The next step after defining the model linking is to create an Extract List that will collect all data external to the CAVI (ie, not coming from another model) into the forecast. In the Eau Galle River NoQPF and QPF forecasts, meteorological grids will be needed that can be downloaded from the Cumulus website.

Create a Cumulus Extract List

From the Setup tab within the CAVI window, select Models → Edit Extract...

The Extract Editor window will appear with a list of Required Input that was set to come from the Extract List within the Model Linking Editor. In this watershed, all of the required inputs are gridded data for the MetVue alternatives.

In order to import the required data, an Extract Group must be set up and a list made within that group.

Select New Extract Group... and set the Name to Cumulus, Type to Gridded Data and ensure that the Source is set to Cumulus as shown below:

Next, select the refresh () button to load the list of available offices and products.

Set the Office to MVP (to filter the Watershed list to only those within the St. Paul District) and the Watershed to Eau Galle River.

Select the ... button next to the Products box to load the list of gridded data products available from Cumulus.

Carefully highlight and select the following datasets that will be used in the Eau Galle River forecasts by holding the Ctrl key down to select more than one dataset at a time:

  • NDGD RTMA AIRTEMP 1hr
  • MBRFC KRF FCT AIRTEMP 1hr
  • NCEP STAGE4 MOSAIC QPE PRECIP 1hr
  • WPC QPF PRECIP 6hr
  • SNODAS-INTERPOLATED SWE 24hr

Select the Add button to add the datasets to the Selected box and then OK.

Leave the Run by Default box checked to run the extract each time a new forecast is created.

Modify the default Time Window to have a Start Offset of -1 days before the Extract Start. 

Leave the Time Window End set to the Forecast Time and select OK to create the extract group.

The Start Time offset was set back -1 days from the Extract Start to extract one extra day back for the 24hr SWE grids.

An offset should not be set on the Forecast Time because Cumulus uses this date to look up the closest available issuance time of the forecast dataset.

For Forecast data sets, set the end of the Time Window to Forecast Time

  • You can include both observed and forecast datasets in the same extract group if the extract Time Window is set from Extract Start to Forecast Time.
  • For a forecast dataset from Cumulus with an extract Time Window of Extract Start to Forecast Time, the entire forecast version of the product will be extracted into the forecast.dss file.
  • Data may be extracted past the end time of the extract window. For example, if the forecast dataset is a 7 day product, then the entire 7 days will be extracted regardless of the end time of the extract window. The HEC-Metvue Alternative Editor provides an option to trim the amount of forecast grids to use (i.e. only use 3 days of forecasted precipitation instead of the given 7 days).
  • You can separate products into different Cumulus extract groups, but you still need to use the extract Time Window of Extract Start to Forecast Time, even for forecast datasets.
  • Cumulus archives all issued forecast data sets based on the issuance date and time. The end time for an extract (in the above case Forecast Time) is used by Cumulus to return the forecast version issued closest to or at that time. Depending on the time of forecast, setting the Extract Time Window from Forecast Time to End of Simulation may result in missing forecast data sets.

Within the Extract Editor Window, highlight all of the datasets listed in the Required Input window up above and select Add To Group.

The selected data will be added to the Cumulus extract list in the lower part of the window. Notice that the From Cumulus and To forecast.dss pathnames are identical by default. The To forecast.dss pathnames are correct since they were added from the Required Input list; they match the pathnames that the CWMS models will be looking for within the forecast.

However, the From Cumulus names must be updated to ensure they match the exact naming that will be imported from Cumulus.

Highlight the top row of data within the Cumulus extract list, then select Browse From... to open the Cumulus DSS Selection dialog which provides a list of the data that was selected as part of the Cumulus extract list when it was created. 

Highlight the corresponding dataset from the Cumulus DSS Selection window and press Select to update the From Cumulus pathname within the extract list.

 

If the pathname of the dataset coming from Cumulus is different than the pathname in the To forecast.dss column, this process will update the From Cumulus pathname to match the name of the raw data arriving from Cumulus.

For example, the following dataset:

From CumulusTo forecast.dss
/SHG/Eau Galle River/PRECIP///FCST-PRECIP//SHG/Eau Galle River/PRECIP///FCST-PRECIP/

would be updated to:

From CumulusTo forecast.dss
/SHG/Eau Galle River/PRECIP///WPC-QPF-2.5KM//SHG/Eau Galle River/PRECIP///FCST-PRECIP/

In this watershed, the input pathnames were set to match the default names that come from Cumulus so this step was shown for demonstration purposes only and it is not necessary to repeat for the rest of the datasets.

Select Apply to save the changes and OK to exit the Extract Editor.

Prior Task(s):

Task 1 - Create New Forecasts with MetVue

Task 2 - Define Model Linking

Current Task:

Task 3 - Create a Cumulus Extract List

Next Task(s):