Diagnostic testing and evaluation of the community WRF-Hydro Modeling System for national streamflow prediction application

Abstract:

A fully-distributed WRF-Hydro modeling system developed at National Center of Atmospheric Research (NCAR) will serve the initial operational nationwide streamflow forecasting needs of the National Water Center (NWC). This paper presents a multi-faceted evaluation of the WRF-hydro modeling system in preparation for operational national streamflow prediction. The testing period encompasses the 2015 warm season which included the National Flood Interoperability Experiment (NFIE) where WRF-Hydro and the RAPID channel routing model were driven by the Multi-Radar Multi-Sensor (MRMS) estimates as the real-time precipitation estimate product and the High Resolution Rapid Refresh (HRRR) for the short term forecast. Here, we validate the MRMS estimates and HRRR precipitation forecasts at national scale using daily precipitation observations from the Global Historical Climatology Network (GHCN). Because WRF-Hydro has several physics options such as surface overland flow, saturated subsurface flow, channel routing as well as conceptual deep groundwater base flow also conducted additional simulations to evaluate WRF-Hydro performance under different processes configurations. Streamflow verification data for model simulations and predictions was completed for a subset of GAGES-II reference basins. Multi-temporal and spatial scale verification is performed in order to test the robustness and skill improvement in WRF-Hydro streamflow simulations under different configuration over a wide range of basins sizes and from short-term (hourly) to longer-term (monthly) flow simulations. Evaluation will be also carried out based on various geographic regions to relate the skill improvement to dominant controls on flow based on the actual physical and climatic properties of the basins. The goal is to inform WRF-Hydro model configuration for the initial operating capabilities (IOC) project and target processes and parameter estimates for improvement.