This article presents several case studies in southwest Germany, which aimed to support land use management decisions by a process-oriented statistical upscaling of point-related environmental monitoring data to the landscape scale. When techniques of data subsetting were used in a sensible way and corresponding to the appropriate scale for the evaluation envisaged, multiple linear regression offered a data mining technique which was able to spatially predict relatively complex environmental patterns with parsimonious, interpretable and accurate models, whereby different evaluation scales were best represented by different DTM resolutions. Scenario models based upon the regression formulas were a valuable tool for visualizing management options and evaluating management impacts (tree species selection) on soil functions (carbon storage), which qualifies the presented methodology as a useful aid in decision making. Such upscaling techniques may be used for forecasting long-term effects of ecosystem management, but they provided no information on temporal dynamics. Therefore, time trends of point information on soil solution data were scaled by linking them to soil chemical data which was available in higher spatial resolution, using both statistical and process-oriented methods.