Assimilation of Null-echo from Radar Observations for Short-term Precipitation Forecasting

A new method to assimilate null-echo from radar observations is developed for short-term quantitative precipitation forecasting (QPF). A null-echo is defined as a region with no precipitating echo(s) within the radar observation range. This information is used to suppress and correct erroneous model precipitation. The model removes excessive humidity and four types of hydrometeor contents (wet and dry snow, graupel, rain water) based on radar reflectivity by a three-dimensional variational (3D-Var) data assimilation technique within Weather Research and Forecasting Data Assimilation (WRF-DA) system. Some procedures of preprocessing radar reflectivity data and the use of null-echo in the assimilation are presented.

A case study of mesoscale convective system that occurred on 23 July, 2013 in South Korea is carried out to test the method for improving short-term convective precipitation. Positive impacts of short-term QPF scores are found. The overall statistical evaluation shows improvements in the analysis and objective conventional forecast scores of up to 12 hours.