Equationing Arabica coffee: Adaptation, calibration, and application of an agrometeorological model for yield estimation

Coffee cultivation is important to Brazil's economy, positioning the country as a global leader in production and export. Given the complex environmental and management factors affecting yields, particularly due to climate change , there is a pressing need from farmers and dealers for more precise crop estimation models. This study aimed to refine and calibrate an agrometeorological model, originally developed by Santos and Camargo (2006) and later adapted by Verhage et al. (2017a), to estimate Arabica coffee yield in the main producing regions of Minas Gerais and São Paulo. Additionally, sensitivity analysis was also performed to identify the most influential model parameters and variables. Yield data from 28 coffee-producing locations (2003−2020) and meteorological data alongside irrigation use were employed. Following calibration and adaptation, a sensitivity analysis was conducted to determine the model's response to variations in coffee plant parameters and environmental conditions. Local sensitivity analysis (LSA) focused on meteorological variables, while global sensitivity analysis (GSA) addressed coffee-related parameters. The adaptations proposed to the original model led to a significant refinement in the yield estimates, emphasizing the complex interactions between climatic variables and agricultural management practices. Key adaptations include the estimation of potential yield (Yp), the incorporation of temporal curves for root growth , leaf area index, available water capacity , and crop coefficient , as well as a water balance that accounts for irrigation and its effect on attenuating high canopy temperatures. Calibration improved the model's accuracy and precision, with the RMSE decreasing from 13.66 (819.6 kg ha −1 ; 1 bag ha −1 = 60 kg ha −1 ) to 8.65 (519.0 kg ha −1 ) bags ha −1 , R 2 improving from 0.62 to 0.65, d-index from 0.79 to 0.88, and NSE from 0.09 to 0.64. During the evaluation phase, with independent data, RMSE was 7.76 bags ha −1 (465.6 kg ha −1 ), d-index 0.85, and R 2 0.55. Sensitivity analysis emphasized the importance of mean temperature and solar radiation on Yp, as well as the impact of irrigation practices and water deficit management under rainfed conditions. Additionally, factors specific to the coffee plant itself directly affect its yield. The findings underscore the importance of a multifactorial and adaptive approach to coffee cultivation, addressing the complexities and challenges posed by varying climatic conditions. This work offers valuable insights into optimizing coffee production, presenting the model as a tool for developing more resilient cultivation strategies and enhancing the sustainability of Brazilian Arabica coffee in future climate change scenarios. • Adaptations made to the model allowed for a refinement in the yield estimates. • Calibration improved accuracy and precision of Arabica coffee yield estimation. • Local sensitivity analysis highlighted temperature, radiation and precipitation. • Global sensitivity analysis emphasized coffee parameters and their impact on yield. • Model captured coffee's biennial production and stable yields with irrigation.