Improvements And Application On WRF 3DVAR System For Short-term Convective Precipitation Forecasting

The short-term quantitative precipitation forecasting(QPF)has long been a challenge to the research and operational communities.The application of high-resolution numerical weather prediction(NWP)systems improves the short-term convective QPF.And the specification of an accurate initial state is critical for NWP.The momentum variables of stream function and velocity potential are used as control variables in a number of operational variational data assimilation systems.However,it is shown that for limited-area high-resolution data assimilation,a ??variational system tends to over-smooth increments and ignore the small-scale information and hence may result in degraded analysis and forecast.In this study the new system with x and y components of wind(UV)as control variables is developed based on WRF 3DVAR.Assimilation and forecast experiments are conducted with multiple cases to investigate the impact of control variable on short-term convective QPF.The main findings are that,the UV variational system is able to assimilate convective-scale information from observations,and meanwhile maintain large-scale background balance.The use of UV control variables allows the analysis wind fields closer fits to radar observations,and improves the subsequent precipitation prediction.The study designs a multi-step assimilation scheme for hourly update cycle system.It is an effective way of assimilating observations of different scales and keeping multi-scale balance.The scheme is to assimilate synoptic and radar observations in different steps and extract both large-scale and convective-scale information from observations typically representing different scales.The implementation of the two-step assimilation scheme in the hourly update system results in improved QPF throughout most of the 12h forecast period in multiple cases.The detailed analysis shows that the two-step assimilation is able to preserve key convective-scale as well as large-scale structures,and it improved precipitation forecast in terms of the area coverage and quantity of rainband.The u and v wind are used as control variables in the new version of WRFDA,but there is no cross-correlation between u and v.The imbalances in the analysis could lead to the occurrence of spurious precipitation in the forecast.So a divergence constraint is added to the cost flunction of WRFDA 3DVAR.This technique aims to suppress noise introduced in the assimilation process.The divergence constraint is applied in hourly update cycle,and shows effective in improving the balance in the analysis,which leads to less spurious convection and improved QPF scores.The improvements of the two-step assimilation scheme and the new constraint were further verified by eight convective cases and shown to be statistically significant.In this paper,a WRFDA system with new control variables is developed,a new assimilation scheme is designed,and a new constraint is added to 3DVAR to improve short-term convective precipitation forecasting.The improvements of the new techniques are verified in multiple cases.The modifications for the WRFDA code are integrated in the latest official version.The new version and the scheme have been running in real-time system.This study provides effective and practical guidance for operational forecast.