Comparisons of a suite of gauge, radar, and gauge-radar blended quantitative precipitation estimates over a mountainous region

Accurate estimation of rainfall rates and amounts is critical to numerical weather prediction and hydrologic applications. To better understand the strategies that could yield the most accurate precipitation estimates over mountainous regions, we created and evaluated a suite of radar and gauge quantitative precipitation estimation (QPEs) products over an area that lies in the NOAA Hydrometeorology Testbed in northern California (HMT-West). In the experiments, we started with an offline version of the Multisensor Precipitation Estimator (MPE) system that is currently used at most of the National Weather Services (NWS) River Forecast Centers (RFCs). The MPE generates a set of rainfall products on the approximate 4km X 4km HRAP grids based on the radar and gauge rainfall observations. The accuracy and statistical properties of a suite of MPE products, including gauge-only, radar-only, bias corrected radar, and gauge-radar blended QPEs, were assessed using independent gauge data, with special attention given to the impacts of mean field, local bias correction, and multisensor blending. Comparisons with similar products from the Multi-radar, multisensor (MRMS) were also conducted to determine the benefit of accounting for vertical profile of reflectivity and relative strengths of MPE and MRMS bias correction schemes.