The aim of presented paper is to get estimated precipitation field as close as possible to the “true” field combining raingauge and weather radar data. The investigations are performed for small, mountainous catchment (the upper Vistula River catchment in Poland, area about 300 sq. km). The works are conducted using software developed in the frame of 5th EU FP Project called MUSIC (“Multi-sensor precipitation measurements integration, calibration and flood forecasting”).

The specific of small mountainous catchment is that precipitation is very changeable in spatial scale and on the other hand small number of raingauges is available. Therefore the problem of precipitation field estimation is more difficult.

The first step is spatial interpolation of raingauge data which is performed using block-kriging technique. Next data from both sources (raingauge measurements after spatial interpolation and radar data) are merged using Kalman filter for Bayesian combination. The both steps are carried out with more simple methods based on weight distance as well.

Quality of estimated precipitation field is controlled basing on statistical analysis where “true” precipitation field is unknown.