Research and application of bias correction method for heavy rainfall forecast of ECMWF model

Abstract

Using the daily precipitation data at 81 stations in Anhui province, NCEP 500 hPa height reanalysis data, EC precipitation and 500 hPa height forecast, on the basis of the objective judgment of rainstorm center and weather type, the forecast bias of precipitation center of 23 heavy precipitation events from 2012 to 2018 is classified and counted. The results show that when the EC forecast precipitation center is located in the north of the 584 dagpm line from 115°E to 120°E, the forecast of precipitation center tends to be biased north. According to the latitude difference between the two and the forecast bias of precipitation center, a correction method of main rain belt position based on weather classification is established. meanwhile, according to the daily averaged biases of the precipitation forecast in the maximum precipitation area of 23 heavy precipitation events, a method for correcting the intensity of rainstorm is established. The bias correction method is applied during the plum flood season in Anhui province in 2020, and the results show that the TS score of rainstorm forecast can be significantly improved regardless of the location or intensity correction. After the correction of both location and intensity, the improvement of TS score is more obvious, especially for the two strongest precipitation events in 2020.

Keywords

• ec rainstorm forecast
• weather classification
• precipitation center
• position bias
• correction method