STATISTICAL ADJUSTMENT OF RADAR-BASED DAILY PRECIPITATION TO GROUND DATA FROM THE CZECH TERRITORY

摘要

Following conclusions can be drawn.: 1. Classified regression (KREG) improves the radar-based estimate of daily precipitation in comparison with the other models tested as well as with the interpolation of ground data. 2. KREG is better than interpolation for a wide range of gauge density (including the 81 on-line gauges on the Czech territory). 3. The first tests with KREG adjustment of 1h precipitation (U.S. data) show improvement in RMSE and BIAS. Nevertheless, additional tests with Czech data are needed. 4. Potential for improvement of the KREG can be expected (1) in considering a more complex radar information than the Zmax values, and (2) in the inclusion of other physical aspects into the classification. Especially separate dealing with convective and stratiform precipitation should be considered.