Assessment of water quality impacts for different management practices using SWAT model.

摘要

The high yield input strategy has been successful in narrowing the gap between food and fiber requirements and the growing population. However, at the same time it has also threatened the sustainability of land and water resources. Best management practices (BMPs) are technically feasible methods for preventing or reducing nonpoint source pollution to a level compatible with water quality goals. Long-term monitoring of BMP impacts is essential to assess their effectiveness under different conditions. However, it is impractical to monitor all BMPs under all conditions due to time and cost constraints. Computer simulation models provide an alternative to evaluate the response of soil and crops to a range of management practices in an efficient and cost effective way. Testing and evaluation of computer models require the use of extensive field data to ensure that they are reliable for the prediction of management effects. This study was designed to: (1) Calibrate and evaluate the subsurface drainage component of SWAT model; (2) Test the ability of SWAT (version 99.2) model for predicting nitrate-nitrogen (N03N) losses with tile flow, by comparing the model output versus measured data; (3) Application of SWAT model on watershed scale.; In general, SWAT adequately tracked the measured tile drain flows, except that the cumulative monthly tile flows were consistently under-predicted. Differences of -8.4 to 6 and 2 to 11% were determined for the annual simulated tile flows as compared to the corresponding measured flows for the calibration and validation period respectively. Calibration of SWAT was performed using tile flow NO3-N loss data measured in 1995 while validation was conducted by comparing the model output with measured NO3-N losses with tile flow observed in 1993--94 and 1996--97. Differences ranging from 2 to 10% and -7.34 to 5.50 were found between annual NO 3-N losses during the calibration period and validation period respectively, indicating that the model tracked the monthly observations reasonably well. However, the peak NO3-N losses were consistently under-predicted for all three combinations of tillage and cropping systems.; The SWAT model was used to estimate the flow and nitrate loading for UMRW watershed. The model was calibrated for stream flow and NO3-N data measured in 1999 at the outlet of the watershed and model was validated for 2000 and 2001 period. The model accurately tracked most of the peak flow events that occurred during the year, although the peaks were usually over predicted. The model tracked the flow reasonably well but model was unable to track the nitrate trend. The underprediction between the simulated and measured annual flow for year 1999 was 24%, while 35% for year 2000 and 12% for year 2001. The NO3-N was over predicted by 25%, 22% and 108% for 1999, 2000, and 2001, indicating the poor performance of SWAT model in NO3-N simulation.