H184-02
A testbed for evaluation of forecast skill for Atmospheric River flooding in California watersheds

Tuesday, 15 December 2020: 16:04
Virtual
Qian Cao, University of California Los Angeles, Los Angeles, CA, United States, Steven A Margulis, UCLA, Department of Civil and Environmental Engineering, Los Angeles, CA, United States, F Martin Ralph, Scripps Institution of Oceanography, Center for Western Weather and Water Extremes (CW3E), La Jolla, CA, United States, Hilary K McMillan, San Diego State University, Department of Geography, San Diego, CA, United States, Forest Cannon, Scripps Institution of Oceanography - Center for Western Weather & Water Extremes, La Jolla, CA, United States, Edwin Sumargo, Scripps Institution of Oceanography, La Jolla, CA, United States and Dennis P Lettenmaier, UCLA, Department of Geography, Los Angeles, CA, United States
Recorded Presentation
Abstract:
Atmospheric River (AR) storms produce both benefits to water supply as well as flood hazards in California watersheds. This places a premium on forecast skill and the use of forecasts for water management and flood control. Here we describe an experimental hydrologic forecast testbed, structured around several watersheds distributed across the state with major flood control, water supply, and hydroelectric power reservoirs downstream. These include the Russian River, the Yuba-Feather Rivers, the American River, and the Santa Ana River. We produce flood forecast at multiple USGS stream gauges upstream of the reservoirs at lead times from several days to four weeks. At sub-weekly lead times, we utilize reforecasts from the Center for Western Weather and Water Extremes (CW3E) West-WRF for the period of 1986-2019. For weekly and longer forecasts, we use hindcasts from the NOAA/Climate Testbed Subseasonal Experiment (SubX) project for the period of 1999-2016. In each basin, we generate a 1/32-degree gridded observation data set synthesized using precipitation gauge data from multiple networks. This data set is used for hydrologic model calibration as well as the downscaling and bias correction of West-WRF and SubX reforecasts. We produce hydrologic forecasts using the Weather Research and Forecasting Hydrological model (WRF-Hydro) from the National Water Model (NWM). The model is implemented for all watersheds at a spatial resolution of 3 km and is run at an hourly time step. The testbed serves as a hydrologic model component for an ongoing CW3E initiative to improve forecasts of Atmospheric River-related flooding in California.
<< Previous Abstract | Next Abstract >>