The overall performance of the Decision Support System for Agrotechnology Transfer-Cropping System Model (DSSAT-CSM) was evaluated for simulating wheat (Triticum aestivum L.) yield, grain N uptake, soil organic C (SOC) and N (SON), soil water and nitrate-N (NO3-N) dynamics. The data used was from a long-term (1967-2005) spring wheat experiment conducted at Swift Current, Saskatchewan in the semi-arid Canadian prairies. Four treatments were selected: (1) continuous wheat receiving N and P fertilizer, Cont-W(NP); (2) continuous wheat receiving P only, Cont-W(P); and each phase of a fallow wheat rotation receiving N and P fertilizer, (3) W-F(NP) and (4) F-W(NP). The simulated grain yields matched the measurements well, with high d (0.74-0.83) and EF (0.16-0.33). The grain N uptake was also simulated satisfactorily with RMSE of 14-17 kg N ha(-1) and d of 0.66-0.81. DSSAT simulated topsoil (0-0.15 m) SOC and SON well in the drier period (1967-1991), whereas it underestimated SOC in the more humid period (1991-2003). The DSSAT successfully simulated soil water and NO3-N dynamics in 0-0.15 m depth, whereas it overestimated soil water and NO3-N in the deep layers and consequently underestimated NO3-N leaching, suggesting that further improvements in soil water module should be made for the semi-arid climatic conditions in Canadian prairies. Sensitivity results showed that soil water content was sensitive to both lower soil water and upper drainage limits in this study. The performances of DSSAT model to yield and soil dynamics were comparable with other models.
