Assessment of winter wheat production under early sowing with supplemental irrigation in a cold highland environment using CropSyst simulation model

The most important question in supplemental irrigation management is when and how much water to apply. It is a laborious and expensive task to develop supplemental irrigation schedules solely by conventional field experimentation. A cropping system simulation model (CropSyst) was evaluated for its ability to simulate growth, biomass, grain yield and evapotranspiration (ET) of wheat sown early with supplemental irrigation (SI). This was computed under rainfed conditions in a semiarid climate with cold winters in the highlands of Turkey. Experimental data from three growing seasons (1998–2001) were used. The experimental design incorporates Bezostia bread wheat cultivar tested under two main treatments: no irrigation at sowing (rainfed-A1) and irrigation at sowing with 50 mm of water (A2); and two sub treatments: rainfed (B1) and spring supplemental irrigation to replenish the total water requirement at 0–90 cm soil profile (B2) at the Ankara Research Institute of Rural Services. Crop input parameters were selected from the model documentation and experimental data. A few cultivar-specific parameters were adjusted within a narrow range of typical fluctuations by model calibration. Results showed that CropSyst was able to simulate yield, biomass and evapotranspiration as observed in the field experiments. Overall, the Willmott Index of agreement between simulated and observed values of grain yield, biomass and ET were 0.98, 0.76 and 0.91, respectively. CropSyst model predicted better the seasonal evapotranspiration under full supplemental irrigated conditions (A2B2) than under rainfed conditions (A1B1), with values of the Willmott index of agreement being 0.97 and 0.89, respectively. The model was run for 20 years (1982–2001) including the 4-year experimental period. Data showed that wheat grain yield could be improved by 15, 19 and 25% with applying only 50 mm of water at the sowing time of 15 October, 1 October and 15 September, respectively. In 80% of the cases, the respective SI applications would give 2.75, 2.7 and 2.95 t ha−1, of the long-term average rainfed yield of 2.1 t ha−1, respectively.