Aim: Vascular epiphytes are ubiquitous components of wet tropical forests where they contribute substantially to local and regional plant diversity. While some basic epiphyte distribution patterns are relatively well studied, little effort has been made to understand the drivers responsible for constraining their global distribution. This study quantifies the substantial contribution of epiphytes to global gradients and centres of vascular plant diversity and explores whether epiphytes vary from terrestrial plants in relation to contemporary and historical environmental variables.

Location: Global.

Time period: Present.

Major taxa studied: Tracheophyta.

Methods: Using a comprehensive epiphyte species list (EpiList 1.0), and distribution information for 27,850 epiphyte species derived from numerous literature sources, we describe the global biogeography of epiphytes. We used generalized linear mixed effects models to assess the relationship between epiphytic and terrestrial plant diversity, and contemporary and historical environmental predictors. 

Results: We find that epiphytes substantially contribute to global centres of vascular plant diversity, accounting for up to 39% of the vascular flora in Neotropical regions. Epiphytes decrease in species numbers with increasing latitude at a rate three times faster than terrestrial plants, a trend that is driven mainly by the distribution of tropical forests and precipitation. Further, large regional differences emerge that are explained by several large endemic angiosperm families (e.g., Neotropical Bromeliaceae) that are absent in other tropical regions.

Main conclusions: Our results show that epiphytes are disproportionately diverse in most global centres of plant diversity and play an important role in driving the global latitudinal diversity gradient for plants. The distribution of precipitation and tropical forests emerge as major drivers of the latitudinal diversity gradient in epiphyte species richness. Finally, our findings demonstrate how epiphyte floras in different biogeographical realms are composed of different families and higher taxa revealing an important signature of historical biogeography.

We related epiphyte richness to 10 contemporary and past environmental variables out of 19 variables that were initially considered. Variables included three contemporary climatic variables derived from CHELSA V1.2 (Karger et al., 2017) – mean annual precipitation (mm, hereafter precipitation), precipitation seasonality (coefficient of variation in precipitation), and mean daily minimum temperature (°C, hereafter temperature), all of which have been previously hypothesised or shown at regional to global scales to be important predictors of both epiphytic and terrestrial plant species richness (Zotz, 2005; Gentry & Dodson, 1987; Kreft & Jetz, 2007; Taylor et al., 2021). As a measure of habitat availability for epiphytes, we included the contemporary extent of tropical forest biomes (km²). Contemporary tropical forest biomes were extracted from a global map of terrestrial biomes (Olson et al., 2001), and overlaid with our botanical country polygons in order to quantify the total area of tropical forest for each botanical country. In addition, we selected area (km², Weigelt et al., 2020) and elevational range (m, from the Global Multi-resolution Terrain Elevation data set by Danielson & Gesch, 2011 at a resolution of 30 arc-seconds), which are important geographical predictors of species diversity.

We further considered three historical factors, reflecting past climate – Last Glacial Maximum (LGM) ice cover (km², Ehlers et al., 2011), and past distribution of tropical forests – LGM tropical forest area (km², Ray & Adams, 2001) and tropical forest area during the Mid-Miocene climate optimum (km², Henrot et al., 2010). The extent of both historical tropical forest biomes was quantified in the same manner as for the distribution of contemporary tropical forested biomes. However, because biome definitions differed between datasets, we first standardised all biomes to match across datasets, delineating “tropical rainforest”, “sub-tropical forest”, and “tropical seasonal forest” to “tropical forest”. Finally, we explore continental differences in epiphyte occurrences, which allow inferences about idiosyncratic historical biogeographic patterns not captured by the environmental predictors (variable biogeographic realm). Similar to the classification of biomes, each region was assigned its respective realm following Olson et al., (2001), in order to explore the richness and relative contribution of different epiphyte families among the different continents.

All references and extended methods can be found in our accompanying manuscript GEB-2021-0338.