Sixth International Symposium on Hydrological Applications of Weather Radar Melbourne, Australia 2-4 February 2004 PRECIPITATION FIELD ESTIMATION BASED ON RADAR AND RAINGAUGE DATA Anna Jurczyk 1 , Katarzyna Osródka 1 , Stanislaw Moszkowicz 1 , Cinzia Mazzetti 3 and Jan Szturc 2 1 Institute of Meteorology and Water Management, Radar Operational Centre, Poland 2 Institute of Meteorology and Water Management, Branch of Katowice, Poland 3 University of Bologna, Department of Earth and Geo-Environmental Sciences, Italy Abstract Aim of the paper is to get the best estimate of precipitation field. The method consists of three stages. The first spatial interpolation of raingauges data is performed using block-Kriging method. Next radar data is corrected with Bayesian technique. Finally both sets of data are merged using weighted mean method. The research was conducted on small mountainous catchment with hourly sums of precipitation from both radar and raingauges. Radar data was significantly underestimated in comparison with raingauges data. Statistical characteristics: bias between raingauges and radar data and their covariance were calculated. After adjustment of radar data based on Bayesian approach statistical characteristics improved. Merging radar and raingauge data resulted in further improvement. Key Words: raingauge, weather radar, precipitation. Method Fig. 1. Scheme of the method Presented work is aimed at getting estimated precipitation field as close as possible to the “true” field combining raingauge and weather radar data (Jurczyk et al., 2003). Raingauges measure rainfall with good accuracy but only in given points. For that reason the first step consisted in interpolation of raingauges data on a lattice of radar pixels. It was achieved by means of block-Kriging method developed in the frame of the MUSIC project. On the other hand radar provides rainfall field with high spatial resolution however suffering from errors of different sources. Therefore it was adjusted with Bayesian technique proposed by Moszkowicz (2001). Finally the two sets of data were combined with the procedure based on the weighted mean method. combined precipitation field estimation of Bayesian formula real-time raingauge data historical data real-time radar data weighted mean method estimation of variogram parameters block-kriging adjustment with Bayesian approach