Using Encyclopedia of Life’s TraitBank to identify plant traits associated with vulnerability

Certain biological traits seem to predispose some species to greater extinction risk than others and, when vulnerability information is limited, could be used as proxies to identify understudied species likely in need of protection. In the past, identifying broadly applicable traits associated with extinction risk has been hampered by the difficulty of collecting information for a broad range of species (both geographically and taxonomically), with most comparative analyses focusing on regional and/or taxon specific patterns. However, efforts to collect and compile existing trait information from regional and taxon specific datasets into a single repository are making it possible to analyze patterns between traits and vulnerability on ever broader scales. We compared trait information from one such repository, the Encyclopedia of Life (EOL) TraitBank, to information on threat status from the International Union for the Conservation of Nature (IUCN) Red List to determine whether such consolidated data can help either clarify previously identified associations or identify new associations between plant traits and threat risk. Using generalized linear mixed effects models (GLMM’s) we found five plant traits that could be used to predict whether a species is threatened or not: ‘plant growth form’, ‘life cycle habit’, ‘low temperature tolerance’, ‘soil depth’ and ‘foliage porosity in winter’. Threatened plant species tend to be trees rather than shrubs or grasses, live for more than one season (i.e. perennials), be less tolerant of cold temperatures, require deep soil for good growth, and have less dense foliage in the winter. Both ‘plant growth form’ and ‘life cycle habit’ have been identified as correlates of vulnerability in past studies and both are relatively easy to use as diagnostic characters, which is ideal for identifying understudied species of potential conservation concern. However, how these two traits relate to vulnerability seems context dependent, with opposite relationships in past regional or taxon specific studies. Unfortunately we could not identify further traits that could add explanatory power to those relationships with the existing data on EOL’s TraitBank but the repository is still growing and the inclusion of additional traits and/or species from other datasets could clarify these relationships further.