with extremely steep channels (slopes > 50 ft/mi) and incised canyons. However, in the vicinity of
Reno and Sparks, NV, channel slopes reduce to nearly 1 ft/mi and a floodplain several miles wide forms.
Additionally, the Truckee River receives two tributaries: 1) Steamboat Creek from the south and 2) North Truckee
Ditch from the north. Downstream of the USGS gage at Vista, NV (10350000; https://waterdata.usgs.gov/nwis/uv?
site_no=10350000), the floodplain contracts and the Truckee River enters another steep canyon, as shown in Figure
2.
While the Truckee River provides a major source of water for its nearby inhabitants, the river is also prone to
significant flooding. Substantial damage to urban areas has occurred during large scale and widespread rain-on-
snow flood events in January 1997, December 2005 - January 2006, and January - February 2017 (amongst others).
The extent of flooding during January 1997 within downtown Reno and at the Reno-Tahoe Airport are shown in
Figures 3 and 4, respectively.
These inaccuracies are due to: 1) the complex hydraulic interactions in and around the cities of Reno and Sparks, NV
and 2) the hydrologic routing simplifications within the existing HEC-HMS model. During this flood (and other large
floods), streamflow exceeded the channel capacity of the Truckee River and Steamboat Creek and began to spread
throughout the very wide floodplain. Also, the large amount of runoff was constricted by the natural topography
downstream of the Vista gage. These complex hydraulic scenarios cannot be accurately simulated using hydrologic
routing. To more accurately route water within the area of interest in and around the cities of Reno and Sparks, NV
during large floods, the existing HEC-HMS model was modified to use the new 2D Diffusion Wave transform. The
following sections detail the processes which were used to update this model.
Streamflow
Streamflow data for major USGS gages within the Truckee River watershed was downloaded from the USGS
National Water Information System (NWIS) database using plugins available within HEC-DSSVue. These plugins
import data from the USGS NWIS database and create new time series records within a DSS file. Specifically, the
instantaneous peak streamflow, daily average flow, and 15- or 60-min flow time series for December 1996 - January
1997 were downloaded.
Terrain
A 1/3rd arc-second digital elevation model (DEM) encompassing the entire Truckee River watershed was
downloaded from the USGS National Map Viewer: https://viewer.nationalmap.gov/basic/. Within the area of
interest, 1/3rd arc-second equates to a horizontal resolution of approximately 10 meters. The NED 10m DEM was
used to make modifications to the subbasin delineations within the existing HEC-HMS project. The NED 10m DEM
was clipped using a buffered polygon that surrounded the Truckee River watershed, projected to the Albers Equal
Area Conic coordinate system, and the vertical units were converted to feet. The resultant NED 10m DEM is shown
in Figure 7.
A 1 meter horizontal resolution DEM encompassing the area of interest in and around the cities of Reno and Sparks,
NV was also downloaded from the USGS National Map Viewer. The NED 1m DEM was used to populate properties
within the two-dimensional (2D) mesh within the area of interest. The NED 1m DEM was clipped using a buffered
polygon that surrounded the area of interest, projected to the Albers Equal Area Conic coordinate system, and the
vertical units were converted to feet. The resultant DEM is shown in Figure 8.
Land Use
Land use classifications were downloaded from the Multi-Resolution Land Characteristics Consortium: https://
www.mrlc.gov/data?f%5B0%5D=category%3Aland%20cover. The National Land Cover Database (NLCD) 2016
version represents the most up-to-date nationwide source for land cover classifications. The NLCD 2016
classifications were used to designate roughness values within the 2D mesh. The land use classifications within the
area of interest are shown in Figure 9.
Hydrography
Stream centerlines from the National Hydrography Data Plus (NHDPlus) data set were downloaded: https://
www.usgs.gov/core-science-systems/ngp/national-hydrography/nhdplus-high-resolution. The stream centerlines
were clipped using a buffered polygon that surrounded the area of interest. The stream centerlines were used to
designate stream channels (which have lower roughness values relative to the overbank areas) within the 2D mesh.
Precipitation
Hourly gaged precipitation for the time period of interest was downloaded from the National Centers for
Environmental Information (NCEI): https://www.ncdc.noaa.gov/. GageInterp was then used to create hourly
precipitation grids which captured the temporal and spatial resolution of precipitation throughout the Truckee
River watershed: https://www.hec.usace.army.mil/training/CourseMaterials/Mongolia_Workshop/
gageInterpUserManual.pdf.
Parameter-elevation Regressions on Independent Slopes Model (PRISM) precipitation for the time period of interest
was also downloaded: https://prism.oregonstate.edu/. In general, this data represents the spatial and temporal
distribution of precipitation much better than the gage precipitation (and resultant grids created by gageInterp).
The PRISM data was first projected to the Standard Hydrologic Grid (SHG), clipped to the area of interest, and
converted to DSS format using Vortex: https://github.com/HydrologicEngineeringCenter/Vortex. The PRISM data
was used to "correct" the gageInterp output at a daily time step. In essence, each grid cell within the gageInterp
grids was adjusted to to equal the corresponding grid cell within the PRISM data set.
Temperature
Hourly gaged temperature for the time period of interest was downloaded from the NCEI. GageInterp, along with a
lapse rate of 4.6 deg C / 1000 meter of elevation, was then used to create hourly temperature grids. This lapse rate
was developed using prevailing temperatures from nearby gages and their respective elevations during the time
period of interest.
Snow Data
Antecedent snow water equivalent (SWE) for 27 December 1996 was developed using gage information from NCEI
and Snow Telemetry (SNOTEL) stations that are maintained by the Natural Resources Conservation Service (NRCS).
137 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsguides/creating-an-hec-hms-model-from-scratch
138 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsguides/geo-referencing-existing-basin-model-elements-using-shapefile-
information
139 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsguides/updating-subbasin-and-reach-delineations-using-hec-hms-gis-tools
The "RenoSparks" subbasin was then exported as a shapefile using the GIS | Export Georeferenced Elements
tool. Finally, every subbasin except "RenoSparks" was set to use the Structured discretization method and grid
cells were computed by clicking GIS | Compute Grid Cells.
Max Iterations 20
Number of Cores 4
Number of Layers 2
GW 1 Fraction 0.22
GW 1 Steps 1
GW 2 Fraction 0.22
GW 2 Steps 1
In order to better simulate the movement of water throughout the area of interest, a spatially distributed Manning's
n Layer was created within RAS Mapper. To allow for faster movement of water through defined stream channels, a
buffered polygon layer surrounding the previously described NHD streams was created and imported to RAS
Mapper. Then, a new Manning's n layer was created by combining the previously mentioned NLCD 2016 land use
classifications and the buffered streams polygon layer; the buffered streams polygon layer was placed at the
highest hierarchy (i.e. overrides the NLCD 2016 land use classifications). Initial estimates for Manning's n were
assigned using the following table:
Streams 0.06
Grassland/Herbaceous 0.1
Pasture/Hay 0.1
Shrub/Scrub 0.1
A new Unsteady Flow file was then created and five boundary conditions were parameterized: 1) 0.1 inch of
spatially and temporally uniform precipitation applied to the 2D mesh, 2) a flow hydrograph using an energy grade
slope of 0.008 ft/ft for the "Truckee_inflow" boundary condition line, 3) a flow hydrograph using an energy grade
slope of 0.007 ft/ft for the "Steamboat_inflow" boundary condition line, 4) a flow hydrograph using an energy grade
slope of 0.007 ft/ft for the "NTruckeeDt_inflow" boundary condition line, and 5) a friction slope of 0.0002 ft/ft for the
"Truckee_outflow" boundary condition line; the precipitation depth is meaningless as it is only used to allow for a
successful simulation. Next, a new Unsteady Plan was created and the previously described geometry and
unsteady flow files were selected. A valid simulation time window and computation settings were set to allow for a
successful simulation. Finally, the unsteady flow simulation was computed, which generated an Unsteady Plan
HDF file, which has an extension of ".p##.hdf" where "p##" corresponds to the specific plan of interest. This file
contains the computational mesh and boundary condition line information which are currently needed by HEC-
HMS.
On the second panel, the "Jan_1997_2D" basin model was selected, as shown in Figure 14.
On the third panel, the "RenoSparks" subbasin was selected, as shown in Figure 15.
On the fourth panel, the "RenoSparks" 2D Mesh was selected, as shown in Figure 16.
Upon clicking Finish, the 2D mesh was imported, the "RenoSparks" subbasin's Discretization was changed to
Unstructured, and four 2D Connection nodes (called "Truckee_inflow", "Steamboat_inflow", "NTruckeeDt_inflow",
and "Truckee_outflow") were added. Also, the 2D Mesh and 2D Connections are shown within the Map Panel, as
shown in Figure 17. Note: you may need to click View | Map Layers and turn on the Discretization and 2D
Connection map layers.
The "Truckee_outflow" 2D Connection type was set to Normal Depth, a Friction Slope of 0.0002 ft/ft was set,
• and a Ratio of Subbasin Baseflow of 1.0 was entered
The Ratio of Subbasin Baseflow is defined as the ratio of baseflow generated within the "RenoSparks" subbasin that
will reach each 2D Connection. In this case, all baseflow generated within the "RenoSparks" subbasin was set to
reach the "Truckee_outflow" 2D Connection, as shown in Figure 18.
The existing "Jan1997" Control Specification was selected and a Start Date, Start Time, End Date, End Time, and
Time Interval of "27Dec1996", "06:00", "15Jan1997", "06:00", and "1 Hour" was used, respectively.
Finally, a new Simulation Run (called "Jan_1997_2D") was created. The previously mentioned Basin Model,
Meteorologic Model, and Control Specifications were selected and the simulation run was computed.
As is evident in this figure, the computed results match the observed streamflow much better than the results
which were previously presented in Figure 6.
Also, time series for each 2D Connection were plotted and investigated, as shown in Figure 21.
Finally, spatial variables like SWE, cumulative precipitation, hydraulic depth, water surface elevation, and velocity
were investigated. A tutorial showing how to turn on and view spatial results can be found here: Viewing Spatial
Results for a Structured Discretization(see page 559). As shown in Figure 22, the spatial results indicate streamflow
exceeding the Truckee River channel capacity starting just downstream of the Reno gage and spreading throughout
the floodplain. Also, the backwater effects caused by the natural topographic constriction in the vicinity of the Vista
gage results in reduced average cell velocities and excessive hydraulic depths throughout the area of interest, as
shown in Figure 23.
The computed hydraulic depths were compared withobservations of both depth and inundation extent, as shown
in Figure 24, and were found to compare well.
Videos of hydraulic depth, average cell velocity, and water surface elevation within the area of interest are shown
below. Also, videos of cumulative precipitation and SWE throughout the entire Truckee River watershed are shown
below.
At this point, the modeler can make changes to the loss, baseflow, and 2D Connection methods/parameterizations
to provide a better response throughout the watershed. However, HEC-HMS does not currently allow the user to
modify Manning's n values within a 2D mesh. This must be accomplished within HEC-RAS by generating a new
Unsteady Plan HDF file and importing to HEC-HMS. In the future, users will be able to create and modify Manning's
n Layers entirely within HEC-HMS in addition to many new features related to 2D flow.
26.3 References
NHD: Buto, S.G., and Anderson, R.D., 2020, NHDPlus High Resolution (NHDPlus HR)---A hydrography framework for
the Nation: U.S. Geological Survey Fact Sheet 2020-3033, 2 p., https://doi.org/10.3133/fs20203033.
MRLC: https://www.mrlc.gov/data?f%5B0%5D=category%3Aland%20cover
3DEP: https://www.usgs.gov/core-science-systems/ngp/3dep/
USGS NWIS: https://waterdata.usgs.gov/nwis
NOAA NCEI: https://www.ncdc.noaa.gov/
PRISM: PRISM Climate Group, Oregon State University, http://prism.oregonstate.edu140, accessed August 2020.
140 http://prism.oregonstate.edu/
HEC-HMS 4.7 Beta was used to create this tutorial. The above example project had to be broken up into multiple
downloads due to file size limitations. After you download all three files to your computer, unzip them and then
move the "data" and "gis" folders to the "MainstemColumbia" folder. The data folder should be located as follows
"...\MainstemColumbia\data" and the gis folder should be located as follows "...\MainstemColumbia\gis". Do not
attempt to open the MainstemColumbia project before correctly moving the "gis" and "data" folders.
The purpose of this tutorial is to provide information about the new spatial results option in HEC-HMS and provide
an example dataset. The figure below shows the example project and the Spatial Results toolbar is highlighted.
The Spatial Results toolbar includes options for selecting output results, an animation toolbar, buttons for
controlling the animation, and an Animation Setting button that opens an editor with options for color ramps,
scale, animation speed, and whether the animation should loop.
In order for spatial results to be computed, subbasin elements must be georeference. Spatial results will be
displayed at the discretization level for the subbasin elements. For example, results will be displayed as subbasin
averages when the None discretization method is selected. Results will be displayed on the SHG or UTM grid when
the Structured discretization method is selected. Results will be displayed on the 2D mesh when the Unstructured
discretization method is selected. The File-Specified discretization method offers mixed results whether results are
displayed at the subbasin or grid level. When using a *.mod file and gridded modeling methods, the program will
display results at the subbasin level since the *.mod file does not contain any geospatial information. The other
File-Specified options will have results displayed on the structured or unstructured grid.
The following steps demonstrate how to turn on Spatial Results and how to use tools on the Spatial Results
toolbar.
1. Open the MainstemColumbia project using HEC-HMS 4.7 Beta, or a newer version. Open the
MainstemColumbia_WY1997 basin model. Some of the GIS files had to be removed from the project due to
file size limitations. Select the GIS→Preprocess Drainage menu option. The program will compute the flow
direction and flow accumulation grids as part of the preprocess drainage step. This information is needed if
the modeler changes information about the structured grid.
2. The model domain is a portion of the Columbia River watershed. The model was configured to simulate part
of the 1997 water year, from October 2, 1996 through May 31, 1997. Gridded precipitation, gridded Hamon
ET, and the Gridded Temperature Index methods were used to compute boundary conditions to the basin
model. Subbasin elements were configured to use the Simple Canopy, Deficit and Constant Loss, ModClark
Transform, and Linear Reservoir Baseflow methods. Model parameters were adjusted to calibrate the model
to flow observations throughout the basin. All locations with observed flow are identified as Computation
Points. HEC-GeoHMS was used to delineate subbasins and reaches; however, terrain data was added to this
project and the flow direction and flow accumulation grids were created using the GIS tools in HEC-HMS.
Then a Structured 2000-meter SHG was created using the GIS→Compute→Grid Cells tool. You can visualize
the 2000-meter SHG by turning on the Discretization map layer as shown below.
