Calibration and evaluation of JULES‐crop for maize in Brazil

Abstract

AbstractMaize (Zea mays L.) is a prominent Brazilian commodity, being the second largest crop produced and fifth exported product by the country. Due to its importance for the agricultural sector, there is a concern about the effect of climate change on the crop. Process‐based models are valuable tools to evaluate the effects of climate on crop yields. The Joint UK Land Environment Simulator (JULES) is a land‐surface model that can be run with an integrated crop model parameterization. The resulting model (JULES‐crop) thus integrates crop physiology principles with the complexity of atmosphere–biosphere coupling. It has been shown to be a valuable tool for large‐scale simulations of crop yields as a function of environmental and management variables. In this study, we calibrated JULES‐crop using a robust experimental dataset collected for summer and off‐season maize fields across Brazil. A targeted local sensitivity analysis was performed to detect parameters of major importance during the calibration process. After calibration, the model was able to satisfactorily simulate both season and off‐season cultivars. Modeling efficiency (EF) was high for leaf area index (EF = .73 and .71, respectively, for summer season and off‐season datasets), crop height (EF = .89), and grain dry mass (EF = .61 and .89, respectively, for summer season and off‐season datasets). The model showed a lower accuracy for simulating leaf dry mass in summer season cultivars (EF = .39) and soil moisture (EF = .44), demonstrating the necessity of further improvements including additional parametrizations of the rainfed conditions.