Application of the DSSAT Model to Simulate Wheat Growth in Eastern China

In order to test the applicability of the DSSAT model in the Chaohu Lake area, an important drinking water catchment of Anhui Province, the model was calibrated based on three year field experiments (2007-2010). Calibrations were based on wheat growth stages, leaf area index (LAI) and yield. The model was used to simulate the effect of different sowing dates and sowing density on wheat yield and the effect of nitrogen fertilizer level on wheat yield and nitrogen losses. The simulation results for the three years 2007 to 2010 agreed well with the measured data from the wheat growth experiment with Nash-Sutcliffe model efficiencies of 0.95 (growth stages), 0.85 (LAI) and 0.92 (yield). These results indicate that plant growth development and yield can be simulated efficiently for the conditions at the experimental site. Simulations with different sowing dates and planting densities showed that early sowing dates correspond with relatively low sowing densities while later sowing dates correspond to medium sowing densities. Compared with the usual sowing date and sowing density for wheat in the experimental region the results indicate that there may be some possible yield increase (and saving of sowing material) with lower densities than actually applied. The simulated N losses according to the model are largely determined by ammonia volatilization and denitrification whereas the simulated losses due to leaching are negligible. A future optimization strategy for fertilization should focus on the type of N fertilizer (urea versus other N fertilizers), the pH value of the soil and tillage and irrigation measures to reduce ammonia volatilization.