3. Go to the Compute tab on the Watershed Explorer and click on the WY1997 Spatial Results simulation to
open the Component Editor. As shown below, make sure the Spatial Results option is set to Yes. You will
not be able to turn on Spatial Results if the subbasin elements are not georeferenced. By default, the output
interval for spatial data is set to the simulation time step. This might generate too much output in some
cases. For this example project, set the Spatial Interval to 1Day.
HEC-HMS will build a cache of animation tiles the very first time you attempt to visualize spatial results. The
message window below will pop up. Click Ok and then a progress dialog will open as shown below. It will
take approximately 2 minutes for the animation tiles to be created. Again, results processing only happens
the first time spatial results are viewed after running a simulation.
]
6. As shown below, drag the animation slider bar to December 19, 1996 to see the accumulated SWE at that
point in time. You can click the play button, , and the animation will loop through the simulation. You
can see the SWE grow in the Fall and Winter as the snowpack becomes larger and then the SWE decreases in
the Spring as the snowpack melts.
7. Click the Animation Settings button, , to open the Animation Setting editor as shown below. The
editor contains options for controlling the speed of the animation and the color ramp and scale used for the
selected output result. Press the Select Color Ramp and Scale button. For SWE, select the "Sequential
PurpleBlue" Color Ramp, set the Transparency to 0.1, and then choose a "Linear" Scale. Click the Apply
and then Close button to close the Color Ramp and Scale editor.
8. Since a Linear scale was chosen, enter 40 (inches) as the Maximum Value in the Animation Settings editor,
as shown below. Change the Frame Rate to 12 and then click the Apply and Close buttons. Play the
animation again to see the new color ramp and scale for SWE results.
9. Try choosing another result type and viewing the animation. After choose another result type, you will likely
need to edit the Color Ramp and Scale for the selected output. The following figure shows incremental
precipitation on December 3, 1996. The "Precipitation NWS Radar" color ramp and the "Precipitation NWS
The purpose of this tutorial is to provide information about discretization and transform methods in HEC-HMS. The
example project can be downloaded and opened to walk through information presented below. You must use HEC-
HMS version 4.9 to open the project. HEC-RAS 6.0 was used to generate the 2D grid and hydraulic property HDF file.
The example project also contains additional precipitation and flow data, for five additional flood events if you are
interested in testing and calibrating models to other floods. The 2D Diffusion Wave transform model and simulation
were developed following the same procedure shown here - Creating a Simple 2D Flow Model within HEC-HMS(see
page 513).
28.1 Introduction
HEC-HMS includes several transform methods. Some methods allow the modeler to enter the unit hydrograph
while synthetic unit hydrograph options allow the modeler to enter information about the subbasin response
(routing response) and then the program computes the unit hydrograph. There are also additional transform
options, the kinematic wave transform option and the ModClark transform method. The ModClark transform
method is a quasi-distributed unit hydrograph option where excess precipitation is routed from grid cells within a
gridded representation of the subbasin to the subbasin outlet using translation and a linear reservoir. The time
component (translation) is based on the travel length for each cell and the time of concentration for the subbasin.
The release of HEC-HMS 4.7 provided a new a new subbasin transform method, the 2D Diffusion Wave transform
method. The 2D Diffusion Wave transform method is a true cell to cell routing model, it is the same routing model
(code) in HEC-RAS.
The release of the new 2D transform method in HEC-HMS also coincides with the requirement for the modeler to
define a discretization method for a subbasin element. The selection of the discretization method helps when
configuring hydrologic processes and process parameters. Subbasins can be modeled as a lumped area, subdivided
into a structured grid (a standard hydrologic grid or a UTM grid), or subdivided into an unstructured mesh.
In addition to breaking a subbasin up into a grid, or mesh, watershed discretization also includes dividing the
watershed into smaller subbasins. The GIS delineation tools in HEC-HMS allow modelers to customize the
watershed delineation with many smaller subbasins or a few larger subbasins. HEC-HMS is flexible, a modeler can
choose only one subbasin to represent the entire watershed or many subbasins and reaches can be used to
represent smaller areas within the watershed. The decision about how to discretize a watershed model has
implications for the model's development, application, and future use. The level of discretization should be made
considering watershed characteristics and the boundary condition data available to calibrate the model. This
tutorial shows different discretization and transform methods and how they can be applied to an HEC-HMS
application.
Terrain data was downloaded from the USGS National Map Viewer: https://viewer.nationalmap.gov/basic/. The
terrain data was reprojected and the vertical units were converted to feet. Published streams from the National
Hydrograph Dataset (NHD) were burned into the terrain as well. The terrain data was used to delineate subbasin
and stream elements within HEC-HMS. The terrain data was also used by the HEC-RAS 2D preprocessor to extract
hydraulic properties for each cell and cell face in the 400-ft unstructured mesh. The following figure shows the
watershed on top of an aerial image; notice the watershed is in the coastal mountains.
Landuse data was gathered from the National Land Cover Database (NLCD) 2016. As shown below, land uses within
the watershed are mixed forest, grassland, and shrub/scrub. Soil data from the SSURGO database was used to
create raster datasets of soil properties, like saturated hydraulic conductivity. The soil property datasets were used
to estimate Green and Ampt parameters.
Precipitation and flow information were provided by staff at the California Department of Water Resources. Both
the precipitation and flow data were recorded at 15-minute intervals. The follow figure shows the precipitation
gage locations. The flow gage is located at the outlet of the watershed area. HEC-GageInterp was used to
interpolate the point rain gage data to create precipitation grid sets for six storm events (a 2000-meter grid spacing
was used for the interpolation). The PRISM 30-year average precipitation dataset was used as a bias grid in the
interpolation process.
The structured grid discretization method breaks the subbasin into uniform grid cells using either the Standard
Hydrologic Grid (SHG) or UTM Grid projections. The modeler can choose the grid cell size for both grid projections.
A 2000-meter SHG is shown on top of the Coyote Creek watershed below. The structure grid option can be used with
gridded or lumped (basin averaged) loss methods and with unit hydrograph and the ModClark transform methods.
When lumped loss and unit hydrograph methods are used (excluding ModClark), gridded precipitation is converted
to basin average precipitation. Gridded or non-gridded precipitation can be used with a structured grid.
The unstructured discretization method is used to create a 2D mesh that can be used with the 2D Diffusion Wave
transform method. Currently, the modeler must create the 2D mesh and compute the mesh properties in HEC-RAS
before referencing the mesh in HEC-HMS. A portion of a 400-ft unstructured mesh is shown on top of the Coyote
Creek watershed. The unstructured mesh option can be used with gridded or lumped loss methods and with unit
hydrograph, ModClark, and 2D Diffusion Wave transform methods. Gridded or non-gridded precipitation can be
used with an unstructured mesh.
The File-Specified discretization option allows the modeler to choose different file types that represent a
structured or unstructured discretization. The file types include *.mod, *.sqlite, *.hdf5 and *.hdf files. The *.mod file
is the old HEC-GeoHMS file type for a structured grid. HEC-HMS creates *.sqlite files to save structure grid
information when the structured grid is computed within HEC-HMS. The *.mod is being phased out as the *.sqlite
file format is the preferred option for structured grids (the *.sqlite file contains geometric information whereas the
*.mod file simply identifies which grid cells are located in each subbasin). Refer to this tutorial for information
about how to quickly create a structured grid for an old project that might have no geospatial information -
Applying Gridded Precipitation to a Non-Georeferenced Project - Structured Discretization(see page 379). The *.HDF
and *.HDF5 file formats are used to store the unstructured mesh information.
The following figures illustrate the linear assumption. The first figure shows a hydrograph generated by 1 unit of
precipitation. Using a linear assumption about runoff response, 2 units of precipitation generate hydrograph
ordinates that are 2 times larger than the 1 unit hydrograph. The peak flow is doubled, but the time of peak flow
does not change. The second figure shows a non-linear response where the hydrograph from 2 units of
precipitation is more than three time larger than the peak flow from 1 unit of precipitation and the peak flow occurs
15 minutes earlier.
The time-invariance assumption means the unit response does not change in time; the watershed will respond the
same for the case where there was one unit of precipitation directly followed by another unit of precipitation vs. the
case where there was a considerable amount of time separating the units of precipitation. It is expected that the
case where 1 unit of precipitation directly following precipitation will have a larger peak flow and shorter time to
peak flow due to excess precipitation from the preceding rainfall already in the channel network. Additional
precipitation while water is in the channel network will result in faster water velocities and an overall different
hydrologic response.
The following figure shows results from two scenarios (the scenarios were modeled with the 2D Diffusion Wave
transform method in HEC-HMS). The smaller runoff hydrograph was generated by one unit of precipitation that fell
on a dry watershed. The larger runoff hydrograph was generated by one unit of precipitation that fell six hours after
an event that also included 1 unit of precipitation. Notice the baseflow is much higher for the scenario with an
antecedent event. There was still water from the antecedent event moving across the watershed when the second
event occurred. This example shows that the antecedent event resulted in a shift of 30 minutes for the time of peak
flow and the antecedent event amplified the magnitude of peak flow.
12Subbasins_ModClark Basin model was divided into 12 subbasins. Green and Ampt and
ModClark parameters were estimated for each subbasin and the model
was calibrated to the 1982 event.
12Subbasins_Clark Basin model was divided into 12 subbasins. The Green and Ampt,
Clark, and linear reservoir parameters were set to be the same as those
in the 12Subbasins_ModClark basin model.
1Subbasin_ModClark One subbasin was used to model the watershed. Green and Ampt and
ModClark parameters were estimated and the model was calibrated to
the 1982 event.
1Subbasin_Clark One subbasin was used to model the watershed. The Green and Ampt,
Clark, and linear reservoir parameters were set to be the same as those
in the 1Subbasins_ModClark basin model.
CoyoteCreek_2D The 2D Diffusion Wave transform method was used with a 400-ft
unstructured mesh. Each cell within the mesh had the same Green and
Ampt parameter values.
Green and Ampt and baseflow parameters were adjusted during model
calibration. The surface roughness values were adjusted in HEC-RAS to
improve timing and magnitude of the routed flood hydrograph.
The following two figures show results from the 1982_1Subbasin_ModClark and 1982_1Subbasin_Clark
simulations for the December 1982 event. Both models used one subbasin to represent the watershed. The
1Subbasin_ModClark basin model used the ModClark transform method and gridded precipitation. The
1Subbasin_Clark basin model used the Clark transform method, which resulted in the gridded precipitation being
transformed to basin average before being applied to the loss computations.
Notice the difference in the simulated hydrographs in the second figure below. Unlike the 12 subbasin model,
discretizing the watershed with only one subbasin resulted in much more difference in the gridded (ModClark) and
non-gridded (Clark) simulations. The Clark basin model does not take advantage of the spatial variation in the 7
precipitation gages. Even though the 1Subbasin_ModClark basin model only has one subbasin representing the
entire watershed, use of the ModClark transform method means HEC-HMS will perform precipitation loss and
excess calculations for each of the 106 SHG 2000-meter grid cells in the subbasin before routing excess
precipitation to the outlet using the grid cell's time of concentration and the storage coefficient.
The approach followed is for illustrative purposes only. The CoyoteCreek_1Subbasin_Clark basin model could have
been calibrated for the 1982 event (the Green and Ampt parameters would be modified). However, this exercise
illustrates the need to re-calibrate the model when changing the model structure.
The following figure shows results from the 1982_2D simulation. Gridded precipitation was applied to a 400-ft
unstructured mesh. Green and Ampt infiltration was used to model loss and excess for each cell in the mesh.
Infiltrated water was routed to the outlet using the linear reservoir baseflow method. The 2D Diffusion Wave
transform method was used to route excess precipitation from grid cell to grid cell. The surface roughness values
were adjusted to improve timing of the peak flow rates and the Green and Ampt saturated hydraulic conductivity
12Subbasins_ModClark 7 seconds
12Subbasins_Clark 4 seconds
1Subbasin_ModClark 1 second
An important consideration when discretizing a model is the availability of boundary condition data to drive the
model. As we saw in the 1Subbasin simulations, using gridded precipitation interpolated from 7 rain gages vs. an
area average precipitation time-series for the entire watershed does result in differences in the computed flow. The
timescale of the boundary condition data is also important. The precipitation data used for this example was
recorded at 15-minute intervals. Subdividing a model into smaller subbasins and/or using the 2D Diffusion Wave
transform method should coincide with use of the most detailed precipitation data as possible, in both time and
space scales. Using coarse boundary condition data will likely result in unwanted parameter adjustments that are
needed to compensate for the quality of the precipitation data. Consider putting more effort into gathering
boundary condition data as the level of complexity of the model increases.
The physical characteristics of the watershed should also be considered when discretizing subbasin elements and
choosing a transform method. When defining subbasins, an important consideration is whether averaging
information across the area appropriate. Large changes in the watershed/stream slope might require delineation of
smaller subbasins. Use of the 2D Diffusion Wave transform method requires a different mindset than using unit
hydrograph methods. When using the 2D Diffusion Wave method, the watershed does not have to be subdivided
into smaller subbasin. One 2D subbasin can be used because physical characteristics of the watershed are
accounted for in the 2D mesh properties, and gridded loss parameters can be used as well.
The intended application of the model is an important consideration for selecting discretization and transform
methods. Any application that requires an uncertainty or sensitivity analysis where many, many simulations are
computed will have to factor in simulation run times. In this case, the 2D Diffusion Wave might not be the best
approach, but the 2D Diffusion Wave method could be used to formulate variable Clark TC and R parameters (HEC-
HMS includes a variable Clark method where a relationship between time of concentration and the storage
coefficient are made with excess precipitation intensity). In some cases, the model will be applied to hypothetical
events where the precipitation magnitude exceeds the magnitude of event used for model calibration. The linearity
assumption with the unit hydrograph approach illustrates the challenge with applying a model using unit
hydrographs to flood simulations larger than the calibration events. In this case, the 2D Diffusion Wave transform
method might be the better choice since the physics of surface flow are being modeled; which should account for
changes in the runoff response with larger precipitation rates. However, it is noted that the roughness coefficients
would not be calibrated for events larger than those used for model calibration.
Finally, where information is needed within the watershed is another important consideration when discretizing a
model. In the case of the Coyote Creek application, if flow information is only needed at the outlet of the watershed
then a simple representation with only one subbasin element provides the necessary information. The watershed in
this example is large enough that spatial precipitation and use of the ModClark transform method could improve
the model. Subdividing the model into smaller subbasins provides additional flow values within the watershed;
however, additional effort is required to calibrate the model due to additional elements. The 2D Diffusion Wave
transform method provides surface flow results and depths throughout the entire subbasin/unstructured mesh.
