Comparison of two LAPS data merging methods for typhoon "HaiKui" forecasting

There are two kinds of data immerging methods in the LAPS system. The early method is based on the Barnes interpolation method, namely the "LAPS method". The latest method is based on the continuous variation method, namely the "STMAS method". These two methods have been applied in analyzing and forecasting a typhoon case (HaiKui, 201211) during 2012. The results are as follow. (1) The typhoon center and the circulation structure cloud be revealed accurately in the output fields of two analysis methods. For the analysis fields based on the "LAPS method", the analyzed intensity of the typhoon is weaker than observed, but more small-scale systems could be found within it. For the results of the "STMAS method", the analyzed intensity of the typhoon is stronger, and the analysis fields are more continuous and smooth than observations. (2) For geopotential height, humidity, low-level wind and temperature fields, the analyses based on the "STMAS method" are closer to soundings than those from the "LAPS method". (3) When WRF model is initialed with the "LAPS method" analysis field, typhoon path forecasting error was bigger than when it is initialed with the "STMAS method". But threat scores for 0-6 h lead time and heavy rainfall center (more 100 mm) 24 h forecasts based on the "LAPS method" are both better than those based on the "STMAS method". When WRF model is initialed with the "STMAS method" analysis field, the typhoon strength forecasting was slightly stronger than observation. Furthermore, the threat scores for 12-24 h lead time and middle rainfall (more than 25 mm and 50 mm) 24 h forecast are both better than traditional method. (4) The threat score of precipitation forecasting for the average of two methods is higher than any one of two methods.