The gauge adjustment technique used to produce 2 x 2 km 3- and 12-hour radar-based accumulated precipitation datasets for the Baltic Sea Experiment is presented. The gauge adjustment technique is based on the gauge-to-radar ratio. A distance-dependent adjustment factor is derived and it is weighted against a spatially analyzed adjustment factor according to the local observation density and estimated spatial decorrelation distance. A preliminary adjustment strategy is applied in order to normalize data from many radars to a common level and to minimize the bias with gauges. The final adjustment field applied to radar accumulations is shown, through validation against independent gauge data, to minimize the bias between radar and gauge sums while raising the explained variance, compared to unadjusted radar sums. Areas not covered by radar are subjected to an optimal interpolation of systematically corrected gauge sums, and this field is merged with the gauge adjusted radar field in order to cover the entire Baltic Sea Experiment region. (C) 2000 Elsevier Science Ltd. All rights reserved.