29.1 Introduction
The purpose of this tutorial is to show how the new 2D Diffusion Wave transform method can be used to create
relationships between excess precipitation and the Clark unit hydrograph parameters. One assumption with the
applications of unit hydrograph transform methods is that the runoff response is linear with the magnitude of
precipitation (2 units of excess precipitation generates a runoff response that is 2 times larger than the runoff
response from 1 unit of precipitation). In natural watersheds, runoff response is not linear. As precipitation rates
increase, channel velocities increase and natural storage in channels and floodplains change as well. The non-linear
runoff response is one reason guidance states that models should be calibrated to the magnitude of event used in
predictive simulations. In some applications, it is necessary to use a model to predict the runoff response for events
much larger than those even available for calibration. In these cases, model parameters could be modified, or
alternative models used.
The HEC-HMS team added an option to the Clark unit hydrograph method that allows modelers to vary the time of
concentration and storage coefficient parameters based on excess precipitation intensity. Developing the
relationship between time of concentration and storage coefficient parameters and excess precipitation can be
challenging. Historical storm events could be used if there were several storm events with varying precipitation
intensities. Another option is to use numerical models, like the 2D Diffusion Wave transform method, that includes
the physics of overland flow and would be able to model the non-linear change in runoff response as precipitation
rates increase.
This tutorial shows how to use the new 2D Diffusion Wave model in HEC-HMS (the same code in HEC-RAS), to create
the variable parameter relationships for the Clark unit hydrograph method. The variable Clark method has
application for studies where many, many simulations will be computed with moderate to extreme precipitation
rates to evaluate uncertainty and parameter sensitivity on a watershed's runoff response. It might not be possible
to run many events through the 2D Diffusion Wave transform solution in an uncertainty analysis. The variable Clark
option can capture the non-linearity in runoff response while computing in a fraction of the time.
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 584
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 585
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 586
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
A copy of the calibrated basin model was made and named CoyoteCreek_2D_peaking. The Loss and Baseflow
methods were set to None. The only hydrologic process turned on was the 2D Diffusion Wave transform method.
Five simulations were created for the hypothetical simulations. The simulation time window for the five
hypothetical simulations was 01 January 2000 at hour 00:00 though 06 January 2000 at hour 00:00. The following
table contains the basin and meteorologic model combinations for the five hypothetical storms.
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 587
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The following figure shows results from the simulation where the main precipitation was 1 inch/hour. Notice the
precipitation depth is a little larger than 0.016 inches in the plot. The simulation time-step was 1 minute, 1 inch/
hour divided by 60 minutes is 0.0167 inches. The meteorologic model applied 0.0167 inches per minute for a 1-hour
period. The resulting hydrograph is the runoff response from 1 inch of precipitation applied to the subbasin
element. The subbasin element only had the 2D Diffusion Wave transform option turned on, so only surface routing
computations were performed during the simulation.
The following figure shows a comparison of the runoff hydrograph from three of the hypothetical simulations. The
hydrographs do not represent unit hydrographs since they have not been normalized to one inch of precipitation.
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 588
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Notice how the time of the peak flow shifts as the precipitation intensity increases.
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 589
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The CoyoteCreek_2D_peaking basin model was copied, and the transform method was changed from 2D Diffusion
Wave to Clark Unit Hydrograph. Four additional copies were created. Five copies were needed in order to identify
a separate set of Clark parameter for each of the hypothetical precipitation simulations in step 2. Five simulations
were created as well that combined the Clark unit hydrograph basin models and the hypothetical meteorologic
models. The following table shows the basin model, meteorologic model, and simulation run names for the Clark
unit hydrograph simulations.
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 590
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The Clark unit hydrograph parameters were adjusted for each hypothetical precipitation event to find a
combination of time of concentration and storage coefficient that reproduced the 2D transform results. The
following figure shows the results from the Clark simulation where the time of concentration and storage coefficient
were adjusted to match results from the 2D simulation that had 1 inch of precipitation. A time of concentration of 3
hours and a storage coefficient of 3.5 hours was found to reproduce results from the 2D Diffusion Wave simulation.
The following figure shows the results from the Clark simulation where the time of concentration and storage
coefficient were adjusted to match results from the 2D simulation that had 0.5 inch of precipitation. A time of
concentration of 4.5 hours and a storage coefficient of 9.0 hours was found to reproduce results from the 2D
Diffusion Wave simulation.
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 591
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The following table contains results from the five Clark unit hydrograph simulations. Notice how the time of
concentration and storage coefficient values decrease as the precipitation intensity increases.
Precipitation TC R
0.25 7 15
0.5 4.5 9
1 3 3.5
2 2 1.75
4 1.5 0.9
When the Clark, Snyder, and SCS synthetic unit hydrographs are selected for a subbasin element, HEC-HMS will
compute the unit hydrograph and save the unit hydrograph as a paired data record to the simulation DSS file. The
following figure shows the records in the Clark_1.dss file. The unit hydrograph is saved to a record with a C-part
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 592
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The following figure shows all five unit hydrographs from the five hypothetical precipitation simulations (the
volume of runoff from each of the hydrographs shown below is 1-inch). Notice the non-linear behavior in the unit
hydrographs. The time of peak flow become much shorter and the shape of the unit hydrograph becomes much
more peaked, runoff is concentrated over a shorter period. In summary, these Clark unit hydrographs were created
for specific precipitation intensity values and were developed by adjusting Clark unit hydrograph parameters to
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 593
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 594
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The following table shows the percentage curves for time of concentration and storage coefficient. The percentages
are keyed from values for the index precipitation rate of 1 inches/hour. Percentage curves were created for both
time of concentration and storage coefficient. The values shown below were entered into percentage paired data
tables. The following figure shows the Concentration percentage curve, it is named TC Curve in the example
project.
2 200% 2 67 1.75 50
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 595
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 596
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Why does the computed hydrograph not match the observed hydrograph? A calibrated 2D Diffusion Wave transform
method was used to develop relationships between excess precipitation and the Clark time of concentration and
storage coefficient parameters. However, the approach used does not mimic the variable nature of precipitation
from natural flood events, like the event being modeled on December 22, 1982. The figure below shows the basin
average precipitation for the December 22, 1982 event. The precipitation rate does not approach the 1 inch/hour
rate, but there are six consecutive hours where the precipitation rate exceeds 0.2 inch/hour. During this extended
period of precipitation, rainfall is falling onto a wet watershed. There is water flowing across the land surface in
streams as precipitation is fall. This natural event is not the same as the hypothetical events used to derive the
variable parameters; the event and results remind us that the time-invariance assumption used with the
application of unit hydrographs has limitation. The main point here is that a unit of precipitation falling onto a
watershed will respond differently with different antecedent conditions.
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 597
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The variable time of concentration and storage coefficient percentage curves were adjusted to improve the model
results for the 1982 event. In addition, the constant loss rate was reduced from 0.25 inches/hour from the 2D
simulation to 0.16 inches/hour for the variable Clark simulation. The following table shows the updated
concentration and storage curves.
200% 67 50
400% 50 26
The following figure shows the updated results after adjusting the variable time of concentration and storage
coefficient curves and loss rate. The precipitation rates for the December 22, 1982 event only allow calibration of
the lower end of the variable time of concentration and storage coefficient curves since the precipitation rate is less
than 0.5 inch/hour. Events with higher precipitation intensities would be needed to improve the upper end of the
variable parameter curves.
The approach and tools in this example provide a method for using unit hydrographs for a range of flood events.
The most useful application for the variable Clark unit hydrograph method could be for extreme events (for those
synthetic flood events that are larger than typical calibration data). The linearity assumption with unit hydrographs
often leads modelers to arbitrarily adjust unit hydrograph parameters when applying models for simulating
extreme floods. In the Coyote Creek example, the 2D Diffusion Wave simulations show the time of concentration
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 598
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
decreases from 4.5 hours for an excess precipitation rate of 0.5 inch/hour to 2 hours for a excess precipitation rate
of 2 inches/hour. The storage coefficient has a much more dramatic reduction (from 9 hours to 1.75 hours), which
translates to a much faster responding hydrograph at higher precipitation rates. The information gathered from the
2D Diffusion Wave simulation can be used to inform the variable Clark unit hydrograph model which could then be
used to quickly model the subbasin response to extreme events.
Creating Variable Clark Transform Method Parameters using the 2D Diffusion Wave Transform Method – 599
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
This tutorial will guide the user through creating a gridded precipitation boundary condition using the Vortex
Importer utility.
The latest Vortex distribution can be downloaded here142. Once downloaded, the distribution should be unzipped
locally on the user's machine. Alternatively, you can launch the gridded data import wizard in HEC-HMS by selecting
File | Import | Gridded Data.
A single day of multi-radar multi-sensor (MRMS) quantitative precipitation estimate (QPE) data has been provided
in the following attachment:
A clipping extent shapefile for the punxsutawney watershed has been provided in the following attachment:
141 https://github.com/HydrologicEngineeringCenter/Vortex
142 https://github.com/HydrologicEngineeringCenter/Vortex/releases
As previously mentioned, there is an abundance of gridded earth science data. The Vortex Importer utility was
created with a generalized framework to read in data from a variety of sources. A table of common gridded data
sources is provided at the bottom of this page(see page 604).
Data can be downloaded manually, but for large datasets it is more efficient to download data using a script. A
repository of community-maintained data retrieval scripts can be accessed on HEC's GitHub page143. Users should
be comfortable with programming before attempting to automate downloads with a script. For those comfortable
with scripting downloads, please consider contributing to the project.
To begin the import process, launch the Vortex Importer utility. The Vortex Importer utility can be launched by
browsing to the <Vortex root>/bin directory and double-clicking importer.exe or importer.bat.
Note: The *.bat launcher will open with a command window that provides more output than would be otherwise
realized during the processing phase. Either launcher will achieve the same end.
On Step 1 of the wizard, select files for import. Use the Open button in the top right to multi-select gridded data
files.
143 https://github.com/HydrologicEngineeringCenter/data-retrieval-scripts
After clicking next, the Importer should go through a processing phase before arriving at import complete.
Close the Importer wizard. Gridded records can now be viewed in HEC-DSSVue.
Note: It is recommended that you use HEC-DSSVue v.3.* for viewing gridded data. Development versions of HEC-
DSSVue can be downloaded here144.
The gridded data can now be referenced in an HEC-HMS grid data object, and used within an HEC-HMS
meteorologic model.
30.1.1 Historical
Name Vari Source Doma Period Ti Res Fo Archive
able in of me olut r
s Record ste ion m
p at
144 https://www.hec.usace.army.mil/software/hec-dssvue/downloads/dev/
145 mailto:hec.hms@usace.army.mil
146 http://mtarchive.geol.iastate.edu
147 https://www.ncdc.noaa.gov/contact
NOAA National Centers Preci NOAA NCEP CONU 2002- 1 4 G NCAR / EOL151
for Environmental pitati S present hr km RI
Prediction (NCEP) Stage on B
IV Data
148 http://mtarchive.geol.iastate.edu
149 https://prism.oregonstate.edu/
150 https://www.ncdc.noaa.gov/contact
151 https://data.eol.ucar.edu/cgi-bin/codiac/fgr_form/id=21.093
152 https://hydrology.nws.noaa.gov/aorc-historic/
153 https://disc.gsfc.nasa.gov/
154 https://github.com/HydrologicEngineeringCenter/nasa_ges_disc_downloader
155 https://www.ncei.noaa.gov/data/cmorph-high-resolution-global-precipitation-estimates/access/
156 https://daymet.ornl.gov/getdata
Livneh CONUS Near- Preci Livneh B., E.A. CONU 1915-2 1 0.06 N FTP157
Surface Gridded pitati Rosenberg, C. S 011 day 25 et
Meteorological and on, Lin, B. Nijssen, V. deg C
Derived Tem Mishra, K.M. ree D
Hydrometeorological perat Andreadis, E.P. F
Data ure, Maurer, and D.P.
et al. Lettenmaier,
2013: A Long-
Term
Hydrologically
Based Dataset of
Land Surface
Fluxes and States
for the
Conterminous
United States:
Update and
Extensions,
Journal of
Climate, 26,
9384–9392.
157 https://www.psl.noaa.gov/data/gridded/data.livneh.html
158 https://nsidc.org/data/g02158
159 https://climate.arizona.edu/data/UA_SWE/
160 https://www.psl.noaa.gov/data/gridded/tables/precipitation.html
161 https://climatedataguide.ucar.edu/climate-data/precipitation-data-sets-overview-comparison-table
31.2 Introduction
HEC-HMS includes several baseflow methods. The linear reservoir baseflow method can be used to model both the
interflow and groundwater flow components that make up subsurface flow. The linear reservoir baseflow method is
unique in HEC-HMS because it is the only baseflow method that guarantees mass conservation. Computed
precipitation losses, computed by the loss methods, are passed to the linear reservoir baseflow method. Only
losses when the soil is in a saturated state are passed to the linear reservoir baseflow method for the Soil Moisture
Accounting, Layered Green and Ampt, and the Deficit and Constant loss methods. All losses from the other loss
methods are passed to the linear reservoir baseflow method.
The figure below shows the linear reservoir baseflow Component Editor within HEC-HMS. The user can choose the
number of linear reservoir layers, up to three layers can be selected. For each linear reservoir layer, the user must
162 file:///C:/confluence/display/HMSGUIDES/Creating+a+Simple+2D+Flow+Model+within+HEC-HMS
163 https://www.hec.usace.army.mil/confluence/download/attachments/46009604/CoyoteCreek.zip?
api=v2&modificationDate=1620409559703&version=2
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 609
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
define the initial condition, the fraction, the linear reservoir coefficient, and the number of steps.
Typically, I like to use two linear reservoir layers. I use groundwater layer 1 to represent interflow, which has
response times 2-5 times longer than surface flow. Interflow represent water that infiltrates, travels down slope in
the soil, within the unsaturated zone, and then makes its way back onto the land surface. After reaching the land
surface, the interflow water travels in the stream network to the subbasin outlet. The figure below shows possible
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 610
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
I use the groundwater layer 2 to represent groundwater flow, which has response times that are much longer than
surface flow. Groundwater flow represents water that percolates to the groundwater table, the saturated zone, and
then flows back onto the land surface within channels and streams. The figure below shows a possible flowpath for
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 611
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
groundwater flow. Notice that the infiltrated water spends much longer in the soil than for interflow.
The fraction parameter was made available starting in HEC-HMS Version 4.4 and is a way to distributed water
passed to baseflow from the loss method. The summation of the fraction values, from all groundwater layers
cannot exceed 1. The total fraction amount can be less than 1. For example, a groundwater 1 fraction of 0.25 and a
groundwater 2 fraction of 0.1 means 25 percent of the percolation losses from the Deficit and Constant loss method
is passed to the groundwater 1 layer, 10 percent of the percolation losses from the Deficit and Constant loss method
is passed to the groundwater 2 layer, and 65 percent of the percolation losses from the Deficit and Constant loss
method are lost to a deep aquifer and do not run off into the stream network.
