A Continuous version of distributed parameter model, ANSWERS (ANSWERS 2000) was applied to a field-sized watershed planted to cotton in the Limestone Valley region of northern Alabama. The field was cultivated for three years with conventional tillage followed by three years of conservation tillage. Overall, the ANSWERS model simulated runoff and nutrient losses in surface runoff within an acceptable range for the conventional tillage system conditions in continuous simulation mode. But the sediment losses predicted by ANSWERS were initially on the order of fifteen times or more higher than measured regardless of tillage systems. In order to duplicate measured data, the sediment detachment coefficient of rainfall and flow had to be reduced for calibration. The model poorly predicted soluble nutrient losses for the conservation tillage system due to the model's weakness in representing the surface application of fertilizer under this practice. The model simulates only one soil layer, in which soil moisture, nutrient concentration, and soil characteristics are assumed homogeneous. Currently, the model does not consider vertical nutrient concentration variation in soil profile. During the conservation tillage system, corn stalk and the residue of a winter cover crop were spread on the soil surface. However, the model did not properly represent surface spreading of crop residue, thus the model was unable to consider the organic-nitrogen contribution from crop residue to the erodible soil surface. This resulted in poor prediction of sediment-bound TKN, especially for conservation tillage system.
