关键词: 集合预报, WRF模式, 强天气过程, 对流尺度

Abstract: As an effective method to reduce the uncertainty of numerical models, the ensemble forecast has been widely applied to numerical weather prediction(NWP) of severe weather processes such as heavy rainfall and severe convective weather. Based on the research progress of WRF model ensemble forecast in China and abroad, this paper summarizes and reviews the improvement of initial field of WRF model, the construction of perturbation method of ensemble forecast and the development of convection-scale ensemble forecast. The results show that the data assimilation technology can provide micro-scale and meso-scale information for the initial field of large-scale model, which is beneficial to improve the forecasting ability about severe weather processes. The construction of reasonable multi-initial values, multi-boundary and multi-physical processes perturbation schemes in ensemble forecast can accurately characterize the uncertainty of atmospheric evolution, and thus improve the dispersion of ensemble forecast and obtain the more accurate result than deterministic forecast. Convection-scale ensemble forecast can simulate severe weather processes better, but how to develop the perturbation methods and evaluation methods to apply to convective-scale ensemble forecast is a difficult problem for the study of convective-scale ensemble forecast.

Key words: ensemble forecast, WRF model, severe weather process, convection-scale