The linear reservoir coefficient parameter can be used to control how quickly water is held in the groundwater
layer. A small linear reservoir coefficient will result in water leaving the linear reservoir more quickly than a larger
coefficient. The figure below shows results from three simulations where the linear reservoir coefficient was varied.
Notice how the magnitude of the baseflow hydrograph is decreased as the coefficient is increased. A larger
coefficient (for the same size watershed) means infiltrated water spends more time in the soil before flowing back
onto the land surface. This example also illustrates that the linear reservoir coefficient does not have much impact
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 612
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The number of steps parameter is a way to control the timing of the peak flow in the baseflow hydrograph. One step
means there is one linear reservoir within the groundwater layer. Two steps means there are two linear reservoirs in
series in the groundwater layer. Water is routed through one linear reservoir and then it is routed through the
second linear reservoir. Each linear reservoir has the same coefficient. The figure below shows results from three
simulations where the number of steps was varied. Not only is the magnitude of the baseflow hydrograph impacted
by the number of steps, but the timing of the baseflow hydrograph is shifted as well.
The release of HEC-HMS 4.7 provided a new a new subbasin transform method, the 2D diffusion wave transform
method. The 2D diffusion wave transform method is a true cell to cell routing model, it is the same routing model
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 613
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
(code) in HEC-RAS. After preliminary application of the 2D diffusion wave transform method to watershed scale
applications, the HEC-HMS team identified the need to provide a 2D subsurface routing option that compliments
the 2D diffusion wave surface routing option. This need for a integrated 2D surface-subsurface model can be
illustrated using the figure below. Within HEC-HMS, baseflow from the linear reservoir baseflow method is only
added to the hydrograph at the outlet of the 2D area (subbasin) and is represented by the brown dashed line in the
model results shown below. Infiltrated water from all grid cells, regardless of their location in the 2D area, is routed
through a linear reservoir and added to the total flow from the 2D area. We found this approach could provide
adequate results at the subbasin outlet, the combination of routed surface flow and baseflow at the outlet could
approximate observed flow (compare the solid blue line, computed flow, and the black line, observed flow, in the
figure below). However, the 2D surface flow at interior cells would be biased low because in reality a portion of the
infiltrated water does flow back onto the land surface throughout the entire watershed, not at some arbitrary
location like a subbasin outlet. Notice that the simulated baseflow hydrograph adds about 20 percent of the flow
magnitude to the 2D diffusion wave hydrograph at the subbasin outlet. This means that the flow/stage results for
cells within the 2D mesh are lower than would be observed in this model application. It is likely that modelers will
have to reduce surface roughness values within the 2D model domain to offset missing groundwater flow when
calibrating a model with this type of configuration.
Until the HEC-HMS and HEC-RAS teams work with development partners to design and fund a true integrated 2D
surface-subsurface solution, the HEC-HMS team made a modification to the linear reservoir baseflow method that
allows the modeler to identify whether the Linear Reservoir groundwater layer is treated as interflow or baseflow. If
the user chooses baseflow, then the routed subsurface water is added to the total flow at the subbasin outlet
(outlet(s) of the 2D area). If the user chooses interflow, then the subsurface water is only routed within the 2D cell,
where the infiltration happened. The routed interflow is added to the cell's surface flow and routed on the 2D
surface to downstream grid cells (where it can be infiltrated again). The new interflow linear reservoir baseflow
option can only be activated when the 2D Diffusion Wave transform option is selected.
It is understood this approach is conceptual and the major assumption is that all interflow leaves the cell on the
surface where the infiltration happens. However, this approach does allow for improved model results for interior
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 614
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
grid cells within the 2D area. Also, the modeler will have to adjust surface roughness values when changing to the
new interflow baseflow routing option because more water will be in the channel network within the 2D area.
This tutorial shows how the new interflow option can be utilized with the 2D diffusion wave transform method, how
to parameterize the linear reservoir baseflow parameters, and then shows additional model calibration that is
needed when using the new interflow linear reservoir baseflow option.
Terrain data was downloaded from the USGS National Map Viewer. The terrain data was re-projected and the
vertical units were converted to feet. Published streams from the National Hydrograph Dataset (NHD) were burned
into the terrain as well. The terrain data was used to delineate subbasin and stream elements within HEC-HMS. The
terrain data was also used by the HEC-RAS 2D preprocessor to extract hydraulic properties for each cell and cell face
in the 400-ft unstructured mesh. The following figure shows the watershed on top of an aerial image; notice the
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 615
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Landuse data was gathered from the National Land Cover Database (NLCD) 2016. As shown below, land uses within
the watershed are mixed forest, grassland, and shrub/scrub. Soil data from the SSURGO database was used to
create raster datasets of soil properties, like saturated hydraulic conductivity. The soil property datasets were used
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 616
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Precipitation and flow information were provided by staff at the California Department of Water Resources. Both
the precipitation and flow data were recorded at 15-minute intervals. The follow figure shows the precipitation
gage locations. The flow gage is located at the outlet of the watershed area. HEC-GageInterp was used to
interpolate the point rain gage data to create precipitation grid sets for six storm events (a 2000-meter grid spacing
was used for the interpolation). The PRISM 30-year average precipitation dataset was used as a bias grid in the
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 617
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
interpolation process.
The unstructured discretization method was used to create a 2D mesh that was used with the 2D Diffusion Wave
transform method. Currently, the modeler must create the 2D mesh and compute the mesh properties in HEC-RAS
before referencing the mesh in HEC-HMS. A portion of a 400-ft unstructured mesh is shown on top of the Coyote
Creek watershed. The unstructured mesh option can be used with gridded or lumped loss methods and with unit
hydrograph, ModClark, and 2D Diffusion Wave transform methods. Gridded or non-gridded precipitation can be
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 618
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
used with an unstructured mesh. There are 18,512 cells within the unstructured mesh.
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 619
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
GW 1 Fraction 0.25
GW 1 Coefficient (HR) 10
GW 2 Fraction 0.1
GW 2 Coefficient (HR) 40
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 620
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
GW 3 Fraction NA
GW 3 Coefficient (HR) NA
The following figure shows results at the outlet of the 2D area along with the observed flow for the 1982_2D
simulation. Below, the black line is the observed flow at the outlet of the subbasin and the smaller dashed blue line
is the total computed flow at the outlet. The total computed flow combines surface flow, solid blue line, and
baseflow, longer dashed line. Notice how the 2D surface runoff does not make it to the outlet of the 2D area until
hour 1400 on December 22. If you look at the excess precipitation time series, you will see excess precipitation was
computed starting at hour 1700 on December 21. Starting the simulation with a dry 2D area causes a delay in
surface runoff, but the linear reservoir baseflow option masks this result by routing the infiltrated water to the
subbasin outlet (baseflow is making it to the outlet before the surface runoff). Overall, the magnitude of the total
flow matches well with the observed peak. However, as discussed above, if you extract results for grid cells within
the 2D area, then the computed flow and stage would be biased low throughout the 2D area.
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 621
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The following table shows the constant loss rate and baseflow parameter values for this model configuration. The
groundwater 1 layer had a much smaller coefficient than the groundwater 2 layer, which results in flow leaving the
groundwater 1 layer more quickly than the groundwater 2 layer. Notice how the storage coefficient values are
significantly smaller than the 2D basin model. The coefficients were reduced because interflow is being modeled
within each 400ft grid cell, not for the entire 108.55 square mile watershed. The surface roughness values were the
same as the 2D basin model. This basin model is used within the 1982_2D_newBF simulation.
GW 1 Fraction 0.1
GW 2 Fraction 0.1
GW 2 Coefficient (HR) 10
GW 3 Fraction 0.1
GW 3 Coefficient (HR) 20
The following figure shows results at the outlet of the 2D area along with the observed flow for the 1982_2D_newBF
simulation. Adding the interflow baseflow output from each grid cell, along with surface runoff, and routing the
surface flow to the downstream grid cells does change results significantly. The simulated peak flow is larger than
in the 1982_2D simulation. This increase in magnitude is due to having more water in the channel network, and
faster velocities, throughout the entire 2D area. The timing of the peak flow was 25 minutes earlier in the
1982_2D_newBF simulation than the 1982_2D simulation. The surface roughness, loss rate, and baseflow
parameter would need adjustment to improve the model results (see the 2D_newbf_Calib section below).
Another note, the delay in runoff from the 2D surface model is more pronounced in this simulation since subsurface
water being routed directly to the subbasin outlet is no longer masking the routing results from the surface 2D
diffusion wave model.
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 622
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
800
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 623
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
GW 1 Fraction 0.1
GW 2 Fraction 0.1
GW 2 Coefficient (HR) 2
GW 3 Fraction 0.1
GW 3 Coefficient (HR) 20
The following figure shows results at the outlet of the 2D area along with the observed flow for the
1982_2D_newBF_Calib simulation. Calibrating the model does result in a better fit to the peak flow, the time of the
peak flow is slightly lagged, due to larger surface roughness values. The more important point with results from this
simulation is that output from 2D surface flow better represents observations at the subbasin outlet. There is some
integration of surface flow and the interflow component of subsurface flow occurring throughout the 2D area.
Therefore, flow and stage at interior locations will be more accurate than the original simulation where all
subsurface flow was added to surface runoff at the outlet of the subbasin. If you only need results at the outlet of
the 2D area, then utilizing the interflow linear reservoir baseflow option might not provide a benefit.
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 624
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The delay in surface runoff is still pronounced in this simulation. The following section shows a method for
improving runoff at the beginning of the storm event.
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 625
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
GW 1 Fraction 0.1
GW 2 Fraction 0.1
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 626
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
GW 2 Coefficient (HR) 2
GW 3 Fraction 0.1
GW 3 Coefficient (HR) 20
The following figure shows results at the outlet of the 2D area along with the observed flow for the
1982_2D_newBF_Calib_preEvent simulation. All basin model parameters were kept the same as the
2D_newBF_calib basin model except the constant loss rate was increased to 0.258 inches per hour. Adding a
warmup period that included artificial precipitation filled streams and channels in the 2D area so that the surface
response to the actual rainfall on December 21 and 22 was more realistic.
Applying the New Linear Reservoir Baseflow Option with 2D Surface Flow – 627
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
32.1 Overview
In this tutorial you will modify a terrain dataset by using the Terrain Reconditioning tool to Burn Streams and Build
Walls. You would use these two options when you want to enforce known information, like the locations of streams
from the National Hydrography Dataset, into the terrain dataset. An example would be if you had subbasin
boundaries and a river reach shapefile from an existing study and you wanted to make sure the automated
delineation tools in HEC-HMS recreated the subbasin and reaches from the existing study.
The figure below shows two subbasin shapefiles. The red polygons were generated by someone digitizing subbasin
boundaries using topographic maps. The black polygons were generated by an automated delineation tool within a
GIS. There are similarities and differences in the subbasin delineations across the watershed. The automated
delineation could be improved by modifying the terrain with known information, like where subbasin boundaries
and stream locations exist.
The Terrain Reconditioning options are described in the HEC-HMS User's Manual, HEC-HMS GIS Menu164. The
following figure illustrates how the burn streams tool works. The black circle represents a stream polyline and the
brown line represents the land surface from the terrain. I set the buffer to 2 cells, the smooth drop to 5 feet, and the
sharp drop to 1 foot. The profile on the left is the original terrain and the figure on the right is the reconditioned
terrain. Finally, the blue squares below the terrain profiles represent the grid cells within the terrain. In this
example, the burning algorithm will buffer the stream line by 2 grid cells on each side. The algorithm will lower the
elevation of the grid cells directly beneath the stream line by 5 feet and then gradually lower the elevation for the
other cells within the buffer so that there is a smooth transition to the existing elevation of the grid cells adjacent to
the buffered area. The sharp drop is an additional drop that is added to the smooth drop. In this example, the grid
cells directly beneath the stream line were lowered an additional 1 foot, for a total decrease in elevation of 6 feet for
164 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsum/latest/geographic-information/gis-menu
The tutorial is divided into two sections. The first section shows how to use the burn streams option only and the
second sections shows how to use the build walls and burn streams options together. This tutorial uses the same
watershed and datasets as the ones described here: Introduction to Creating an HEC-HMS Model from Scratch using
HEC-HMS GIS Tools165.
5 https://www.hec.usace.army.mil/confluence/hmsdocs/hmsguides/creating-an-hec-hms-model-from-scratch-using-hec-hms-gis-tools/
16
introduction-to-creating-an-hec-hms-model-from-scratch-using-hec-hms-gis-tools
166 https://www.hec.usace.army.mil/confluence/download/attachments/48136412/startPunxsutawney.zip?
api=v2&modificationDate=1611915228879&version=1
5. In the next step, select the NHD_River_albersft shapefile, it is in the GIS Data folder in the project directory (
…\Punxsutawney\maps\GIS Data). I set the buffer to 10 cells, a smooth drop height of 50 feet, and a sharp
drop height of 20 feet. The amount of terrain reconditioning might be excessive for this example, I want to
make sure the terrain modification was visible for illustration. In many cases, editing the terrain is an
iterative process. The goal is to reproduce a known drainage network (subbasin and reach delineation);
therefore, you might have to try different options and datasets to get appropriate results.
6. Click the Finish button. A new terrain dataset is created and is referred to as the "Reconditioned Terrain"
dataset, see the Map Layers editor in the figure below. You can identify cell values in the Basin Model map
by selecting a map layers in the Map Layers editor list and then hovering your mouse on top of the
reconditioned terrain in the map.
7. Both figures below show the reconditioned terrain. The figure on the right has the NHD_River_albersft
stream shapefile turned on to help see which grid cells would have been lowered in the burn streams
process. If the amount of burning in is large, then you will be able to visually see the cell elevations below
and within the buffer tolerance of the stream shapefile have been lowered.
The figures below show the elevation for a grid cell that fell along the stream shapefile, the location is
approximately at the confluence of the two reach lines. The figure on the left shows the original terrain, the
elevation for this location was 1370.5 feet. The figure on the right shows the reconditioned terrain and the
elevation was 1300.5 feet. The total elevation drop for this cell, which is collocated with the stream line was
70 feet. The buffer was set to 10 cells around the stream shapefile and the smooth drop was 50 feet. The
buffer cells farthest out from the stream line had the smallest adjustment while the cells collocated with the
stream line were adjusted by a total of 50 feet. For those grid cells collocated with the stream line, the sharp
drop of 20 feet was added to the smooth drop of 50 feet, for a total elevation drop of 70 feet.
Each time you run the Terrain Preprocessing tool, the program will automatically use the original terrain
dataset. The program does not continue to edit the reconditioned terrain dataset. After you modify the
terrain, the next step is to run the Preprocess Sinks step before running the remaining tools to delineate
subbasin and reach elements.
