Submitted to: Transactions of the ASAE
Publication Type: Peer reviewed journal
Publication Acceptance Date: 12/15/2000
Publication Date: 3/1/2001
Citation: Bakhsh, A., Kanwar, R.S., Jaynes, D.B., Colvin, T.S., Ahuja, L.R. 2001. Simulating effects of variable nitrogen application rates on corn yields and NO3-N losses with subsurface drainage water. Transactions of the ASAE. 44(2):269-276. Interpretive Summary: Agriculture is a major source of nitrate contamination of surface and ground water across the U.S. New nitrogen management strategies need to be developed that reduce off-site movement of nitrate, but maintain crop production. Because the soils and climate are extremely variable across the U.S., it is not cost effective to develop and fine-tune nitrogen management strategies for every combination of the two through field tests. Instead, our best strategy is to develop computer models that simulate the interaction of soil, weather, plants, and management that can be used to test and identify best management practices for any combination of factors. The Agricultural Research Service is developing the model RZWQM that can be used for this purpose. In this research, we tested the model's ability to predict the effect of different nitrogen fertilizer rates on corn and soybean yield and off-site nitrate movement by comparing the model's prediction of crop yield and drainage water quality to 4 years of measured data from a farmer's field. The model predicted well the yield response to varying rates of nitrogen fertilizer and predicted the amount of nitrate leaching to the subsurface drains accurately. These results give scientists confidence in the use of the model to develop best nitrogen-fertilizer management practices for a wide range of soils and climates, thereby improving our Nation's water quality while maintaining the profitable production of food and fiber.
Technical Abstract: Using a model as a management tool requires testing of the model against field measured data prior to its application for solving natural resource problems. Therefore, the Root Zone Water Quality Model (RZWQM98) was tested using 4 years (1996 to 1999) of field-measured data to predict the effects of different N-application rates on corn yields and NO3-N losses with subsurface drainage "tile" water. Three N-application rates (low, medium, and high), each replicated three times, were applied to corn in 1996 and 1998 under randomized complete block design at a tile-drained corn-soybean rotation field in central Iowa. Model calibration and evaluation was based on field measurements of tile flows, NO3-N losses with tile water, and corn-soybean yields. On the average, the model predicted tile flow, NO3-N losses with tile water, and yields adequately by showing a percent difference of -7%, 21%, and -2%, respectively, between measured and dpredicted values. The model successfully incorporated the climatic effects, and predicted corn yields were similar to the measured values in 1998 (validation year) despite lower N-fertilizer rates applied in 1998 than in 1996. Predicted corn yields remained constant when N-application rate exceeded 170 kg-N/ha in 1998. Model predictions of NO3-N losses with tile water as a function of N-application rates were better in 1998 than in 1996, probably due to better estimation of initial conditions in response to continuous model simulations from January 1, 1996 through December 31, 1999. However, the model overestimated NO3-N losses with subsurface drainage water during the soybean growth period, which may require further refinements in the N-cycling algorithm in relation to N2-fixation and N-uptake processes.