A strategy for controlling error of distributed environmental models by aggregation

In this paper a strategy for controlling error when calculating model param eter values from low-resolution low-accuracy data (LRLA-data) is presented. The basic principle of the proposed strategy is that the error in the mode l output can be reduced by output aggregation. By comparing the output from LRLA input data and high-resolution, high-accuracy input data (HRHA-data) at various aggregation levels, the error due to a decrease in accuracy and resolution of the input data can be assessed. An aggregation level can then be chosen so that an output is generated with the desired level of error a nd sufficient spatial detail. The method may also be used to evaluate the i nformation loss when simplified models are used. The calculation of the top ographic factors in an erosion model is taken here as an example applicatio n.