Four-Dimensional Variational Assimilation of Precipitation Data With the Large-Scale Analysis Constraint in the 21.7 Extreme Rainfall Event in China

In this study, the four-dimensional variational data assimilation (4D-Var) method in the Weather Research and Forecasting model is applied to directly assimilate hourly precipitation data to predict an extreme rainstorm process in Henan Province, China. Three simplified microphysics schemes available in 4D-Var are assessed first, revealing that the new regularized WSM6 scheme performed relatively better in precipitation prediction. Meanwhile, precipitation data assimilation (DA) utilizing the China Meteorological Administration Land Data Assimilation System (CLDAS) V2.0 precipitation reanalysis product is evaluated against the experiments with conventional observations in DA and no assimilation. Results demonstrates that it seems that DA with precipitations is able to enhance the accuracy of precipitation forecasts. In addition, it is well known that one of the challenges in convective-scale DA is to extract small-scale information from the observations while maintaining the large-scale balance and mitigating the growth and propagation of large-scale errors. Therefore, the large-scale analysis constraint (LSAC) is further introduced to improve precipitation forecasting. Results indicate that LSAC could effectively adjust large-scale information, including temperature, humidity, and dynamic conditions, thereby improving the precipitation forecasting skills to some extent.