32.3 Using the Build Walls and Burn Streams Options in Sequence
1. Start HEC-HMS (version 4.8 and newer) and open the Punxsutawney project.
2. Open the Mahoning Creek basin model. A terrain dataset has been added to the basin model. Three
shapefiles have been added and displayed in the basin model map. The NHD_River_albersft shapefile is a
portion of the stream layer that was extracted from the National Hydrography Dataset. The streams were re-
projected to the same projection as the terrain model. The subbasins shapefile represents subbasin
polygons from an existing model of the watershed. The subbasin shapefile was also re-projected to the same
projection as the terrain model.
5. In the next step, select the NHD_River_albersft shapefile, it is in the GIS Data folder in the project directory (
…\Punxsutawney\maps\GIS Data). I set the buffer to 10 cells, a smooth drop height of 50 feet, and a sharp
6. Click the Finish button. A new terrain dataset is created and is referred to as the "Reconditioned Terrain"
dataset, see the Map Layers editor in the figure below. You can identify cell values in the Basin Model map
by selecting a map layers in the Map Layers editor list and then hovering your mouse on top of the
reconditioned terrain in the map.
7. Zoom into the area the red arrow is pointing to in the figure below. This area is at the confluence of two
rivers (Stump Creek and East Branch Mahoning Creek).
8. Turn on display of the subbasins and NHD_River_albersft shapefiles and make sure the reconditioned
terrain is displayed as well. As shown below, you can see the terrain cells where the elevation was raised
based on the subbasin layer's polygon outline (see the "Wall Added" area in the figure below). You can also
see the terrain cells that were lowered because of their proximity to the stream layer's polyline (see the
"Stream Burned into Terrain" area in the figure below). You can also see where the stream polylines acted as
a breach in the walls (see the "Breach in Wall due to Stream Buffer" area in the figure below).
It is suggested when using the build walls option that you always specify a stream layer to breach each of the
subbasin polygons. You do not have to burn in the streams (smooth and sharp drops are not required). If you do not
specify a stream polyline layer when building walls, the Preprocess Sinks tool will fill in a large amount of the terrain
in order to create a hydrologically corrected DEM. As an example, I ran the steps shown above, but did not choose a
stream layer to burn in or to ensure that a breach was added to the walls. Then I ran the Preprocess Sinks option.
The following figure shows the Sink Locations that were filled in by the preprocess sinks tool. Notice a significant
amount of the reconditioned terrain had to be filled in for water to flow across the terrain.
Each time you run the Terrain Preprocessing tool, the program will automatically use the original terrain
dataset. The program does not continue to edit the reconditioned terrain dataset. After you modify the
terrain, the next step is to run the Preprocess Sinks step before running the remaining tools to delineate
subbasin and reach elements.
Background: For this workshop, measurements at gages within the lower Columbia River watershed will be used.
Figure 1 shows the Columbia River watershed, and the three District boundaries within the Columbia River
watershed. The watershed was divided into 12 basins and separate HEC-HMS models were developed for each of
the basins. One of the initial steps in the HEC-HMS model development was calibration of the HEC-HMS
temperature index model to SNOTEL gages. Figure 2 shows the SNOTEL gages located within the Mainstem
Columbia River watershed. The gages were selected to span the watershed in both location and elevation.
Tasks: The following major tasks will serve as an outline for the workshop:
1. Download precipitation, temperature, and snow water equivalent data for SNOTEL gages within the
Mainstem Columbia River watershed.
2. Create an HEC-HMS project.
3. Add time-series data to the HEC-HMS project.
4. Create a Basin Model that has subbasin elements representing three SNOTEL gages.
5. Create a Meteorologic Model and parameterize the precipitation and snowmelt components.
6. Create a simulation for the 1997 and 1998 water years and evaluate results.
7. Adjust the Temperature Index parameters to improve model results for the three SNOTEL gages.
Initial project files: maps.zip167
167 https://www.hec.usace.army.mil/confluence/download/attachments/54661347/maps.zip?
api=v2&modificationDate=1616596532388&version=1
Columbia River watershed. These gages are named Arbuckle Mtn, Red Hill, and Tipton, and they are located in
Oregon.
1. Open HEC-DSSVue and create a new DSS6 file.
2. Go to Data Entry Import and select SnoTel. The SnoTel Download editor will open.
3. Select the Daily data type, enter a start date of 01Oct1996 and an end date of 30Sep1998, and then select
the button to Get Site List By State. Choose Oregon. The table should fill in at the bottom with all gages in
the state of Oregon.
4. Check the boxes to download data for the Arbuckle Mtn, Red Hill, and Tipton gages as shown in Figure 3.
5. Click the Import to DSS button at the bottom of the editor. You should get a message that 22 datasets from
3 of 3 gages was downloaded. Close the SnoTel Download editor.
6. Examine the downloaded records, you should notice SWE, Temperature, and Precip-Inc-Adjusted records for
each of the three gages. Figure 4 shows observed SWE for the three gages. The Precip-Inc-Adjusted dataset
will be used as the precipitation in HEC-HMS. This dataset was created by adjusting measured precipitation
to match observed SWE volumes.
7. Some changes will need to be made to the SnoTel data. With the data opened in HEC-DSSVue, select the
three SWE gages and select the Tabulate option. The data for the selected records will open up in a separate
window. Edit the data by selecting the Allow Editing option found under the Edit tab. While in the edit
session, change the UNITS for all three gages to IN and the Type to INST-VAL. Save the edits and close the
window. Select the Temperature-Air-Avg option for all three gages and tabulate the data. There will be
missing data for the gages. 24Oct97 will be missing for the Arbuckle MTN gage and 21Mar98 will be missing
for ALL three gages. Edit the data and fill in the missing values by interpolating between the data points
around the missing values. Save the edits, close the window, and close HEC-DSSVue.
Figure 4. Observed SWE data for the Arbuckle Mtn, Red Hill, and Tipton gages.
Figure 8. The Specified Hyetograph component editor with precipitation gages selected for each subbasin.
1. Add the ATI-Meltrate and ATI-Coldrate paired data to the project using the Paired Data Manager under the
Components tab. HEC-HMS 4.4 allows you to define a separate meltrate and coldrate function for each
subbasin; however, create only one curve that will be selected by all three subbasins initially. You can create
separate paired data curves for each subbasin if needed during calibration. Enter the values in Table 1 for
the ATI-Meltrate curve and the values in Table 2 for the ATI-Coldrate curve.
0 0.04
100 0.08
500 0.12
0 0.02
500 0.03
1000 0.03
1. The next step is to enter temperature index boundary condition information and parameters for each
subbasin. Click on one of the subbasin nodes (within the meteorologic model portion of the watershed
explorer) to open the Atmospheric Variables component editor as shown in Figure 9. Select the appropriate
temperature gage for the subbasin. Then, select the Temperature Index node, below the subbasin node, to
open the component editor shown in Figure 10. Enter a lapse rate of 0 DEG F / 1000 FT, a lapse rate is not
required since the temperature and SWE measurements were made at the same gage site. Enter the
temperature index parameters using values in Table 3. Repeat this step for the other two subbasins.
Parameter Value
PX Temp (Deg F) 32
1. An elevation band must be added to each of the subbasins. Right click on the subbasin's Temperature
Index node in the watershed explorer. A pop-up menu should open, click the option to add an elevation
band. Open the elevation band's component editor by clicking on Band 1 in the watershed explorer. With a
single elevation band per subbasin, you need to set the Percent (%) parameter to a value of 100. Many of the
initial snow model parameters required will be set to zero because the simulation will start at the beginning
of the snow season, when there is no snowpack established. The elevation for the band is the same elevation
defined for the temperature gage, 5000 ft. Table 4 contains the information to enter for elevation Band 1.
Create an elevation band for all three subbasins and enter the information in Table 4.
Table 4. Initial snow conditions for elevation Band 1, assuming no initial snowpack.
Parameter Value
33.6
Task 6: Create a simulation for the 1997 and 1998 water years and
evaluate results.
1. Create a Control Specifications named Snow Time Window with a start date of 02Oct1996 at time 00:00
and an end date of 01Oct1998 at time 00:00. Set the Time Interval to 1 Day (the SNOTEL data is at a 1-Day
time step).
2. Create a simulation run named SWE Calibration that combines the basin model, meteorologic model, and
control specifications.
3. Run the simulation. The simulation should take about two seconds to complete.
4. Computed and observe SWE results can be accessed from the Results tab of the watershed explorer. Expand
the watershed explorer for the Red Hill subbasin. Select the Snow Water Equivalent result first, hold down
the control key and then select the Observed SWE result second. Then click the View Graph button to open
a larger plot of the results. Figure 11 shows the results at the Red Hill gage using the parameters defined
above. Figure 12 shows the summary table with metrics describing how well/poor the model performed.
Results for the Red Hill Gage should look similar to the other subbasins, the simulated results show that the rate of
melt is slower than the observed melt rate in the spring-summer time frame. The magnitude of peak SWE is close
between the observed and simulated time-series.
Figure 11. Modeled and observed SWE for the Red Hill gage.
Using the initial parameterization, the timing of when the snowpack completely disappears is delayed in the model
results. There are a few factors that could contribute to this. Cold content and liquid water content both act to
delay the onset of melt, though typically the impact is small. The most likely factor contributing to the delay in melt
is the dry meltrate. Increasing the dry meltrate values, especially at the lower end (lower ATI values) was found to
give much better results.
Make note of your NSE score for each of the SnoTel gages. The team with the highest NSE score per gage will
be awarded a prize. Other metrics that are important with SWE are date of peak SWE and date when the SWE
melts off.
Question 2: What is your final parameter set per subbasin/gage? Which parameters were more influential for
simulating SWE and matching observations?
Answer
Besides the ATI-Meltrate, another sensitive parameter was the PX temperature. The PX temperature was increased
to 33 degrees F for the Red Hill and Tipton SNOTEL gages. The increase in the PX temperature resulted in more
accumulation of snow during the season. The ATI-Coldrate paired data curve was modified to reduce the amount of
delay in melt when temperature goes above the Base Temperature and when melt begins.
Question 3: Create a plot for the Arbuckle MTN gage that includes Computed SWE, Observed SWE, and
Temperature. Zoom into the time frame between January 20 and February 15 - 1997. Do the observed SWE and
Temperature datasets seem reasonable in this time frame?
Answer
The following figure shows the Observed Temperature and SWE for January 1997, look inside the rectangle. Notice
that the observed SWE decreases from 20 inches to 15 inches even though the temperature is below freezing. There
could be other processes that are impacting the snowpack, like sublimation. The HEC-HMS model will not be able
to track the trend in observed SWE during this period if other processing not included in HEC-HMS are impacting the
snowpack. Another possibility are errors in the gage measurements.
Question 4: Describe some of the limitations when using daily accumulated precipitation and averaged
temperature data as boundary conditions.
Answer
Daily data does not include the diurnal pattern in temperature, when temperatures are colder at night and warmer
during the day. Cold temperatures at night can add cold content to the snowpack that offsets some melt during the
day. Using daily average temperature does not capture the diurnal temperature pattern, the model will not
simulate the increase/decrease in cold content due to the diurnal pattern in temperature. Precipitation can be
dynamic within an event. A storm can start out warmer, with mostly liquid precipitation. Then, the temperature
can drop as a cold front passes, and liquid precipitation turns into snow. Using daily accumulated precipitation
with a daily average temperature might not capture the dynamic nature of rain/snow during some storm events.
168 https://www.hec.usace.army.mil/confluence/download/attachments/54661347/MainstemColumbia.zip?
api=v2&modificationDate=1616596607946&version=2
Objectives. In this workshop, you will gain experience using the gridded temperature index snowmelt/
accumulation method within the John Day River watershed. You will refine initial temperature index parameter
estimates for both the deficit constant and soil moisture accounting loss methods.
The following major tasks will serve as an outline for the workshop:
1. Add temperature index parameters from the previous workshop to an existing HEC-HMS model for the John
Day River.
2. Compute the HEC-HMS model and compare the initial results vs observed results.
3. Refine the temperature index parameters to provide better agreement between computed and observed
SWE.
4. Refine the linear reservoir baseflow parameters to provide better agreement between computed and
observed streamflow. There will be a prize awarded to the team with the highest Nash-Sutcliffe Efficiency
score at the Monument gage!
An additional task is included at the end of this workshop, if time allows.
Spend approximately 25 minutes per task to accomplish all tasks in this workshop within the time allotted.
Initial Project Files: John_Day_River.zip169
Parameter Value
PX Temperature (deg F) 33
169 https://www.hec.usace.army.mil/confluence/download/attachments/54661374/John_Day_River.zip?
api=v2&modificationDate=1616598241901&version=3
Parameter Value
• Expand the Paired Data node in the tree and populate the "meltrate_WY2011" and "coldrate_WY2011"
curves with the data shown in Table 2 and Table 3. These best fit curves were determined within the
previous workshop
0 0.03
10 0.08
100 0.2
500 0.2
0 0
500 0.02
1000 0.02
34.2 Task 2: Compute the HEC-HMS model and compare the initial
results vs observed results.
Now that the HEC-HMS model has been parameterized, it's time to make the first simulation and see how the initial
computed results compare against observed data. When it comes to observed snow water equivalent (SWE) data,
there are several options for use in hydrologic models. These data sources include snow courses,
hydrometeorological gages, model output, and remotely sensed data, amongst others. Within this workshop,
basin-average SWE from the National Operational Hydrologic Remote Sensing Center (NOHRSC) Snow Data
Assimilation (SNODAS) model will be used to compare against the SWE computed from the HEC-HMS model. For
more information regarding the NOHRSC SNODAS model and output, visit this webpage: https://
www.nohrsc.noaa.gov/nsa/. The SNODAS output has already been selected for each subbasin within the John Day
River HEC-HMS model.
• Compute the model by selecting the "WY2011" simulation from the compute selection drop down menu and
clicking the exploding rain drop: .
• Once the simulation has finished, click on the Results tab and select any subbasin element.
• For each, subbasin, there will be multiple time series available for selection. Click on the "Snow Water
Equivalent" object, hold the Ctrl key, and click on the "Observed SWE" object. This will create a plot of
the two selected time series. The selected time series can be shown in a separate plot window by clicking the
plot button: .
• Additional time series can be added to an existing plot window by left-click and holding on the time series of
interest, dragging into the existing plot window, and releasing the left mouse button in the grey space
around the plot.
• Marker lines can be added on either the x- or y-axis of an existing plot window by selecting the pointer (
), right clicking in the viewport, and selecting one of the Add Marker options.
• Line styles/colors can be edited by clicking the Edit Graph Properties button ( ) and selecting various
plot options.
Question 1: How does computed SWE compare against observed SWE for the "JohnDayRv_S70" subbasin? Why is
the computed SWE different than the observed SWE? Try plotting precipitation, temperature, computed SWE, and
observed SWE in the same window (change the line styles of various curves to better visualize them). What
temperature index parameters should be changed to better match observed SWE?
