Calibration and validation of the DNDC model to estimate nitrous oxide emissions and crop productivity for a summer maize-winter wheat double cropping system in Hebei, China

The main aim of this paper was to calibrate and evaluate the DeNitrification-DeComposition (DNDC) model for estimating N₂O emissions and crop productivity for a summer maize-winter wheat double cropping system with different N fertilizer rates in Hebei, China. The model's performance was assessed before and after calibration and model sensitivity was investigated. The calibrated and validated DNDC performed effectively in estimating cumulative N₂O emissions (coefficient of determination (1:1 relationship; r²) = 0.91; relative deviation (RD) = -13 to 16%) and grain yields for both crops (r² = 0.91; RD = -21 to 7%) from all fertilized treatments, but poorly estimated daily N₂O patterns. Observed and simulated results showed that optimal N fertilizer treatment decreased cumulative N₂O flux, compared to conventional N fertilizer, without a significant impact on grain yields of the summer maize-winter wheat double cropping system. The high sensitivity of the DNDC model to rainfall, soil organic carbon and temperature resulted in significant overestimation of N₂O peaks during the warm wet season. The model also satisfactorily estimated daily patterns/average soil temperature (^o C; 0-5 cm depth) (r² = 0.88 to 0.89; root mean square error (RMSE) = 4 °C; normalized RMSE (nRMSE) = 25% and index of agreement (d) = 0.89-0.97) but under-predicted water filled pore space (WFPS; %; 0-20 cm depth) (r² = 0.3 to 0.4) and soil ammonium and nitrate (exchangeable NH₄⁺ & NO₃^-; kg N ha^-1; r² = 0.97). With reference to the control treatment (no N fertilizer), DNDC was weak in simulating both N₂O emissions and crop productivity. To be further improved for use under pedo-climatic conditions of the summer maize-winter wheat double cropping system we suggest future studies to identify and resolve the existing problems with the DNDC, especially with the control treatment.