资源环境科技发展态势分析平台

Rainfall separation methods are widely used in literatures, with some algorithms applied to surface rainfall and some applied to radar or other observations. How the effectiveness or practicability of these rainfall partitioning algorithms when applied to model results is not well known. In this study, three convective-stratiform rainfall partitioning methods are compared and evaluated by analyzing cloud properties and associated precipitation budgets based on both observations and grid-scale simulations from a three-dimensional Weather Research and Forecasting modeling of Typhoon Fitow (2013). Results show that convective-stratiform rainfall separation based on surface rain rate is generally more reasonable than that based on radar reflectivity in terms of statistics of vertical velocity and raindrop microphysical budgets. The radar reflectivity partition method does not conduct an exact demarcation between convective and stratiform regions, since the rainfall in stratiform region has similar characteristics as convective rainfall, whereas the rainfall classified as convective rainfall has similar characteristics as stratiform rainfall. Observational analysis based on satellite and lightning confirms these results and shows that the rainfall separation based on surface rain rate generally detects the cellular convective rainfall near eyewall and isolated convection in spiral rainbands, whereas the rainfall separation based on radar reflectivity not only fails to identify the scattered convective rainfall but also overestimates the convective rainfall embedded among the stratiform regions. Above all, the method based on surface rainfall produces more reasonable statistical and microphysical characteristics for both convective and stratiform regions of precipitation, which generally agree with the real rainfall distributions from satellite and lightning observations.