Verification of GEFS Precipitation Forecasts and the Implications of Statistical Downscaling over the Western United States

Quantitative precipitation forecasts (QPF) from operational medium-range ensemble modeling systems currently lack sufficient resolution to adequately resolve orographic influences on precipitation in mountainous regions. In this study, we verify 55km-resolution Global Ensemble Forecast System (GEFS) cool-season QPF over the western U.S. relative to Climate Prediction Center (CPC) precipitation analyses and Snow Telemetry (SNOTEL) observations. Results indicate that GEFS significantly under-predicts events greater than about 12.7 mm, with biases and accuracy measures such as equitable threat score generally worsening towards the interior. In an effort to improve QPF without exacerbating computing demands, we explore statistical downscaling based on 30-arc second (~800m) climatological precipitation analyses from the Parameter-elevation Relationships on Independent Slopes Model (PRISM). Such downscaling generally improves model biases and accuracy measures, especially over the western interior, despite some increase in false alarm ratios. These results, combined with ongoing work examining probabilistic GEFS QPF, should assist operational forecasters and hydrologists over the western U.S. and other regions of complex terrain.