Abstract The cultivation of common bean in Brazil typically occurs in three annual cropping seasons across the country’s major geographic regions, exposing genotypes to contrasting environmental conditions across time (seasons and years) and space (locations). With this aim, this study investigated the influence of the G × E interaction on common bean yield across different sites, seasons, and years. It evaluated the contributions of fixed and random effects. Data from 424 multi-environment trials (METs) conducted by Embrapa between 2011 and 2018 were used, involving 87 genotypes across three cropping seasons (Wet, Dry, and Winter) distributed over 71 locations. Genetic and environmental effects were estimated through linear mixed models fitted with the REML/BLUP method. In addition, multivariate analyses, including the GGE Biplot, were used to decompose and visualize G × E effects, while missing data were imputed via Principal Component Analysis (PCA). Environmental stratification and the identification of mega-environments enabled grouping sites with similar characteristics based on the presence or absence of G × E interactions. The stability and adaptability analysis of the cultivars, based on a ranking that considered the particularities of the PRVG, MHPRVG, Lin and Binns, Wricke’s Wi, and Finlay–Wilkinson indices, revealed distinct patterns of behavior across the three seasons.
Cruz et al. (Mon,) studied this question.