Crop model validation and sensitivity to climate change scenarios

Field measurements of land surf ace-atmosphere heat and water exchanges, leaf area index, crop height, dry matter accumulation, and crop yield at Bondville, an agricultural site of the AmeriFlux network located in the USA Midwest, were used to evaluate the performance of the process-oriented crop model Agro-IBIS under a corn (Zea mays L.)-soybean (Glycine max (L.) Merr.) crop rotation. Simulations of drainage and nitrate/nitrogen leaching and concentration are also discussed. The Simulated fertilization effect of carbon dioxide (CO2) on corn and soybean crops was evaluated against results from a recent synthesis of Free Air-CO2 Enrichment (FACE) experiments. Two contrasting climate change projections from the Canadian Global Climate Model v.2 (CGCM2) and from the Hadley Centre coupled Model v.3 (HadCM3) were further used to assess the sensitivity of Agro-IBIS simulations of corn and soybean yields to potential climate perturbations. Overall, our results clearly show model strengths and allowed us to identify some model deficiencies. The model was able to reproduce net radiation, latent and sensible beat fluxes within 11, 17 and 2%, of observations, respectively. Yields of corn and soybean were underestimated on average by 11 and 10%, respectively. In leaf area index, the model was able to simulate its peak within < 10% of observation, but had difficulty accurately simulating its development during the period of leaf senescence. Our results suggest that Agro-IBIS needs to be improved with regard to its simulation of the effect of rotation on corn and soybean yields, as well as its estimation of the fertilization effect of elevated atmospheric CO2 concentration on soybean yield.