In Brazil, nearly half of the country’s eucalyptus plantations are located in regions where seasonal dry periods lasting three to five months bring monthly rainfall below 50 mm. Thus, the main goal of this study was to evaluate the potential benefits of incorporating multi-year and multi-environmental data, as well as to investigate the impact of modeling different covariance structures on the selection of drought-tolerant families, and thus compare the results with the standard analysis model framework used in forest breeding. For this purpose, we evaluated the diameter at breast height (DBH, cm) in 232 Eucalyptus families and six commercial clones, installed in different experimental trials distributed in three distinct locations in Brazil. Three measurements were performed at 18, 30, and 42 months after planting. Finally, we identified promising families for cultivation under drought conditions. Our results show that spatial row–column corrections improved model fit (as indicated by lower AIC values) in single-environment analyses and were therefore adopted for subsequent multi-environment and multi-age analyses. Among the selected best-fitting models, the heterogeneous compound symmetry covariance structure effectively accommodated the data complexity for predicting interaction effects, while diagonal first-order heterogeneous autoregressive structures were more suitable for modeling residual effects. In this context, the selected model exhibited significant variability among families in drought tolerance. Models that incorporated multi-year and multi-environmental data and adjustments in the variance-covariance structures for genetic and residual effects allowed for a better fit to the data, resulting in more consistent variance component estimates and increasing the accuracy of genetic and spatial interaction assessments. Furthermore, the selected families promise to continue regional breeding programs to develop drought-tolerant eucalyptus. Our findings underscore the potential of Brazilian eucalyptus germplasm for drought-tolerance improvement and highlight the importance of robust data integration in analyzing trials under abiotic stress.
Canal et al. (Fri,) studied this question.