Influence of management history and landscape variables on soil organic carbon and soil redistribution

Controlled studies to investigate the interaction between crop growth, soil properties, hydrology, and management practices are common in agronomy. These sites (much as with real world farmland) often have complex management histories and topographic variability that must be considered. In 1993 an interdisiplinary study was started for a 20-ha site in Beltsville, MD. Soil cores (271) were collected in 1999 in a 30-m grid (with 5-m nesting) and analyzed as part of the site characterization. Soil organic carbon (SOC) and Cesium (Cs) were measured. Analysis of aerial photography from 1992 and of farm management records revealed that part of the site had been maintained as a swine pasture and the other portion as cropped land. Soil properties, particularly soil redistribution and SOC, show large differences in mean values between the two areas. Mass C is 0.8 kg m greater in the pasture area than in the cropped portion. The pasture area is primarily a deposition site, whereas the crop area is dominated by erosion. Management influence is suggested, but topographic variability confounds interpretation. Soil organic carbon is spatially structured, with a regionalized variable of 120 m. Cs activity lacks spatial structure, suggesting disturbance of the profile by animal activity and past structures such as swine shelters and roads. Neither SOC nor Cs were strongly correlated to terrain parameters, crop yields, or a seasonal soil moisture index predicted from crop yields. SOC and Cs were weakly correlated (r ~0.2, F-test P-value 0.001), suggesting that soil transport controls, in part, SOC distribution. The study illustrates the importance of past site history when interpreting the landscape distribution of soil properties, especially those strongly influenced by human activity. Confounding variables, complex soil hydrology, and incomplete documentation of land use history make definitive interpretations of the processes behind the spatial distributions difficult. Such complexity may limit the accuracy of scaling approaches to mapping SOC and soil redistribution.