Answer
A plot of the precipitation (blue line in the upper viewport), temperature (red line in the lower viewport), computed
SWE (blue line in the lower viewport), and observed SWE (dashed black line in the lower viewport) for the
“JohnDayRv_S70” subbasin are shown in the figure below.
The shape of the computed SWE curve largely matches the shape of the observed SWE curve. However, the
magnitude of the computed SWE curve is larger than the observed SWE curve. The computed SWE begins to
deviate from the observed SWE in the beginning of December 2011. During this time, the computed SWE increases
while the observed SWE decreases. Also, during this time, the basin-average temperature is very close to the initial
base and PX temperatures of 32 deg F and 33 deg F, respectively. Since the shape of the two curves are very similar
while the magnitudes are different, only minor changes to the base and PX temperatures are needed to match the
observed data to an acceptable degree.
Question 2: How does computed SWE compare against observed SWE for the following subbasins:
NFJohnDayRv_S30, JohnDayRv_S50, and JohnDayRv_S20? Which subbasin/region should receive more attention
when calibrating temperature index parameters?
Answer
The initial results (computed = blue lines, observed = dashed black lines) for NFJohnDayRv_S30, JohnDayRv_S50,
and JohnDayRv_S20 are shown in the figure below.
Within this figure, the computed SWE appears to not adequately match the observed SWE in any of the subbasins.
However, when the magnitude of the observed SWE in each subbasin is taken into account, it’s evident that the
differences within the NFJohnDayRv_S30 subbasin are much larger than the other two. The NFJohnDanRv_S30
subbasin has a peak observed SWE of approximately 18 inches compared against 1.4 and 1.0 inches of SWE within
the JohnDayRv_S50 and JohnDayRv_S20 subbasins, respectively. For these reasons, much more time should be
spent adjusting parameters within the NFJohnDanRv_S30 subbasin (and other similar subbasins) than the
JohnDayRv_S50 and JohnDayRv_S20 subbasins.
Question 3: Using the figure in the previous answer and any other information that you can gather, provide at least
two reasons why NFJohnDayRv_S30 subbasin may have larger SWE magnitudes than the other two subbasins
mentioned in Question 2.
Answer
1. Predominant weather patterns may be causing more precipitation to reach the NFJohnDayRv_S30 subbasin
than other subbasins within the modeling domain.
2. As weather systems move west to east, they will follow the terrain. As mountains are encountered, any air
mass must rise. If the air mass contains moisture, precipitation will likely fall. Temperatures also tend to
decrease as elevation increases. Since the NFJohnDayRv_S30 subbasin is one of the higher elevation
subbasins within the modeling domain, it likely is subject to orographic influences to a larger degree than
most other subbasins. Finally, when precipitation falls, the likelihood of it falling as snow is amplified due to
expected decreases in temperature.
Remember to think critically before you make any parameter changes. Ask yourself several questions:
• How different are the computed and observed results?
• What parameters should be changed to better align the computed and observed results?
• What magnitude change should be made (small or large)?
• Can I make the same change(s) within multiple subbasins before computing again to save myself time?
• Am I making things too complicated?
Question 4: The most common metric used to ascertain the acceptability of SWE calibration is comparing the peak
computed SWE against the peak observed SWE. List at least three additional ways in which SWE calibration can be
measured/determined.
Answer
Additional metrics that could be used to establish the acceptability of SWE calibration include:
1. Date of peak observed and computed SWE
2. Melt out date (i.e. date when SWE is completely melted) of observed and computed SWE
3. Nash-Sutcliffe Efficiency (NSE)
4. Ratio of the Root Mean Square Error to the Standard Deviation Ratio (RSR)
5. Percent Bias (PBIAS)
Question 5: How close is "close enough" when it comes to comparing the HEC-HMS model results against the
SNODAS data?
Answer
It is important to keep the goals of the study, accuracy of the computed data, and accuracy of the observed data in
mind when assessing the acceptability of calibration. For instance, the “observed” SWE data used within this
workshop is actually derived from another model. As such, inaccuracies within the SNODAS model processes and
other data used in assimilation are present. For this reason (and others), agreement to a few inches in magnitude
and a few days within the subbasins that develop the largest snowpack magnitudes is more than adequate. Also,
remember that calibrating SWE is not the same thing as calibrating both SWE and streamflow. There is a give-and-
take relationship between the two results and computed streamflow is a much more important parameter within
the vast majority of the studies that USACE undertakes.
Question 6: When calibrating SWE within the aforementioned subbasins, did you notice any parameters that could
be regionalized?
Answer
The law of parsimony states that simpler solutions to a problem tend to be more appropriate than complex
solutions. In layman’s terms: the solution to a problem should be as complicated as need be to adequately solve it
but no more complex. An extension of this law to the realm of hydrology would imply that simpler methods are
more appropriate than complex methods. Regionalization of model processes and parameters better adheres to
this law. Some important temperature index parameters that could be regionalized include:
• Range of PX and base temperature
• Base Temperature
• Wet Meltrate
• ATI-Meltrate
• ATI-Coldrate
All of these parameters can be regionalized, but may still vary a little bit throughout a region. Keep in mind that any
model is a simplification of the natural world.
While you are calibrating the precipitation-runoff processes to the aforementioned gages, take note of the Nash-
Sutcliffe Efficiency (NSE) score, which is a common metric used to ascertain goodness of fit. The Summary Results
contains this information for elements that have an observed time series linked to them. The Summary Results
table can be opened by selecting the element of interest and clicking on the summary table button near the top of
the frame . The NSE score location is shown in Figure 5. The team with the highest NSE score at the
Monument gage location will be awarded a prize. In the event of a tie, the tiebreaker will be the team with the
lowest absolute value Percent Bias.
Question 8: When calibrating streamflow within the aforementioned subbasins, did you notice any parameters that
could be regionalized?
Answer
There are a few parameters that could be regionalized. However, the most impactful to this analysis are GW1 and
GW2. These two parameters can be related to the watershed storage coefficient (Clark “R”), which itself was
determined from physically-measureable characteristics of the individual subbasins. The relationship between
these parameters should be maintained (to a reasonable degree) within a region.
34.5 Additional Task 1: Complete calibration of the entire John Day River
HEC-HMS model.
Once you've calibrated SWE and streamflow to a few subbasins/gage locations, try calibrating the entire model.
Specifically, streamflow at the Monument, ServiceCreek and McDonaldFerry gage locations are of prime
importance. Prioritize model performance at these locations.
Final Project Files: John_Day_River.zip170
170 https://www.hec.usace.army.mil/confluence/download/attachments/54661374/John_Day_River.zip?
api=v2&modificationDate=1616598241901&version=3
35.1 Introduction
The number of gridded data products that can be used for hydrologic modeling applications is increasing.
Modeling with gridded data is beneficial because gridded products are widely available and they have the ability to
better capture the temporal and spatial distribution of gridded parameters, like precipitation, across a watershed
when compared to gage measurements at a single point.
This tutorial demonstrates how Multi-Radar Multi-Sensor (MRMS) Quantitative Precipitation Estimate (QPE) gridded
precipitation data can be obtained, imported, and used within HEC-HMS. This tutorial will also show how
simulated runoff changes when different data is used in an existing model of the Cannon River watershed.
171 https://www.hec.usace.army.mil/confluence/download/attachments/54662087/CannonRev_v1_Start.zip?
api=v2&modificationDate=1617056144375&version=2
172 https://www.hec.usace.army.mil/confluence/download/attachments/54662087/Grib2_Clipper.zip?
api=v2&modificationDate=1617050110761&version=1
173 https://www.hec.usace.army.mil/confluence/download/attachments/54662087/Data_Grib.zip?
api=v2&modificationDate=1617050188715&version=1
174 https://mtarchive.geol.iastate.edu/
The data was unzipped and the resulting GRIB files (“.grib2” file extension) were available to import into HEC-HMS,
as shown in Figure 2.
Once the raw GRIB files are added, click “Next” to confirm the selection of the files, as shown in Figure 6.
Alternatively, the clipping shapefile can be created in HMS by selecting “GIS | Export Georeferenced
Elements” and using the subbasin element type. Be sure to buffer any polygon used for clipping by 1 to 2
miles.
The target wkt is a coordinate reference system (CRS) in well-known text (WKT) format. The globe button can be
used to select a typical CRS. Alternatively, a PRJ projection file from a shapefile can be used. The CRS of the
Standard Hydrologic Grid (SHG) was selected, as shown in Figure 8.
If the discretization is a file-specified “.mod” file, the gridded boundary conditions have to be the same cell
size and projection that was used to generate the “.mod” file. “.mod” files were commonly used within
HEC-HMS models prior to version 4.7 that used gridded data. The “.mod” files in earlier versions of HEC-
HMS were typically generated using the Standard Hydrologic Grid (SHG) with a 2,000 meter cell size.
The target cell size is the cell size of the resampled grids. It is recommended that the target cell size be similar to
the cell size used by the MRMS QPE precipitation data. A target cell size of 2,000 meters was selected to correspond
with the original “.mod” file in this case, as shown in Figure 9. A resolution of 2,000 meters is also reasonable for
modeling this watershed which has a total drainage area of 1,440 square miles.
Alternatively, the model discretization could be changed from a file-specified “.mod” file to a structured
discretization and a 1,000 meter grid cell size could be used.
The resampling method is the method used to resample the grid. The Import Wizard allows the user to select
between the Nearest Neighbor, Average, and Bilinear method. A description of each method is included below.
Select the bilinear resampling method for continuous data like precipitation. Click “Next” once the clipping source,
CRS, cell size, and resampling method are selected.
• Nearest Neighbor – A resampling technique for discrete (categorical) data which does not alter the value of
the input cells. This sampling technique is used for nominal or ordinal data where each value represents a
class, member, or classification. Examples of this type of data include land-use, soil, or forest type (ESRI,
2021).
• Average – The value of the output cell is computed from the values of all input cells which overlap the
output cell (ESRI, 2021).
• Bilinear – Bilinear interpolation uses the value of the four nearest input cell centers to determine the value
for the output raster cell. The value of the output raster cell is a weighted average of these four values based
on their distance from the center of the output cell. This resampling technique is well suited for continuous
data or surfaces, like temperature or precipitation (ESRI, 2021).
For HEC-HMS meteorologic models that use gridded boundary conditions, all grids should be resampled to
the same CRS, extent, and grid cell size.
The Gridded Data Import Wizard will now allow the pathnames of the data to be defined. Note that the C-, D-, and
E-Parts of the pathnames use the wildcard asterisk because the program automatically names these parts based on
the type and timestep of the data. The user can specify the A-, B-, and F-Parts which represent the grid type,
geographic area, and data source, respectively. An example for this study is shown in Figure 11 for the Cannon River
watershed study area. Click “Next” to run the Import Wizard which creates the DSS file and converts the MRMS QPE
GRIB data to DSS format for use in HEC-HMS. The time to complete this process varies depending on the amount of
data but typically takes less than one minute.
The Gridded Data Wizard will indicate when it has finished importing the data, as shown in Figure 12. After the
import is complete, click “Close” to exit the Gridded Data Import Wizard.
represents. The data type should be “Precipitation Gridsets”. Click “Create” to add the new precipitation grid to
the HMS model, as shown in Figure 13.
Select the “2016 MRMS QPE” meteorologic model from the Meteorologic Models tree. Select the “Gridded
Precipitation” option within the Precipitation drop down menu and set the Replace Missing option to “Set to
Default”; this allows the simulation to continue if a timestep with missing data is encountered, as shown in Figure
17.
Click the “Basins” tab of the “2016 MRMS QPE” Meteorological Model and ensure that the basin model that will be
used for a simulation is set to “Yes”. In this case the basin model used is the “CannonRev1_Cal-2016” model, as
shown in Figure 18.
Within the “2016 MRMS QPE” Meteorological Model, highlight the “Gridded Precipitation” icon. Select the
“MRMS_QPE_2016” option within the Grid Name drop down menu, as shown in Figure 19.
Provide a descriptive name for the simulation run, as shown in Figure 21. Click “Next”.
Select a basin model to use with the simulation run. In this case, the calibrated basin model for the 2016 event
(“CannonRev1_Cal-2016”) is selected for an easy comparison to show how different precipitation sources affect the
results, as shown in Figure 22. Click “Next”.
Select the meteorological model which uses the 2016 MRMS QPE data, as shown in Figure 23. Click “Next”.
The final step is to select the control specification. For this case, the existing “2016-Sept” Control Specifications
were used for this simulation, as shown in Figure 24. Click “Finish”. The simulation run is now created and ready to
use.
175 https://www.hec.usace.army.mil/confluence/download/attachments/54662087/CannonRev_v1_Solution.zip?
api=v2&modificationDate=1617056247219&version=1
176 http://mtarchive.geol.iastate.edu/
177 https://www.weather.gov/abrfc/pcpn_methods
35.6 References
ESRI. (2021). ArcGIS Desktop. Retrieved from Cell size and resampling in analysis: https://desktop.arcgis.com/en/
arcmap/latest/extensions/spatial-analyst/performing-analysis/cell-size-and-resampling-in-analysis.htm
NOAA NSSL. (2021, 2 24). MRMS QPE. Retrieved from https://inside.nssl.noaa.gov/mrms/
36.1 Introduction
This tutorial demonstrates how Parameter-elevation Regressions on Independent Slopes Model (PRISM) gridded
precipitation can be obtained, imported, and used within HEC-HMS. This tutorial will also show how simulated
runoff changes when different data is used in an existing model of the Cannon River watershed.
178 https://www.hec.usace.army.mil/confluence/download/attachments/54662281/CannonRev_v1_Start.zip?
api=v2&modificationDate=1617075570098&version=1
179 https://www.hec.usace.army.mil/confluence/download/attachments/54662281/Clip.zip?api=v2&modificationDate=1617075638225&version=1
180 https://www.hec.usace.army.mil/confluence/download/attachments/54662281/Data_asc.zip?
api=v2&modificationDate=1617075643921&version=1
181 https://prism.oregonstate.edu/
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 684
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 685
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Once the raw ASC files are added, click “Next” to confirm the selection of the files, as shown in Figure 5.
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 686
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Alternatively, the clipping shapefile can be created in HMS by selecting “GIS | Export Georeferenced
Elements” and using the subbasin element type. Be sure to buffer any polygon used for clipping by 1 to 2
miles.
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 687
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The target wkt is a coordinate reference system (CRS) in well-known text (WKT) format. The globe button can be
used to select a typical CRS. Alternatively, a PRJ projection file from a shapefile can be used. The CRS of the
Standard Hydrologic Grid (SHG) was selected, as shown in Figure 7.
If the discretization is a file-specified “.mod” file, the gridded boundary conditions have to be the same cell
size and projection that was used to generate the “.mod” file. “.mod” files were commonly used within
HEC-HMS models prior to version 4.7 that used gridded data. “.mod” files in earlier versions of HEC-HMS
were typically generated using the Standard Hydrologic Grid (SHG) with a 2,000 meter cell size.
