Setting up an Ensemble Extract Group
Just like deterministic forecast runs, ensemble forecast runs require input data. This data must be added to an Extract Group so that HEC-RTS can properly access the required inputs during computation. Extract Groups containing ensemble data can be used with both Ensemble Forecast Runs or standard deterministic Forecast Runs.
This guide walks you through the process of creating a new Ensemble Extract Group and configuring it with the necessary information to ensure your forecasts run successfully.
Prerequisites
Before starting, ensure you have an HEC-DSS file that contains ensemble collection data.
Step 1: Open the Extract Editor
- Launch HEC-RTS and open your watershed.
- Navigate to the Setup tab.
- From the Models menu, select Edit Extract.
- The Extract Editor dialog will appear.

Step 2: Create New Extract Group
- In the Extract Editor, click New Extract Group.
- A new group is required since ensemble time series collections must be stored in local .dss files.
- In the New Extract Group dialog:
- Enter a Name for the extract group.
- Set Type to Time Series.
- Set Source to DSS.
- In the Data Store field, enter the name of your HEC-DSS file that contains the collection data.
- Configure the Storage Options and Time Window settings as needed.
- Since ensemble data typically falls within the forecast period, this is often set from Forecast Time to End of Simulation, but the exact settings depend on your specific forecast configuration.
- Optionally, check Run by Default if you want the extract to run when a new forecast is created.

- Click OK to create the extract group.
Step 3: Select Required Inputs
- In the Required Input table, select all boundary conditions necessary for your model run that will be using collection data.
- Click Add to Group.
- The selected input pathnames will now appear in the Extract Linking section below.
Step 4: Assign Collections to Required Inputs
Now you need to link the appropriate ensemble collections from your HEC-DSS file to the required boundary conditions.
- In the Extract Linking table, double-click the first row under the From: filename column, then click the ellipsis (...) button.

- The DSS Time Series Record Chooser will open.
- From the View menu, select Condensed - Group Collections to view your ensemble collections.

- Select the collection pathname you want to associate with the boundary condition and click Set Pathname.
- The pathname of the first member of the collection will appear in the Extract Editor, as seen below.

- Repeat this process for each input in the table.
- When all inputs are assigned, close the DSS Time Series Record Chooser.
Step 5: Add Data Using the Adhoc Method
If needed, you can manually add data to the extract group using the Adhoc method. This allows you to include any HEC-DSS record and optionally modify the destination pathname in the To forecast.dss column. If this step isn't required, skip ahead to Step 6.
- Click Add Extract Ensemble.

- The Ensemble Adhoc: DSS Time Series Record Chooser will appear.
- Open the HEC-DSS file containing the ensemble collections you want to add.
- Select a single member from the ensemble collection and click Add Pathnames.

- A dialog will appear prompting you to choose which ensemble member to extract. You can:
- Select a specific subset, or
- Leave the default option (represented by *) to extract all members.
- Click OK to add the record.
- Repeat the above steps to add more ensemble records if needed.
- When finished, close the Ensemble Adhoc DSS Time Series Record Chooser to return to the Extract Editor.
- If necessary, modify the pathname for any adhoc record in the To forecast.dss column.

Step 6: Save the Extract Group
- Review your setup to ensure all required inputs are included and correctly linked to the appropriate collections.
- Make any necessary adjustments.
- Click Apply to save the extract group.
- Click OK to close the Extract Editor.
Final Review
You’ve now created and configured a new Ensemble Extract Group in HEC-RTS. This ensures your ensemble forecast runs have the data they need to compute successfully.