Global change and population dynamics: Understanding heterogeneity in European beech demography across its rear edge distribution

Abstract

Climate and land use change are altering species distributions across the globe. Understanding the driving mechanisms behind species range dynamics is essential to improving predictions of species distributions. The performance of rear edge populations (the equatorial or low altitude margin) is an important determinant of a species’ long-term persistence, yet observed heterogeneity in demographic performance among rear edge populations remains poorly understood. Using European beech (Fagus sylvatica L.) populations in northeastern Spain as a case study, this thesis aims to disentangle potential interactions between global change drivers to enhance understanding and future prediction of heterogeneous rear edge performance. This study is the first to account for multiple dimensions of marginality and interactions with land use legacies to investigate performance of rear edge populations.

In Chapter 2, I found that demographic rates of beech were differently impacted by interactions between human land use legacies and marginality gradients: on a multiplicative scale, intense historic management either magnified negative relationships between demography and marginality, or reduced demographic performance in favourable climates to levels observed in unfavourable climates. In Chapter 3, I developed an Integral Projection Model (IPM) to determine how these interactions impacting beech demography combined to influence regional-scale population dynamics, alongside creating future population projections. I found the direct effects of some global change drivers on population dynamics depended on land use legacies and life stage, alongside projected decline of rear edge beech populations regardless of future climate. In Chapter 4, I found that machine learning predicted the outputs of the IPM with over 90% accuracy, demonstrating its potential use as a metamodelling tool for scaling up process-based models. Overall, I show that heterogenous performance across rear edge populations may be the combined result of many, potentially interacting variables, and that disentangling human and environmental factors can facilitate better understanding of the underlying mechanisms of species range dynamics.