For more than 50 y ecologists have believed that spatial heterogeneity in habitat conditions promotes species richness by increasing opportunities for niche partitioning. However, a recent stochastic model combining the main elements of niche theory and island biogeography theory suggests that environmental heterogeneity has a general unimodal rather than a positive effect on species richness. This result was explained by an inherent tradeoff between environmental heterogeneity and the amount of suitable area available for individual species: for a given area, as heterogeneity increases, the amount of effective area available for individual species decreases, thereby reducing population sizes and increasing the likelihood of stochastic extinctions. Here we provide a comprehensive evaluation of this hypothesis. First we analyze an extensive database of breeding bird distribution in Catalonia and show that patterns of species richness, species abundance, and extinction rates are consistent with the predictions of the area–heterogeneity tradeoff and its proposed mechanisms. We then perform a metaanalysis of heterogeneity–diversity relationships in 54 published datasets and show that empirical data better fit the unimodal pattern predicted by the area–heterogeneity tradeoff than the positive pattern predicted by classic niche theory. Simulations in which species may have variable niche widths along a continuous environmental gradient are consistent with all empirical findings. The area–heterogeneity tradeoff brings a unique perspective to current theories of species diversity and has important implications for biodiversity conservation.

The thresholds of drought duration and intensity required to provoke pulses of tree mortality across Earth's biomes remain unclear. Using globally-extensive updated databases of drought-associated tree mortality, we report substantial diversity in the types of drought events that cause tree death in different forest types. Tree-killing droughts are longer, more intense and have higher completeness (proportion of extreme drought within long-lasting droughts) in dry versus wet biomes. Mortality-inducing droughts are more intense and show higher completeness in angiosperm-dominated forests. We find a marked tendency towards long-lasting and more severe and complete droughts in recent years, particularly in more arid sites. Warming-amplified aridity is a main factor underpinning these variations. Differences in "sampling effort" across regions make it challenging to characterize the high variability in drought-induced tree mortality events. In this work we demonstrate the need to create, continuously update, and refine more extensive field-based tree mortality monitoring programs globally.