Pairwise comparisons of significant variation for each bioclimatic variable, component and drought severity estimator in relation with population structure (p-value<0,001).

Variable abbreviations: P1–P19: main variables as defined in table 1, DIT: Annual Mean Drought Index (Thornthwaite), DIT, Max: Maximum Drought Index (Thornthwaite), DIH: Annual Mean Drought Index (Hamon), DIH, Max: Maximum Drought Index (Hamon), F1P, F2P: Main two factors for bioclimatic precipitation variables (P12–P19), F1S, F2S: Main two factors for drought-related variables.Kruskal-Wallis tests were applied in all cases except for DI (Drought Index) estimations, where an ANOVA followed by a Tukey’s-b post-hoc test was used. A, B and C are different ranks. Populations with more than one letter could not be assigned to a single rank. Mean for each variable for each population: A>B>C.Bold variables: Drought-related variables. Selected variables for further analysis based on their conceptual power to describe drought tolerance.