Which traits predispose species to extinction? A review

Abstract

Biodiversity is shrinking rapidly and despite our efforts, only a small part of it has been redlisted. Identifying the traits that make species vulnerable might help us predict the outcome for those less known. We used machine learning algorithms to filter relevant publications among 2700 potential ones and collected a final list of 559 statistical models within 122 publications, from which we gathered information on trait responses to extinction risk, across all taxa, spatial scales and biogeographical realms, in what we think it is the most complete compilation up to date. This talk will explore gaps in the research: are taxa, or biogeographical regions equally sampled? Then we will answer to the question in the title by identifying which traits have been tested and which of those have been successful in explaining extinction risk patterns. Our meta-analyses reveal that habitat breadth and geographical range size are always negatively related to extinction risk, but body size, even taking the taxonomical group into account, does not show a clear unidirectional response. In the end, we will propose an alternative way to understand which traits the conservationists should be looking for, without having to deal with the lack of data and methodological constraints that current statistical models impose.