Simulation And Evaluation Of DSSAT Model Under Water And Nitrogen Management In Crop Production System

Nitrogen is a major limiting factor for crop production under normal field condition. Irrational irrigation and fertilization usually lead to resources waste, yield reduction, product quality decrease, and environment pollution. Therefore, it is important to develop best management practice to improve nutrient and water managements for sustained crop production and healthy environment. In this study, based on field experiment data, Decision Support System for Agrotechnology Transfer (DSSAT) Crop System Models were used to simulate different field experiments, and then the DSSAT models simulation performance and the model adaptability to the experiment condition were evaluated. The main findings are as following:1. From 1959 to 2008, A 50 year long-term experiment was carried out in southern Canada with continuous maize with, with fertilization (CC-F) and without fertilization (CC-NF) treatments. Study on maize cultivar coefficients calibration in the simulation indicated that radiation use efficiency (RUE) should considered in maize cultivar coefficient calibration to be done for different cultivars, experiments and areas. In this study, 50 years experiments were divided into 4 periods. Based on the 4 periods, 4 set of maize cultivars were defined, and the cultivar coefficients were calibrated. Using these calibrated coefficients, the maize yields of CC-F and CC-NF could be better simulated, compared with that using only one set of cultivar coefficient in 50 year.From the simulation results, DSSAT could simulated the yields of CC-F (with n-RMSE = 39%, R~2 = 0.36) more precisely than that for CC-NF (with n-RMSE = 82%, R~2 = 0.40). The model did not provide accurate predictions for annual maize yields. However, the model can simulate the yield change trend well. For the CC-F and CC-NF treatments, there were marked differences between simulated soil mineral nitrogen content(0-15 cm and 15-30 cm)of and the actual measured data, with the n-RMSE for CC-F ranged from 58% to 60%, and that for CC-NF ranged from 64% to 89%.The simulated cumulative nitrate loss from CC-F treatment from 1998 to 2000 by DSSAT model was consistent with the actual measured values, with n-RMSE为29%,EF为0.63. However, For the CC-NF treatment, the DSSAT model overestimated the cumulative nitrate loss, relative to the actual measured values, with n-RMSE of 160%,EF of 0.04. This indicates that the DSSAT model could not well simulate soil nitrogen situation in the nitrogen depleted soil after long term nitrogen depletion without nitrogen fertilizer application.2. From 2000 to 2004, in southwestern Ontario, Canada, a 5-year field experiment was carried out to investigate the effect of regular free tile drainage (TD) and controlled tile drainage with optional subsurface irrigation (CDS) on crop yield, soil water content and nitrate loss. The simulation results showed that DSSAT model could simulate the maize and soybean yield excellently (n-RMSE:4.3%～14.0%,d:0.985～0.998); well simulate soil water content (0-30cm) in root zone (n-RMSE:9.9%～14.8%,d: 0.724～0.831);reasonably good to simulate nitrate loss from tile drainage (n-RMSE: 17.8%～25.2%,d: 0.529～0.979). Thereby, DSSAT can be a useful tool for simulating the effects of TD and CDS on environmental quality, crop productivity, and soil profile processes (5-8 years).3. Maize experiments were conducted on a black soil in Gongzhuling, Jilin, Northeast China during 2006 and 2007 to study the effect of different fertilizer application on maize growth, nitrogen uptake and yield. After calibrating maize cultivar coefficients, radiation use efficiency(RUE)and soil fertility factors (SLPF), the DSSAT model could simulate aboveground biomass (such as d = 0.95～0.98 in the calibrated year), yield(n-RMSE = 4.6%～9.0%;d = 0.68～0.95)very good, but it could not well simulate maize nitrogen uptake. By doing sensitive analysis on nitrogen content change coefficient, b, it showed that DSSAT model could simulate maize nitrogen uptake more precisely after b value adjustment. Therefore, The DSSAT 4.5 model is a useful decision support tool to help decision makers, scientists and farmers to optimize maize management using different fertilizer N application strategies.