Years of data collection at the bases of 130 lake watersheds in Iowa have enabled us to study the relationships between nutrient concentrations and land use/land cover, as well as to measure the fluxes of key nutrients through hydrologically-linked agroecosystems within the landscape. These ecosystems are dominated by cropland and intensively managed fields with high nutrient export coefficients. Satellite image-derived land cover maps, digital elevation maps and soil mapping units were interpreted and manipulated using geographic information software. We evaluated the relationships between landscape characteristics and lake water quality in Iowa (USA) by regressing four lake water quality responses on landscape variables that were measured for whole watersheds and for three different buffer distances. The response variables were total nitrogen (TN---the combination of dissolved inorganic and organic nitrogen, particulate organic inorganic nitrogen), total phosphorus (TP---includes all forms, dissolved and particulate reactive as well as dissolved and particulate unreactive phosphorus), chlorophyll-a (Chl-a) and Secchi transparency (SD). We developed regression models with a stepwise protocol selecting significant explanatory variables. We tested hypotheses concerning the influence of watershed composition and configuration on the water quality. The models predicted between 15 and 67 percent of the variability of the response variables. Configuration variables were very important regardless of scale. Spatially explicit data improved the statistical power of the whole watershed models for variables TN, Chl-a and SD, but not for TP.