资源环境科技发展态势分析平台

Snow falling into a melting layer will eventually consist of a fraction of meltwater and hence change its characteristics in terms of size, shape, density and fall speed. Most microphysical parameterizations in numerical weather prediction models typically only represent purely solid or liquid hydrometeors. Generally, this has been an acceptable compromise since the melting layer is typically very shallow and adding a mixed solid/liquid particle type would result in increased computational time. This research shows how improvements were made to the treatment of melting snow in a microphysical parameterization within the Weather Research and Forecasting (WRF) model by implementing an approximation of snowflake melted fraction together with a physically-based expression for melting particle terminal velocity. In addition, the more appropriate definition of melting level defined by the wet-bulb temperature was consistently used in various process rates, all while not adding additional prognostic variables that would add computational cost. Multiple events observed during the 2015–2016 Olympic Mountain Experiment (OLYMPEX) were used to compare with the WRF model results. The modified scheme is able to represent disdrometer observations of joint particle size and fall velocity during wet snow events, as well as fall velocity profiles through the melting layer derived from a vertically-pointing radar. The improved scheme removes ‘bulls’ eyes' of snow accumulation in lee-side areas within the melting zone, and should result in better predictions of surface precipitation phase and amount.