The target cell size is the cell size of the resampled grids. It is recommended the target cell size be similar to the cell
size used by the PRISM precipitation data. In this particular case, the HMS model was developed in HEC-HMS
version 4.2.1 and upgraded to version 4.7.1. The discretization was a file-specified “.mod” file so the projection and
cell size must match the projection and cell size used to create the “.mod” file. The original modeling effort used
the SHG projection with a 2,000 meter cell size. A target cell size of 2,000 meters was selected because this is the
cell size associated with the SHG, as shown in Figure 8.
Alternatively, HEC-HMS versions 4.7 and newer have the ability to update the discretization used in the
model. The model could be updated to use a structured discretization which would allow the user to alter
the grid cell size and projection instead of relying on the grid cell size and projection used to create the
original “.mod” file.
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 688
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The resampling method is used to resample the grid to a user specified size. The Import Wizard allows the user to
select between the Nearest Neighbor, Average, and Bilinear method. A description of each method is included
below. Select the bilinear resampling method for continuous data like precipitation. Click “Next” once the clipping
source, CRS, cell size, and resampling method are selected.
• Nearest Neighbor – A resampling technique for discrete (categorical) data which does not alter the value of
the input cells. This sampling technique is used for nominal or ordinal data where each value represents a
class, member, or classification. Examples of this type of data include land-use, soil, or forest type (ESRI,
2021).
• Average – The value of the output cell is computed from the values of all input cells which overlap the
output cell (ESRI, 2021).
• Bilinear – Bilinear interpolation uses the value of the four nearest input cell centers to determine the value
for the output raster cell. The value of the output raster cell is a weighted average of these four values based
on their distance from the center of the output cell. This resampling technique is well suited for continuous
data or surfaces, like temperature or precipitation (ESRI, 2021).
For HEC-HMS meteorologic models that use gridded boundary conditions, all grids should be resampled to
the same CRS, extent, and grid cell size.
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 689
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The Gridded Data Import Wizard will now allow the pathnames of the data to be defined. Note that the C-, D-, and
E-Parts of the pathnames use the wildcard asterisk because the program automatically names these parts based on
the type and timestep of the data. The user can specify the A-, B-, and F-Parts which represent the grid type,
geographic area, and data source, respectively. An example for this study is shown in Figure 10 for the Cannon River
watershed study area. Click “Next” to run the Import Wizard which creates the DSS file and converts the PRISM
“.asc” data to DSS file format for use in HEC-HMS. The time to complete this process varies depending on the
amount of data but typically takes less than one minute.
The Gridded Data Wizard will indicate when it has finished importing the data, as shown in Figure 11. After the
import is complete, click “Close” to exit the Gridded Data Import Wizard.
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 690
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 691
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
pathname from the DSS file needs to be selected, as shown in Figure 13. Any pathname within the DSS file can be
selected here.
Select the “2014 PRISM ASC” meteorologic model from the Meteorologic Models tree. Select the “Gridded
Precipitation” option within the Precipitation drop down menu and set the Replace Missing option to “Set to
Default”; this allows the simulation to continue if a timestep with missing data is encountered, as shown in Figure
16.
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 692
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Click the “Basins” tab of the “2014 PRISM ASC” Meteorological Model and ensure that the basin model that will be
used for a simulation is set to “Yes”. In this case the basin model used is the “CannonRev1_Cal-2014” model, as
shown in Figure 17.
Within the “2014 PRISM ASC” Meteorological Model, highlight the “Gridded Precipitation” icon. Select the
“PRISM_ASC_2014” option within the Grid Name drop down menu, as shown in Figure 18.
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 693
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Provide a descriptive name for the simulation run, as shown in Figure 20. Click “Next”.
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 694
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Select a basin model to use with the simulation run. In this case, the calibrated basin model for the 2014 event
(“CannonRev1_Cal-2014”) is selected for an easy comparison to show how different precipitation sources affect the
results, as shown in Figure 21. Click “Next”.
Select the meteorological model which uses the 2014 PRISM data, as shown in Figure 22. Click “Next”.
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 695
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The final step is to select the Control Specifications. For this case, the existing “2014-July” Control Specifications
were used for this simulation, as shown in Figure 23. Click “Finish”. The simulation run is now created and ready to
use.
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 696
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The “STR05” subbasin in the model was selected to evaluate how precipitation and loss differs when daily PRISM
data is used instead of hourly NCRFC MPE data. The two factors which influence the difference in the runoff
response the most are the amount of total precipitation and how losses are computed in the model. Table 1 shows
that the PRISM data resulted in more total precipitation but also resulted in significantly more loss and less direct
runoff.
Table 1. STR05 Subbasin Precipitation and Loss Summary
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 697
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
Figure 27 compares results from using NCRFC hourly precipitation data (on the left) to PRISM daily precipitation
data (on the right) for the “STR05” subbasin in the model. The blue bar in each graph represents precipitation and
the red bar represents loss. The original model simulation time step was setup and calibrated using an hourly time
step. The constant loss rate estimated for the “STR05” subbasin was 0.07 inches per hour. In the example below,
the PRISM dataset is a daily dataset that HMS automatically converts to hourly values to match the HMS simulation
time step. Representing the daily total precipitation in even hourly increments results in the incremental
precipitation PRISM data rarely exceeding the constant loss rate of 0.07 inches per hour. Runoff occurs in HMS
when the precipitation rate exceeds the infiltration rate. When NCRFC MPE hourly precipitation data is used, there
are many timesteps where the precipitation rate is significantly greater than the constant loss rate and and has a
higher intensity than the PRISM data.
Cumulative losses for the original analysis using hourly NCRFC MPE data are compared against cumulative losses
using daily PRISM precipitation data in Figure 28. The image shows that the PRISM version of the model results in
higher losses because the hourly representation of daily PRISM precipitation data in HMS for this model is rarely
intense enough to overcome the constant loss rate resulting in more loss.
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 698
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
The purpose of this tutorial is to show how alternate precipitation data sources can be imported into HEC-HMS
using the Gridded Data Import Wizard. Daily PRISM precipitation data was successfully imported and used to
perform a rainfall runoff simulation in HEC-HMS. This tutorial also highlights the importance of selecting an
appropriate dataset for the watershed. In this case, the Cannon River watershed is relatively small and using hourly
NCRFC MPE precipitation data is beneficial to accurately capture the rainfall runoff response and losses.
36.6 References
ESRI. (2021). ArcGIS Desktop. Retrieved from Cell size and resampling in analysis: https://desktop.arcgis.com/en/
arcmap/latest/extensions/spatial-analyst/performing-analysis/cell-size-and-resampling-in-analysis.htm
PRISM Climate Data. (2021). Retrieved from PRISM Climate Group, Oregon State University: https://
prism.oregonstate.edu/
182 https://www.hec.usace.army.mil/confluence/download/attachments/54662281/CannonRev_v1_Solution.zip?
api=v2&modificationDate=1617077371661&version=1
183 https://prism.oregonstate.edu/
184 https://www.weather.gov/abrfc/pcpn_methods
Importing and Using Parameter-elevation Regressions on Independent Slopes Model (PRISM) Precipitation Data – 699
HEC-HMS Tutorials and Guides – HEC-HMS Tutorials and Guides
37.1 Introduction
This tutorial demonstrates how Daymet gridded precipitation products produced by the National Aeronautics and
Space Administration (NASA) can be obtained, imported, and used within HEC-HMS. This tutorial will also show
how simulated runoff changes when different data is used in an existing model of the Black Earth Creek watershed.
185 https://www.hec.usace.army.mil/confluence/download/attachments/54662424/HMS_model_Tutorial.zip?
api=v2&modificationDate=1617115329686&version=1
186 https://www.hec.usace.army.mil/confluence/download/attachments/54662424/Watershed_Boundary.zip?
api=v2&modificationDate=1617115498526&version=1
187 https://www.hec.usace.army.mil/confluence/download/attachments/54662424/Precip_Data.zip?
api=v2&modificationDate=1617115504067&version=1
Data was downloaded for a period of years from the Daymet website as a compressed TAR file and then unzipped in
a 3-step process. First, the TAR file was unzipped using the 7-Zip File Manager. Then, the TAR.GZ file shown in
Figure 2 was unzipped.
Lastly, a final TAR file was unzipped, and a subfolder that contained the gridded Daymet datasets was made
available. Grids for day length, precipitation, shortwave radiation, snow water equivalent, maximum air
temperature, minimum air temperature, and water vapor pressure were all packaged with the Daymet archive, as
shown in Figure 3.
Once the raw NetCDF file is added, click “Next” to confirm the selection of the file, as shown in Figure 7.
Alternatively, the clipping shapefile can be created in HMS by selecting “GIS | Export Georeferenced
Elements” and using the subbasin element type. Be sure to buffer any polygon used for clipping by 1 to 2
miles.
The “target wkt” is a coordinate reference system (CRS) in well-known text (WKT) format. The globe button can be
used to select a typical CRS. Alternatively, a PRJ projection file from a shapefile can be used. For this tutorial, the
CRS of the Standard Hydrologic Grid (SHG) was selected, as shown in Figure 9. If no projection was selected, the
data would retain its original projection.
If the discretization is a file-specified “.mod” file, the gridded boundary conditions have to be the same cell
size and projection that was used to generate the “.mod” file. The “.mod” files were commonly used
within HEC-HMS models that used gridded data and were developed in versions which pre-date HMS
version 4.7. The “.mod” files in earlier versions of HMS were typically generated using the Standard
Hydrologic Grid (SHG) with a 2,000 meter cell size.
The target cell size is the cell size of the resampled grids. It is recommended the target cell size be similar to the cell
size used by the gridded precipitation data. A target cell size of 2,000 meters was selected because this is the cell
size associated with the SHG, as shown in Figure 10. However, a smaller cell size may improve the model results
given the small size of the Black Earth Creek watershed (272 km2).
The resampling method is the method used to resample the grid. The Import Wizard allows the user to select
between the Nearest Neighbor, Average, and Bilinear method. A description of each method is included below.
Select the bilinear resampling method for continuous data like precipitation. Click “Next” once the clipping source,
CRS, cell size, and resampling method are selected.
• Nearest Neighbor – A resampling technique for discrete (categorical) data which does not alter the value of
the input cells. This sampling technique is used for nominal or ordinal data where each value represents a
class, member, or classification. Examples of this type of data include land-use, soil, or forest type (ESRI,
2021).
• Average – The value of the output cell is computed from the values of all input cells which overlap the
output cell (ESRI, 2021).
• Bilinear – Bilinear interpolation uses the value of the four nearest input cell centers to determine the value
for the output raster cell. The value of the output raster cell is a weighted average of these four values based
on their distance from the center of the output cell. This resampling technique is well suited for continuous
data or surfaces, like temperature or precipitation (ESRI, 2021).
For HEC-HMS meteorologic models that use gridded boundary conditions, all grids should be resampled to
the same CRS, extent, and grid cell size.
The Gridded Data Import Wizard will now allow the pathnames of the data to be defined. Note that the C, D, and E-
Parts of the pathnames use the wildcard asterisk because the program automatically names these parts based on
the type and timestep of the data. The user can specify the A, B, and F-Parts which represent the grid type,
geographic area, and data source, respectively. An example for this study is shown in Figure 12 for the Black River
Creek watershed study area. Click “Next” to run the Import Wizard which creates the DSS file and converts the
NetCDF data to DSS format for use in HEC-HMS. The time to complete this process varies depending on the amount
of data but typically takes less than one minute.
The Gridded Data Wizard will indicate when it has finished importing the data, as shown in Figure 13. After the
import is complete, click “Close” to exit the Gridded Data Import Wizard.
from the DSS file needs to be selected, as shown in Figure 15. Any pathname within the DSS file can be selected
here.
Select the “2018_08_Daymet” meteorologic model from the Meteorologic Models tree. Select the “Gridded
Precipitation” option within the Precipitation drop down menu and set the Replace Missing option to “Set to
Default”; this allows the simulation to continue if a timestep with missing data is encountered, as shown in Figure
17.
Click the “Basins” tab of the “2018_08_Daymet” Meteorological Model and ensure that the basin model that will be
used for a simulation is set to “Yes”. In this case the basin model used is the “BlackEarthCreek_Tutorial” model, as
shown in Figure 18.
Within the “2018_08_Daymet” Meteorological Model, highlight the “Gridded Precipitation” icon. Select the
“2018_Daymet” option within the Grid Name drop down menu, as shown in Figure 19.
Provide a descriptive name for the simulation run, as shown in Figure 21. Click “Next”.
Select a basin model to use with the simulation run. In this case, the calibrated basin model for the 2018 event
(“BlackEarthCreek_Tutorial”) is selected for an easy comparison to show how different precipitation sources affect
the results, as shown in Figure 22. Click “Next”.
Select the meteorological model which uses the Daymet data, as shown in Figure 23. Click “Next”.
The final step is to select the control specification. For this case, the existing “2018_08” Control Specifications were
used, as shown in Figure 24. Click “Finish”. The simulation run is now created and ready to use.
37.6 References
ESRI. (2021). ArcGIS Desktop. Retrieved from Cell size and resampling in analysis: https://desktop.arcgis.com/en/
arcmap/latest/extensions/spatial-analyst/performing-analysis/cell-size-and-resampling-in-analysis.htm
NASA. (2021). Daymet Daily Surface Weather and Climatological Summaries. Retrieved from https://daymet.ornl.gov/
188 https://www.hec.usace.army.mil/confluence/download/attachments/54662424/HMS_model_Solution.zip?
api=v2&modificationDate=1617122079400&version=1
189 https://daac.ornl.gov/DAYMET/guides/Daymet_V4_Stn_Level_CrossVal.html
190 https://www.weather.gov/abrfc/pcpn_methods
191 https://www.hec.usace.army.mil/confluence/display/HMSUM/.Normalizer+v4.9
Precipitation Data:
Note that files are provided in a compressed folder and must be unzipped.
The precipitation grids used in this tutorial were converted to HEC-DSS format using the Gridded Data
Import Wizard in HEC-HMS. There are several tutorials showing how to use this tool with different gridded
data products.
In the Select Source Grids panel, select the grids for the time period 25DEC2005:1200 through 10JAN2006:1200.
When the grids are highlighted, press the right arrow in the center of the dialog box to move the files to the second
panel and select Next.
The source grids were selected to begin at 1200 hours because the PRISM data is recorded at 1200 hours.
If, for example, the normal grids began at time 0000, the first source grid should begin at time 0000.
In the next window, under Select Normal File, click the folder icon, navigate to the PRISM.dss file in the Precip_Data
folder, select the file, and click Open.
In the Select Normal Grids panel, select the grids for the time period 24DEC2005:1200 through 10JAN2006:1200.
When the grids are highlighted, press the right arrow in the center of the dialog box to move the files to the second
panel and select Next.
