Effects of changes in N-fertilizer management on water quality trends at the watershed scale

In this study, the ADAPT (Agricultural Drainage and Pesticide Transport) model was calibrated and validated for monthly flow and nitrate-N losses, for the 2000–2004 period, from two minor agricultural watersheds in Seven Mile Creek (SMC-1 and SMC-2) in south-central Minnesota. First, the model was calibrated and validated using the water quality data from the SMC-1 and again independently validated with the SMC-2 dataset. The predicted monthly flow and associated nitrate-N losses agreed reasonably with the measured trends for both calibration (r2 = 0.81 and 0.70 for flow and nitrate-N losses, respectively) and validation (r2 = 0.85 and 0.78 for flow and nitrate-N losses from SMC-1, and 0.89 and 0.78 for flow and nitrate-N losses from SMC-2, respectively) periods. The model performed less satisfactorily for the snowmelt periods than it did for the entire simulation period. Using the calibrated model, long-term simulations were performed using climatic data from 1955 to 2004 to evaluate the effects of climatic variability and N application rates and timing on nitrate-N losses. The predicted nitrate-N losses were sensitive to N application rates and timing. A decrease in the fall N application rate from 179.3 to 112 kg/ha decreased nitrate-N losses by 23%. By changing application timing from fall to spring at a rate of 112 N kg/ha, nitrate-N losses decreased by 12%. The predicted nitrate-N losses showed a linear response to precipitation with larger losses generally associated with wet years. A 25% increase in mean annual precipitation would offset reductions in nitrate-N loss achieved using better N fertilizer management strategies described above.