Variety-specific sugarcane yield simulations and climate change impacts on sugarcane yield using DSSAT-CSM-CANEGRO model

Crop simulation models are still little used for sugarcane crops due to the lack of understanding of their capabilities and experience in calibration as compared to other crops. Realistic assessment of future environmental change effects on crop production is also necessary for successful agricultural management and planning. The objective of this study is to provide DSSAT-CANEGRO simulated variety-wise sugarcane yield models for twelve different locations of Muzaffarnagar District of India. The CANEGRO model calibration is performed under three different dates of planting (early, mid, late) during the spring season 2013–2014. The results of the models are validated for consecutive two years datasets (2014–15 and 2015–16). The model shows the best calibration and validation results under the mid planting date scenario (simulated yield +0.85 %, +2.80 % and +5.20 %). The impact of climate change (sensitivity analysis) on yield of sugarcane has also been made putting different values of Tmax ( ± 1 to ± 3 ºC), Tmin ( ± 1 to ± 3 ºC), solar radiation ( ± 1 to ± 3MJ/m2/day), and atmospheric CO2 concentration of 380 ppm (720 ppm scenario A2, 500 ppm scenario B2). The study highlighted that Sugarcane yield simulation mid-planting model presenting the highest R (0.81, 0.83), and D (0.88, 0.91), and the lowest errors (RMSE = 8.37 q ha-1, 10.70 q ha-1 and MAPE = 1.10 %, 1.30 %) for the years 2014–15 and 2015–16 as compared to the other two models. The CANEGRO model simulated yield under incremental values of Tmax and Tmin in the range of (+1 to +3 ºC) shows gradual decrement in the yield ranges while, gradual increment of solar radiation from 1 to 3 MJ/m2/day showed a yield increment. •Our studies investigated the simulation of sugarcane yield under three different dates of planting using the CANEGRO model.•Data on growth, development, and yield are computed for five sugarcane varieties at twelve different locations.•Model simulates the effect of physical-chemical properties of soil, such as texture information, pH, OC%, N%, and CEC.•Model calibration is done using one year datasets and validation is performed for the two consecutive year datasets.•Model’s accuracy assessment is done using statistical parameters (R, R2, RMSE, MAPE, and Willmott index of agreement).•Model sensitivity to change climatic parameters (Tmax, Tmin, solar radiation and atmospheric CO2 concentration).•Complexity in climatic pattern, due to change in individual parameters and in combination wit