The time periods for the source and normal grids should match. In this example, the normal grids begin at
1200 and end at 1200 the next day. Therefore, the first source grid begins at 1200, and the last source grid
ends at 1200.
Set the Start and End times for the Normalization Period equal to the start and end times of the source grids (25
Dec, 2005: 1200 – 10 Jan, 2006: 1200). Enter dates in mm/dd/yyyy format. Set the Normalization Interval to 1 day.
Then, select Next.
It is very important to select a start time for the Normalization Period that corresponds to the start time of
the normal grids. It is also important to select a Normalization Interval such that a whole number of
normal grids make up one interval.
If for example, the normal grids span 0600-1800 but the Normalization Interval is 1200-1200, some data will not be
included in the normalization computation for a given interval, as shown below:
After selecting the filename and location, the user is prompted to enter the A-, B-, and F-Parts of the normalized
grids. Note that the C-, D-, and E-Parts of the pathnames use the wildcard asterisk because the program
automatically names these parts based on the type and timestep of the data. Enter "Stage IV Norm" for the F part
and select Next. After selecting "Next" the Normalizer Tool computes the normalized grids and exports them to the
user-defined location. The computations are complete when the status bar reaches 100%. After the computations
are complete, the tool can be restarted or closed.
pathname from the DSS file needs to be selected. Any pathname within the DSS file can be selected.
Select the "Stage_IV_corrected" meteorological model from the Meteorological Models tree. Select the "Gridded
Precipitation" option within the Precipitation dropdown menu and set the Replace Missing option to "Set to
Default." This allows the simulation to continue if a timestep with missing data is encountered.
Click the "Basins" tab of the "Stage_IV_corrected" meteorological model and ensure the basin model that will be
used for the simulation is set to "Yes."
Within the "Stage_IV_corrected" meteorological model, highlight the Gridded Precipitation icon. Select the
Select a basin model to use with the simulation run. In this case, use the existing basin model. Click Next.
Select the meteorological model that uses the normalized data. Click Next.
The final step is the select the control specification. For this case, the existing "Dec_2005_Jan_2006" control
specifications can be used. Click Finish. The simulation run is now created and ready to use.
The simulation results show the hourly Stage IV data effectively captures the timing of the flood peaks, but the
magnitude of runoff is significantly less than what was observed at the gage. A greater volume of runoff is generated
when the daily PRISM data is used, but the timing and shape of the hydrograph do not match the observed data
well.
To view the difference in cumulative precipitation volume between the Stage IV and PRISM datasets over the
simulation window, expand the "SonomaCr_S10" basin element for the "Dec2005_Jan2006_base" and
"Dec2005_Jan2006_PRISM" simulations. Then, hold the CTRL key, select "Cumulative Precipitation" for each
simulation, and select the plot icon on the toolbar.
The PRISM data accounted for over twice as much precipitation volume as the Stage IV data during the simulation
period.
The above plot illustrates a common issue with using radar-generated precipitation grids in the western
United States. While it is not clear exactly why there is such a large difference in accumulated volume
between the two datasets, one explanation could be radar beam blockage from mountainous terrain.
Since PRISM data is gathered from a wide range of monitoring networks and applies a more sophisticated
quality control system, it may capture the total volume of precipitation more accurately.
Now, add the cumulative precipitation volume for the normalized grid to the plot by expanding the
"SonomaCr_S10" basin element for the normalized simulation, holding the CTRL key, selecting "Cumulative
Precipitation" for all three simulations, and selecting the plot icon on the toolbar. Note how the cumulative
precipitation for the normalized grid now closely follows the cumulative precipitation for the PRISM dataset.
To view the simulated runoff hydrograph at the Sonoma Creek gage using the normalized grid, expand the
"SonomaCr" element within the normalized simulation and select "Graph." The simulated runoff using normalized
precipitation grid closely matches both the timing and the magnitude of the observed data.
192 https://www.emc.ncep.noaa.gov/mmb/ylin/pcpanl/stage4/
193 https://prism.oregonstate.edu/
Note that files are provided in a compressed folder and must be unzipped.
194 https://www.hec.usace.army.mil/confluence/display/HMSUM/.Normalizer+v4.9
For each normalized grid, the Normalization Period began 10/1/2016 at time 1200 and ended 9/30/2017 at
time 1200. 1200 was selected as the start and end time because the normal (PRISM) grids span time 1200
on day one through 1200 on day two.
While holding the CTRL key, select "Cumulative Precipitation" in the "UpTruckeeRv_S10" results tree for the
"WY2017_MRMS_raw" and "WY2017_PRISM_raw" simulations. Then, select the plot icon on HEC-HMS toolbar.
By the end of the simulation, the cumulative precipitation volume of the uncorrected hourly MRMS data is much
lower than the cumulative precipitation volume of the daily PRISM data.
Now, add the cumulative precipitation for the "UpTruckeeRv_S10" subbasin for the
"WY2017_MRMS_1day_corrected" and "WY2017_MRMS_2day_corrected" simulations to the plot by holding the
CTRL key, highlighting the "UpTruckeeRv_S10" cumulative precipitation result for all four simulations, and
selecting the plot icon on the toolbar. The cumulative precipitation of the corrected, or "normalized," grids more
closely matches the cumulative precipitation of the PRISM data.
Zooming in to the plot during the Mar-Sep 2017 time period shows the normalized grids that used a Normalization
Interval of 1 day resulted in a lower cumulative volume than the PRISM grids at the end of the simulation period.
The normalized grids that used a Normalization Interval of 2 days had a higher cumulative precipitation volume,
but the volume was still ultimately less than the PRISM data.
The lower cumulative volume of the normalized grids relative to the PRISM grids is due to how the normalized grid
cells are computed. For each normalized grid cell computation, if the cumulative volume during the Normalization
Interval of either the source grid or the normal grid is zero, the normalized grid cell value will be zero. This can be
seen in the two computation examples below. In Example 1, the cumulative volume of the source grids over the
Normalization Interval (1 day) is zero. In Example 2, the cumulative volume of the normal grids over the
Normalization Interval (1 day) is zero.
39.3.1 Example 1
39.3.2
Example 2
While small differences in cumulative volume between the corrected, or "normalized," grid and the normal grid may
be negligible over the course of a short simulation, these differences in volume can add up through space and time.
To see how selecting a longer Normalization Interval reduces the difference in cumulative volume between the
normalized grids and the normal grids, add the cumulative precipitation for the "UpTruckeeRv_S10" subbasin for
the "WY2017_MRMS_30day_corrected" simulation to the plot by holding the CTRL key, highlighting the
"UpTruckeeRv_S10" cumulative precipitation result for all five simulations, and selecting the plot icon on the
toolbar. The normalized grids that used a Normalization Interval of 30 days resulted in a similar cumulative
precipitation volume over the subbasin as the normal (PRISM) grids at the end of the simulation.
Using the magnifying glass tool on the resulting plot, zoom in to the January 1-6 storm event. Note the slight
differences in timing between the MRMS data and the PRISM data. During the 05 Jan: 1200 – 06 Jan: 1200 interval,
the PRISM data records a significant amount of precipitation, but the MRMS data records very little. If a
Normalization Interval of 1 day was selected, the normalized grid may overestimate the amount of precipitation
that fell during the 05 Jan: 1200 – 06 Jan: 1200 interval solely due to the difference in timing between the source
and normal grids. However, if the user selected a Normalization Interval that encompassed the entire storm event,
such as a 5-day interval, errors due to differences in timing between the source and normal grids could be
minimized, since the volume from the entire storm event would be accounted for in the computation.
To view the error created by using a 1-day Normalization Interval, plot precipitation for the "UpTruckeeRv_S10"
subbasin for the "WY2017_MRMS_raw," "WY2017_PRISM_raw," and "WY2017_MRMS_1day_corrected,"
simulations. Using the magnifying glass tool, zoom into the January 1-6 storm event. Note the unreasonably high
precipitation value on January 5th.
The image below shows the hyetograph for the same storm event using a normalized grid with a 5-day
Normalization Interval. Because a longer Normalization Interval was used, the shape of the hyetograph of the
normalized grid more closely matches the shape of the raw, uncorrected MRMS data.
The image below shows how each simulation replicated observed discharge at the "Truckee" junction during the
largest flood event in the simulation. While each normalized simulation offered an improvement over the raw,
uncorrected MRMS simulation, the 30-day Normalization Interval resulted in the best fit to the 2017 flood peak.
39.6 Conclusion
Depending on the quality of available gridded precipitation data, creating a normalized precipitation grid may be a
necessary step to achieving a well-calibrated, robust HEC-HMS model. When using the Normalizer Tool in HEC-HMS,
selecting a reasonable Normalization Interval is a critical step in the normalization process. The Normalization
Interval that should be selected for a given HEC-HMS model depends on the size of the watershed, the length of the
simulation(s), the purpose of the model, and the gridded data making up the source and normal grids. While the
decision largely depends on the modeler's engineering judgment, it is often prudent to select a Normalization
Interval that at least spans the duration of a typical rainfall event in the basin. For an event-based model, the
Normalization Interval could be set equal to the length of the simulation. For a long-term, continuous model, a
weekly or monthly interval could be selected. Longer Normalization Intervals minimize error associated with
differences in timing between the source and normal grids and ensure the cumulative volume of the corrected grid
closely matches the cumulative volume of the normal grid at the end of the simulation.
40.1 Overview
This tutorial will step through how to import and apply a gridded curve number dataset in HEC-HMS. The Curve
Number (CN) method is a widely used procedure for estimating precipitation excess/loss that takes into effect land
use and soil types. The CN procedures were empirically derived from studies of small agricultural watersheds. Most
applications of the CN method lumps the CN value as an average of the subbasin or watershed. HEC-HMS allows
the modeler to either apply a subbasin average CN value or use a gridded approach where different CN values can
be applied on a grid cell basis. Each each grid cell within a subbasin receives its own precipitation and losses/excess
are computed based on the grid cell's CN value. To utilize the gridded curve number method in HEC-HMS, the
modeler is required to convert their gridded curve number raster into a gridded DSS record. In the past, the
asc2DSS.exe tool was used to convert ascii format grids to gridded DSS records. Recent development in HEC tools
allow the HEC-HMS modeler to convert gridded CN rasters from common GIS formats to the DSS format using HEC-
Vortex.
This tutorial was tested using HEC-HMS beta 4.9195 and Vortex 10.20196 and requires the users to have access to
these software. We will use the Pilot Creek watershed (watershed located approximately 12 mile south of Knoxville,
TN) for this tutorial. The tutorial assumes you have already created your gridded curve number raster in a GIS. This
tutorial does not describe how to delineate a watershed or how to set up a gridded precipitation simulation. You
must also know how to create an empty DSS file. The project and associated files can be downloaded here:
PistolCreek_Tutorial.7z197
195 https://www.hec.usace.army.mil/software/hec-hms/downloads.aspx
196 https://github.com/HydrologicEngineeringCenter/Vortex/releases/tag/v0.10.20
197 https://www.hec.usace.army.mil/confluence/download/attachments/58630848/PistolCreek_Tutorial.7z?
api=v2&modificationDate=1622137427784&version=1
If you do not have a watershed shapefile, you can easily export one in HEC-HMS by selecting on the GIS
menu → Export Layers
The next step is to import the CN raster file to DSS. Open HEC-DSSVue and create an empty DSS file. Name the file
"CNgrid" and save the file in the data folder of the project directory (\PistolCreek_Tutorial\data\). Open Vortex tool
importer.exe which is located in the ...\vortex-0.10.20\bin folder. In the first window, select the folder to the right
and add the CNgrid.tif located in the gis folder of the project. Click Next to continue to the next screen. Double
click on CNgrid under Select Variables and click Next. In the next screen, leave the Clipping Datasource blank since
the .tif file has already been clipped. For Target wkt, select the image of a globe on the right and choose SHG as the
projection. In the same window, select 50 as the Target cell size. Click Next. In the Select Destination section, select
the empty DSS file, CNgrid. Label the DSS paths as shown in the figure below. Check the Override DSS units and in
the Units string type in UNDEF. Check the Override DSS data type and select INST-VAL as the Data-type. This is a
critical step, units and data type must be set correctly for HEC-HMS to properly read the CN grid.
Click next and the CN grid import should finish very quickly. Open your CNgrid.dss file and plot the CN grid to
ensure the import process completed. You should see something like the image shown below in the Grid Display
tab.
In the Grid Info tab, check that the values match the figure below paying close attention to the Grid Type, Data
Units, Data Type, and cell size.
Once those are confirmed, you will need to rename the Part C of the DSS record to CURVE NUMBER. This lets HEC-
HMS know this grid is a curve number grid. Once these are checked, you are ready to import the grid into HEC-HMS.
You can select SHG or one of the UTM projections as well as any of the available cell sizes listed; however,
your selected projection and cell size must be consistent with your discretization in the HEC-HMS model.
This tutorial has the discretization set to SHG and 50 meters as the cell size. If the discretization has been
set to 2000 meters, then the CN grid would need to be resampled to a 2000 meter cell size. If using gridded
precipitation, the gridded precipitation projection must also be consistent with the CN grid and
discretization. The precipitation grid size can be different from the CN grid cell size where you can have a
50 meter CN grid cell size and a 2000 meter precipitation grid size.
Under the Basin Model, select Pistol Creek basin model. Check through all of the methods and parameters. You'll
notice that all of the methods and parameters have already been provided except for the Loss Method. Under the
Loss Method, select Gridded SCS Curve Number.
In the Discretization tab, make sure the Projection and cell size is set to SHG and 50 meters. These values should be
familiar to you since we set them during the CN grid import process. In the GIS menu, select Compute → Grid Cells
to create your Discretization grid layer. You can view the created layer in the View → Map Layers editor. Make sure
you check "Discretization".
Once you have your Discretization computed, add your newly created DSS CN grid file into the HEC-HMS project.
This can be done by selecting Components→Grid Data Manager.
Under Data Type, select SCS Curve Number Grids. Click "New" and type in Pistol Creek as the CN Grid name.
Once you click OK, the SCS Curve Number Grids folder will appear under the Grid Data folder. Expand the SCS Curve
Number Grids and select Pistol Creek. In the Component Editor, navigate to the location of your DSS file in the data
folder (or wherever you saved your DSS file with the CN grid). In the DSS Pathname, select your CN grid record and
save the project. In the Component editor of your subbasin element, select the CN grid in the loss tab.
Check that the CN grids were set up correctly by running a simulation. Head over to the compute tab and run the
Apr 2017 simulation. The simulation should run to completion.
Although the gridded precipitation method is selected, non-gridded meteorological methods can also be
used with the gridded SCS Curve Number method.
198https://www.hec.usace.army.mil/confluence/download/attachments/58630848/PistolCreek_Tutorial-Complete.7z?
api=v2&modificationDate=1622137480917&version=1