Abstract

Aim We mapped global patterns of tree phylogenetic endemism (PE) to identify hotspots and test hypotheses about possible drivers. Specifically, we tested hypotheses related to current climate, geographical characteristics and historical conditions and assessed their relative importance in shaping PE patterns. Location Global. Time period We used the present distribution of trees, and predictors covering conditions from the mid-Miocene to present. Major taxa studied All seed-bearing trees. Methods We compiled distributions for 58,542 tree species across 463 regions worldwide, matched these to a recent phylogeny of seed plants and calculated PE for each region. We used a suite of predictor variables describing current climate (e.g., mean annual temperature), geographical characteristics (e.g., isolation) and historical conditions (e.g., tree cover at the Last Glacial Maximum) in a spatial regression model to explain variation in PE. Results Tree PE was highest on islands, and was higher closer to the equator. All three groups of predictor variables contributed substantially to the PE pattern. Isolation and topographic heterogeneity promoted high PE, as did high current tree cover. Among mainland regions, temperature seasonality was strongly negatively related to PE, while mean annual temperature was positively related to PE on islands. Some relationships differed among the major floristic regions. For example, tree cover at the Last Glacial Maximum was a positive predictor of PE in the Palaeotropics, while tree cover at the Miocene was a negative predictor of PE in the Neotropics. Main conclusions Globally, PE can be explained by a combination of geographical, historical and current factors. Some geographical variables appear to be key predictors of PE. However, the impact of historic and current climate variables differs considerably among the major floristic regions, reflecting their unique histories. Hence, the current distribution of trees is the result of globally relevant geographical drivers and regional climatic histories.