Improving Heavy Rain Forecast by Using Strongly Constrained Radial Wind Analysis

Two data assimilation schemes of LAPS and STMAS have been applied for initializing Advanced Research Weather and Forecast Model (WRF-ARW) forecasts investigating a heavy rain event causing severe flooding and damages in 2016. They used identical datasets and boundary conditions in this experiment for comparing these data assimilation schemes. We conclude that (1) LAPS and STMAS analysis can capture mesoscale information, but their distribution characteristics are different. (2) TS scores in 24h/6h accumulated precipitation forecasts show that rainfall forecasted by STMAS scheme is better than LAPS scheme above 10mm,especially in 100mm/24h and 50mm/6h. (3) TS scores of 6h radar reflectivity forecasted by STMAS are much higher than by LAPS scheme at 5、10、20、30、40dBz. (4) Radar reflectivity forecast by STMAS is closer matched with the observation compared with the LAPS scheme, since the STMAS scheme adjusts the humidity directly according to the observation in the variational assimilation, while the impact of the radar echo on the humidity in the LAPS scheme through cloud analysis module. (5) There is significant difference between the wind analyses of the two schemes due to the different methods. The LAPS scheme uses a continuous correction method and STMAS adopts 3DVAR method. STMAS analyzes 3-d wind with a strong constraint for assimilating radial wind.The strong constraint has the effect of the dynamic constraint, which makes the analysis possess more physical significance, which has resulted improved impact on NWP forecasts.