While the effect of cultivation on soil properties has been well documented, its effect on the spatial distribution of soil properties is less well understood. The purpose of this study is to use GIS classes, soil map units and landscape positions, and geostatistics to characterize the spatial distribution of soil properties in a native prairie and agricultural field. A secondary purpose is to use soil color in combination with these techniques to predict soil organic carbon (SOC) content to 0.2 and 1.0m depths across each land use. Each land use was sampled in an unbalanced hierarchical nested grid for a total of 406 cores. Soil color was measured with Munsell Soil Color Book and chroma meter with three types of samples: (a) prepared samples, ground to <2mm, (b) horizon peds, and (c) split cores (measurements taken at horizon and depth increment mid-points). Standard techniques were used to describe all cores and analyze a subset (63 in each land use) for soil organic carbon (SOC), bulk density, percent water stable aggregates (WSA), pH, and surface horizon texture. Bulk density, pH, and WSA are not spatially dependent using any technique. Using GIS classes, the prairie has more significant differences in soil properties between classes. The agricultural field is more homogenous, but geostatistics show it has spatial dependence with small-scale continuity. SOC content distribution is related to localized, mid-slope wetness in the prairie that no longer occurs in the agricultural field due to artificial drainage. Only a few models in this study were generally satisfactory for predicting SOC contents. SOC content is significantly related to soil color, on individual samples, but not entire cores. The best predictor of SOC content is topographic wetness index in the agricultural field and kriging and co-kriging in the prairie. Average land use predictions vary by 2.4 kg m-2 for 0.2m and 3.8 kg m-2 for 1.0m in the agriculture field and 6.2 kg m-2 for 0.2m and 19.0 kg m-2 for 1.0m in the prairie. Agricultural cultivation has changed the distribution of SOC across the landscape and thus different models are needed to make accurate predictions.