本文利用中尺度模式Weather Research and Forecasting Model (WRF) 3.1版本及National Centers for Environmental Prediction (NCEP)分析资料,就2003年6月下旬我国江淮及南方地区的强降水事件,以24 h短期天气模拟的方式,研究了模式中四个不同陆面方案对降水模拟的影响.结果表明,此次暴雨事件模拟对不同陆面方案是比较敏感的,模拟区域内雨量级别越高,不同方案的TS评分差异就越大,较大范围雨量可存在30%的差异,四种方案的暴雨中心值可存在100%~150%的较大差别;不同陆面方案还导致了模拟平均感热通量及潜热通量的系统性差异,这些差异的分布具有地域特点;陆面方案通过两种机理对模拟降水产生重要影响,即主要影响地表蒸发量,以及主要影响低层环流及水汽辐合,从而分别影响模拟的较大范围降水(如,平均约7%、最大约30%的较大范围雨量差异)及包含模拟降水中心的较小范围暴雨(如,方案间暴雨中心雨量可存在100%~150%的较大差别).可见,不同陆面过程可从不同空间尺度、不同程度上影响暴雨天气,改进陆面方案可以提高WRF模式对暴雨的模拟能力.
The simulation of a heavy rainfall event occurring in the Yangtze-Huaihe valley and the south of China during late June, 2003 was conducted to examine the effects of different land surface schemes on simulated precipitations using the Weather Research and Forecasting Model (WRF) Version 3.1 and National Centers for Environmental Prediction (NCEP) analysis. The simulation was performed in the short-range mode for 24-h integrations. The results showed that, generally, the simulated heavy rainfall event is sensitive to different land surface schemes, the scheme-induced difference of threat score becomes larger as the level of rainfall gets higher within the relatively large study subarea, where the scheme-induced relative difference of precipitation can amount up to 30% with an average of 7%, while the maximum values of daily precipitation differences can be as large as 100%-150%, and different schemes lead to simulated systematic differences in averaged sensible and latent heat fluxes that are characterized by regional distributions. Finally, the land surface schemes show to substantially affect the simulated precipitations via two mechanisms, i. e. , by affecting land surface evaporation, and low-level atmospheric circulation and water vapor convergence, the schemes exert great influences, respectively, on the simulated rainfall over a relatively large area of model domain (e. g. , with an average difference of 7% and a maximum difference of ~ 30%), and on the simulated heavy rainfalls within small areas including rainfall centers (e. g. , up to differences within 100M 150M). All these suggest that different land surfaces affect heavy rainfall weather at different spatial scales and to different extents, and that improving the land surface schemes can lead to better simulation of the weather with the WRF model.