MONTHLY WATER-BALANCE OF UNGAUGED CATCHMENTS OBTAINED BY GEOGRAPHICAL REGIONALIZATION

Among 24 monthly water balance models, one proved to be suitable for all 75 Belgian basins studied. However, their respective parameter values proved to be significantly different. Each of these 75 basins in turn was considered ungauged and two techniques (kriging and the use of parameter values of a few neighbouring basins) were applied to compute their parameter values. Kriging gives good results in 72% of the basins, whereas the second technique gives good results only in 44%. Thus kriging is significantly better. Even if the requirements for 'goodness' are relaxed, a limited number of basins are badly modelled even by kriging. This may be due to the intrinsic weakness of the technique, poor data quality, or lack of data.