An Ensemble Concept that May Reduce the Inherent Dry Bias in Complex Terrain

In areas of complex terrain, significant variations may exist between probability of precipitation and qualitative precipitation forecasts in a relatively short distance. Not only is this problematic for the meteorologist, but it is also a challenge for the users who interpret the forecast. Most populated centers or transportation routes in complex terrain exist at drier valley bottoms, with the higher terrain receiving the bulk of the precipitation. Verification studies or results have depicted an inherent dry bias in areas of complex terrain, especially in valley locations where most of the populated centers and transportation routes exist. The use of model ensembles may have improved the forecasts, but the inherent dry bias remains. This presentation will examine a two-pronged approach that may reduce this bias. This methodology separates the big picture model data (e.g. GFS, ECMWF, etc…) from the higher resolution model data (e.g. RAP, WRF, etc…). Instead of one ensemble blend, this concept combines two ensemble blends. The first ensemble identifies significant synoptic scale features; the second ensemble is composed of higher resolution model data that provides finer details such as mesoscale features and terrain enhancements. This presentation will check the viability of such an approach and whether this reduces or eliminates the dry bias in complex terrain.