Revealing ecological processes of range dynamics through space and time

Abstract

In ecology, process-explicit models represent the dynamics of ecological systems as explicit functions of the mechanisms and drivers that produced them. Process-explicit models are therefore able to link observed ecological patterns, such as species spatial abundance patterns, directly to their causes, such as climate and environmental change. In this PhD thesis, I show how processexplicit models can be used to establish determinants of range collapses and extinction by unpacking complex interactions between ecological lifestyles, biological traits, climate change, and human-driven threats. By providing a more complete understanding of the ecological mechanisms that regulate species’ responses to climate and environmental change, my PhD research provides information needed to better predict vulnerability to future climate and environmental change. In Chapter I, I reviewed and interpreted the techniques used to unlock ecological and evolutionary mechanisms responsible for spatial and temporal patterns of biodiversity ranging from the gene to the ecosystem. By revealing how models can codify the generalisable mechanisms responsible for the distributions of life on Earth, this review will help to enable important advances in macroecology, evolutionary biogeography and conservation biology, strengthening both basic and applied science. Chapter II is a sensitivity analysis of the Climate Informed Spatial Genetic Model (CISGeM), a process-explicit model of human migration out of Africa. While it is well-known that correlative models of species ranges, such as environmental niche models, are highly sensitive to the climate dataset used for parameterisation, the sensitivity of process-explicit models of human migration to climate data and other model parameters has never been tested. I found that the outputs of CISGeM are robust to the choice of palaeoclimate simulation data, but sensitive to the values for key demographic processes. In Chapter III, I used process-explicit models to reconstruct the late Quaternary range dynamics of the steppe bison (Bison priscus) using a new R package, paleopop, that I co-developed. The approach linked spatially explicit population models with inferences of demographic change from fossils and ancient DNA to continuously simulate 45,000 years of steppe bison extinction dynamics. The models included dispersal and demographic processes responding to human harvesting and rapid deglacial warming. I found that deglacial warming interacted with hunting pressure from humans to cause the range of the steppe bison to contract to refugial highland populations, which became extinct in the early Holocene. Chapter IV used a related approach to reconstruct the range dynamics of the European bison (Bison bonasus) from the last ice age to the year 1500. The European bison became extinct in the wild in 1927 and has been bred back from captive animals. It is a goal of European Union policy to reintroduce the bison more broadly, but there is a debate about the optimal locations and habitats for reintroduction. I inform this debate by showing where bison became extinct due to hunting, land use change, and climate change. General findings from my PhD will help macroecologists to better model and understand species range and extinction dynamics, providing important theoretical and applied insights for conserving vulnerable species in the Anthropocene.