Application of Different Observations in Numerical Weather Prediction of Southwest China

Using the 3. 4 version of WRF assimilation and forecasting system, the assimilation comparison tests about automatic weather stations (AWS), radiosonde, GPS atmospheric precipitation and atmospheric motion vectors (AMVs) data in southwest China were carried out. The purpose is to understand the application effect of these observations used in Southwest Regional Centre numerical prediction system, give a better assimilation scheme when using these observations, and provide a reference for Southwest Regional Centre numerical prediction improvement. Tests show that: 1) There are many AWS stations, but these stations are distributed unevenly. Assimilating AWS mainly adjusts the low elements of first guess field. There are relatively few radiosonde stations, but the vertical observations are rich, so all levels of first guess field can be adjusted after assimilating radiosonde. AMV data are mainly in the 200 ~ 300 hPa and the main effect lies in 200 ~ 300 hPa to first guess field.GPS data have only 25 stations and more than half are excluded, so a little of adjustment to first guess field; 2) The tests of assimilation alone find that radiosonde and AMVs data can improve precipitation forecasts, other two observations have the negative effects to NWP; 3) Using radiosonde and AMVs data at the same time can improve precipitation forecasts in southwest region, especially heavy precipitation center's forecast.