Climate, topography and soil factors interact to drive community trait distributions in global drylands

The skewness and kurtosis of community trait distributions (CTDs) can provide important insights on the mechanisms driving community assembly and species coexistence. However, they have not been considered yet when describing global patterns in CTDs. We aimed to do so by evaluating how environmental variables (mean annual temperature [MAT] and precipitation [MAP], precipitation seasonality [PS], slope angle and sand content) and their interactions affected the mean, variance, skewness, kurtosis of the plant CTDs in global drylands. We gathered specific leaf area and maximum plant height data from 130 dryland communities from all continents except Antarctica. Over 90% of the studied communities had skewed CTDs for SLA and height or had kurtosis values differing from those of normal distributions. Higher MAT and/or lower MAP led to a shift toward plant communities over-represented by “conservative” strategies, and a decrease in functional diversity. However, considering interactions among environmental drivers increased the explanatory power of our models by 20%. Sand content strongly altered the responses of height to changes in MAT and MAP (climate × topo-edaphic interactions). Increasing PS reversed the effects of MAT and MAP (climate × climate interactions) on the four moments of CTDs for SLA, particularly in dry-subhumid regions. Our results indicate that the increase in PS forecasted by climate change models will reduce the functional diversity of dry-subhumid communities. They also indicate that ignoring interactions among environmental drivers can lead to misleading conclusions when evaluating global patterns in CTDs, and thus may dramatically undermine our ability to predict the impact of global environmental change on plant communities and associated ecosystem functioning.