Transition from FLDWAV to HEC-RAS; Forecast Implications and Transition Tools

8/1/2008

The "Flood Wave Dynamic Model" FLDWAV and the "Dynamic Wave Operation" DWOPER are the two main operational hydraulic models currently supported by the Office of Hydrologic Development (OHD) as part of the National Weather Service River Forecast System (NWSRFS). FLDWAV supersedes DWOPER, although DWOPER is still used at some River Forecast Centers (RFCs). The recent report on "Evaluation of Hydraulic Models in Support of NWS Operations" (Hydraulic Model Evaluation Team, 2007) includes the following recommendation:

"It is the consensus of the evaluation team that HEC-RAS be considered for inclusion into the suite of NWS hydraulic models."

HEC-RAS is the Hydrologic Engineering Center-River Analysis System (HEC-RAS) developed by the United States Army Corps of Engineers (USACE). HEC-RAS has some advantages over both FLDWAV and DWOPER including more functionality, a better user interface, better documentation, and easier troubleshooting. Because it will not be cost effective to continue supporting two very similar hydraulic models (FLDWAV and HEC-RAS) in operations, OHD plans to replace FLDWAV with HEC-RAS in the new Community Hydrologic Prediction System (CHPS). In order to accomplish this, two projects are underway: (1) the OHD Hydrologic Software Engineering Branch is coordinating an effort to integrate HEC-RAS software into CHPS, and (2) (this project) the hydraulics group is developing guidance and procedures to efficiently replace existing FLDWAV and DWOPER models with HEC-RAS models without losing accuracy or functionality.

Although FLDWAV and HEC-RAS solve the same basic hydraulic equations, the solution algorithms, representations of cross-section geometry and structures, and methods for calculating conveyance are not identical. One obvious difference is that HEC-RAS implementations most often use more explicit cross-section geometries and different methods to specify boundary roughness. To make the transition, we are evaluating the viability of two approaches: (1) convert existing DWOPER and FLDWAV models to HEC-RAS, including closely approximating the FLDWAV roughness parameterizations, and (2) obtain existing HEC-RAS models for the same rivers of interest (e.g. developed for FEMA studies) and modify them to meet RFC needs. We are working with models on two rivers for this evaluation, the lower Tar River in North Carolina and the lower Columbia River which forms the border between Oregon and Washington. Initial results suggest that either approach is viable and that the first approach may be less work in the short-term but the second approach may have benefits for future flood forecast mapping applications.

The modeling experiments in this project directly address another recommendation from the hydraulic model evaluation team report:

"The team recommends performing a test focusing on comparing development of FLDWAV and HEC-RAS models using identical data sets, assessing difficulties in converting data sets from one program to another and differences in results."

We plan to complete the analysis and documentation for this project by January 31, 2009.