Using HEC-FIA to compute holdouts within a single forecast run is the simplest and most efficient approach to computing FDR. However, there are limitations and assumptions with this process that make it inappropriate for some watersheds. This method works best for watersheds that contain only reservoirs, or watersheds that contain reservoirs and small levees.

There are some other limitations to keep in mind. First, this method only computes FDR for structures, and does not allow for computing agricultural FDR at this time; a future enhancement is planned for such. Second, when cross-watershed reservoir damages are desired, additional steps are needed and are described later on this page.

In this context, "small levee" describes a levee system with negligible hydraulic impacts to mainstem river stages when comparing with-levee versus without-levee scenarios. The assumption is not that these relatively small levees do not have hydraulic impacts, rather their impacts are negligible in the larger scope of the calculations being performed in FDR. For example, if the hydraulic impacts of the levees are less than an elevation difference determined to be reasonable for the watershed/levees, it is likely the hydraulic differences are within the elevation error tolerances of the terrain used in River Analysis System (HEC-RAS) and HEC-FIA models.

Figure 1 is an example of what might be classified as a "small levee." The floodplain behind the levee is relatively small and has a definitive line of high ground, thus small changes in water surface elevation would not result in large changes in the inundated area. Additionally, the volume behind the levee relative to the volume of a typical flood event is not large enough to impact in channel water surface elevations with its removal. 

Example of a Small Levee

To compute FDR for levees, HEC-FIA makes use of a simplifying assumption. The HEC-FIA Holdouts compute assumes that if the levee was removed, water surface elevations (WSEs) within the leveed area are equal to those found along the river centerline. This is generally a reasonable assumption for small levees having negligible hydraulic impacts to mainstem river stages, but not appropriate for large levees with significant hydraulic impacts to mainstem river stages.

To compute FDR for reservoirs, HEC-FIA uses output from two HEC-RAS alternatives—one simulating the regulated (observed) conditions in the watershed and one that simulates the unregulated (natural) conditions in the watershed. The difference in damages between the two alternatives at any given structure (not protected by a levee) is the total flood damage reduced for that structure due to the regulation of all reservoirs. If there are multiple reservoirs reducing the WSEs downstream, HEC-FIA can apportion FDR to each reservoir in one of two ways. Users can specify fixed percentages for each reservoir, or HEC-FIA can use holdout results from the Reservoir System Simulation (HEC-ResSim) to calculate the percentages in HEC-FIA.

The apportioning of FDR between multiple reservoirs becomes more complicated when reservoirs in upstream watersheds provide benefits to downstream watersheds. There are ways to handle this cross-watershed scenario using this approach. One solution is to create a new HEC-ResSim-only watershed combining the individual HEC-ResSim watersheds to compute holdouts. The combined model can be used to compute holdouts for all reservoirs for apportioning FDR between watersheds.

The remainder of this section describes CAVI setup required beyond a typical CWMS watershed setup for real-time forecasting. If the reservoir apportioning methods (including cross-watershed benefits) and/or the levee WSE assumptions are not appropriate for the watershed, see HEC-FIA Individual Computes Approach for a description of the alternative approach.

Forecast Runs

In this approach, only one forecast run is executed to compute the FDR results. The basic flow of data through the models in this approach is shown in Figure 2, which represents the case where the watershed contains at least one reservoir and may or may not include levees. The first half of the model sequence of the FDR forecast run (Meteorologic Forecast Processor (MFP)/Meteorological Visualization Utility Engine (MetVue), Hydrologic Modeling System (HMS), and ResSim) is identical to that of a typical real-time forecast. However, the FDR forecast run diverges at the River Analysis System (HEC-RAS), where two alternatives are required. The results of the two HEC-RAS simulations (regulated and unregulated) are passed to the HEC-FIA Holdouts alternative. 

Forecast Run Schematic - Holdouts Approach

Model Alternatives

At least two new model alternatives are required beyond the typical CWMS real-time forecasting setup. A new HEC-RAS alternative must be created to represent the unregulated (without-reservoir) condition. The second new model alternative required is an HEC-FIA Holdouts alternative. This a new HEC-FIA configuration that computes damages for four scenarios and processes them to return FDR results. You can find more details on how to set up these alternatives in HEC-RAS Model Setup and HEC-FIA Model Setup.

Program Order

Since this forecast run requires two HEC-RAS alternatives, you need to create a new program order. Creating multiple program orders is not enabled by default in the CAVI, but can be accomplished under the Tools menu. On the CAVI Options tab as shown in Figure 3 , check the Advanced Options box, which allows for multiple program orders.

CAVI Interface Options Dialogue - CAVI Options Tab

Once the option is checked, you can create a new program order that includes the additional HEC-RAS alternative. You can then create a new forecast run to compute FDR using the new program order. Alternative keys, model linking, and other setup steps are required to successfully create each FDR forecast run. Figure 4 provides an example forecast run setup within the CAVI.

Example Forecast Run Setup - Holdouts Approach