Crowd-sourced trait data can be used to delimit global biomes

Abstract

Terrestrial biomes and their biogeographic patterns have been derived from a large variety of variables including species distributions and bioclimate or remote sensing products. However, classifying the biosphere into biomes from a functional perspective using biophysical traits has rarely been tested. Such a trait-based biome classification has been limited by data availability. Here, we aimed to exploit crowd-sourced plant observations and trait databases to systematically assess which traits are most suitable for biome classification. We derived global patterns of 33 biophysical traits covering around 50 % of the land surface by combining crowd-sourced species distribution data from the Global Biodiversity Information Facility (GBIF) and trait observations from the TRY database. Using these trait maps as predictors for supervised cluster analyses, we tested to what extent we can reconstruct 31 published biome maps. A sensitivity analysis with randomly sampled combinations of traits was performed to identify the traits that are most appropriate for biome classification. Performance was quantified by comparing modeled biome maps and the respective observation-based biome maps. Finally, spatial gaps in the resulting biome maps were filled using species distribution models to obtain continuous global biome maps. We showed that traits can be used for biome classification and that the most appropriate traits are conduit density; rooting depth; height; and different leaf traits, including specific leaf area and leaf nitrogen content. The best performance of the biome classification was obtained for biome maps based on biogeographic zonation and species distributions, in contrast to biome maps derived from optical reflectance. The availability of crowd-sourced plant observations is heterogeneous, and, despite its exponential growth, large data gaps are prevalent. Nonetheless, it was possible to derive biome classification schemes from these data to predict global biome patterns with good agreement. Therefore, our analysis is a valuable approach towards understanding biome patterns based on biophysical traits and associated ecological strategies.