Fig 1. Variation partitioning of the three-diversity metrics (SD, FD and PD) explained by abiotic, present-day and historical variables.

Variance partitioning was used to calculate the relative importance of each predictor set (using PCA axis as explanatory variables for abiotic, present-day and historical data) either alone or in combination based on GLM models. The black bars indicate the results for the old patches and the light grey bars indicate the results for the new patches. Abi, abiotic variables; Pre, present-day variables; His, historical variables; Conjoint, the effect of all predictors set together (i.e. abiotic + present-day + historical variables).