Regionalizing crop varieties improves modeling of crop production and greenhouse gas emissions with process-based ecosystem model

Abstract:

Agriculture is a significant source for greenhouse gases (GHGs), and process-based models can be applied to estimate and evaluate global emissions from cropland management activities. Accurately simulating crop production is critical for process-based models' performance and determines the amount of plant residue and exudate production and crop uptake of fertilizers, influencing soil carbon storage, CH₄, and N₂O emissions. However, studies modeling global agricultural GHG emissions often use a single crop variety in global assessments, implying that major crops are identical across all regions of the world. This study applied a Bayesian approach to calibrate the DayCent ecosystem model to produce regional crop varieties for corn, soybean, rice, and wheat. We used a global crop production dataset between 2001 and 2013 at 0.5° resolution to calibrate and evaluate the model. We selected major cropping regions from the FAO Global Agro-Environmental Stratification as a basis for the regionalization and identified each crop's most important parameters through a global sensitivity analysis. We calibrated DayCent using the sampling importance sampling algorithm and produced posterior distributions for the parameters. We found significant improvement in DayCent simulations of crop yields and GHG emissions with calibrated parameters for regional varieties of the major crops. This study highlights the importance of representing regional variation in crop types for achieving accurate predictions of crop yields and agricultural GHG emissions.