Climate Change-Resilient Conservation of Cacti

Abstract

Anthropogenic climate change is rapidly eroding biodiversity across the globe. While the proliferation of biodiversity databases and modeling tools has enabled us to perform large-scale assessments of climate change vulnerability, the uncertainty associated with projections is rarely investigated. This renders us unable to make confident generalizations across taxa and ecosystems, and impedes robust conservation planning in the context of global change. Furthermore, we are increasingly confronted with the inability to make predictions for the large number of rare species, ignoring a disproportionate component of biodiversity. Here, I describe the results of a systematic assessment of future climate change impacts in Cactaceae, a plant family of almost 2,000 species mostly endemic to the Americas. While the negative impacts of habitat destruction and overcollection have resulted in the family being one of the most threatened groups of organisms on the planet, the lack of a family-wide study of climate change vulnerability limits our understanding of the conservation of this iconic clade. Leveraging biodiversity databases and community science resources, my collaborators and I compiled the most comprehensive, curated Cactaceae biodiversity dataset. With these data, I built species distribution models for a significant portion of the family, finding that approximately 60% of species are expected to decline under climate change. These declines are concentrated in current hotspots and important bioregions of cactus diversity. I thoroughly quantified the uncertainty of these projections, observing that model complexity and dispersal capacity are the most important sources of uncertainty. However, the overall patterns of decline are robust to these modeling decisions. Importantly, the extent of decline is greatly impacted by the inclusion of variables describing the severe drought regime. While extreme climatic events are known to play an important role in shaping species’ distributions, a lack of large-scale, robust forecasts previously made it impossible to account for this observation. Using novel predictions of severe drought, I found that relevant variables were included for 87% of models, emphasizing their importance. Notably, the fraction of species projected to lose over a quarter of their suitable climatic area increases from 25% to 47-51% when severe drought is accounted for. Failing to integrate extreme climatic events therefore leads to a systematic underestimation of extinction risk. Combining my projections with a newly available comprehensive phylogenetic tree for Cactaceae, I investigated whether climate change vulnerability is phylogenetically conserved. I found a significant phylogenetic signal for niche size and range size, and was able to confidently predict range sizes from niche sizes. However, I did not recover a signal for projected changes in range size, and was unable to predict such changes using niche and range size. These results suggest that evolutionary relatedness cannot reliably predict vulnerability to climate change. This limits the utility of phylogenetic extrapolation to infer climate change responses for the large number of rare species for which we are unable to build traditional species distribution models.