Using the process-based JULES-crop model for forecasting off-season maize yield in Brazil

• Ano de defesa: 2023
• Autor(a) principal: Prudente Junior, Amauri Cassio
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Zea mays L.; Calibração de modelos; Climate change; JULES-crop; Models calibration; Mudanças climáticas; Previsão de safra; Yield forecasting
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/11/11152/tde-23032023-111110/

Maize (Zea mays L.) is an important Brazilian commodity, being the second most produced crop and the fifth most exported in Brazil. In view of its relevance for many sectors of the economy, studies that deepen the consequences of climatic effects are imperatives in face of a climate change scenario for the next decades. For this purpose, process-based biophysical models has been used to evaluate the weather effects on crop yield. However, there is a gap in the science of models able to perform in large-scale due to limitations in the integration of energy, CO2, water and momentum fluxes with crop physiology. In view of this lacuna, the land surface model Joint UK land environment simulator (JULES) was integrated with a parametrization of different crops, among which maize, however, the model was not calibrated and evaluated in Brazil. This thesis brings in two chapters the use of a large-scale model in maize and its application to predict the off-season maize yield in Brazil. In the first chapter, the objective was to calibrate and evaluate the JULES-crop model for maize, obtaining a high performance to simulate leaf area index (LAI), canopy height and grain dry mass both for irrigated or rainfed conditions, in different regions of Brazil and sowing dates. In the second chapter, it was possible to use the calibrated JULES-crop, in addition to agro- climatic indicators such as air temperature, rainfall and diffuse radiation, to develop a large scale yield forecasting model for off-season maize in Brazil. The conjunction of agro-climatic indicators and JULES-crop outputs resulted in high performance predictions for maize yield from the 80th day of the cycle. Therefore, it is possible to confirm a skillful model to simulate in a large scale, and that it is able to improve the forecasting for maize yield in Brazil.