Quantifying the vulnerability of biodiversity to climate change.

Abstract

Species distribution models (SDMs) have emerged as a powerful research tool for predicting the impact of climate change on global biodiversity and are increasingly being used to guide international and local decision-making for conservation. However, the accuracy of these models is limited by their reliance on seasonally aggregated climate data derived from weather stations, which fail to reflect conditions experienced by organisms in nature. This is because climate conditions vary in multiple dimensions: geographical space, vertical space, and through time. In this thesis, I explore what this means for predicting the impact of climate change on global biodiversity. I demonstrate that the conclusions reached about where, and which species are most threatened by climate change is fundamentally altered when the discrepancy between standard climate data and conditions experienced by species is considered. In Chapter 2, I show that global diversification has, in part, been driven by spatiotemporal variation in climate, such that the most diverse regions have historically experienced relatively stable conditions and are thus more sensitive to changes that do occur. In Chapter 3, I show that the choice of near-ground or free-air temperatures to fit SDMs significantly influences which places are perceived as most vulnerable to climatic changes. In Chapters 4 and 5, I investigate the temporal and vertical dimensions of climate change and demonstrate that even modest changes can result in widespread and pervasive novel conditions across the tropics. Finally, in Chapter 6, I incorporate spatiotemporal climate gradients into SDMs to investigate whether recent climate changes have resulted in declines of climate suitability and species richness for neotropical bird species. I conclude by emphasising the need to realistically capture the conditions experienced by organisms when predicting how they are responding to climate change. My thesis demonstrates the importance of accurately capturing multi-dimensional climate gradients and considering how they are relevant to organisms when predicting the impact of climate change on global biodiversity. By doing so, predictive modelling can better inform conservation strategies to protect the most vulnerable species and regions.