부산대학교 도서관

Demography aims to understand the changes in population numbers or density arising from individuallevel variation in fertility, mortality, and migration. Understanding and predicting population dynamics remains an important keystone for ecologists in order to identify population management strategies and explore evolutionary optimisation of timings and energetic allocations of life history strategies. In chapter two, I examine how comparative analysis of published plant matrix population models (MPMs) from the COMPADRE Plant Matrix Database can be used to determine axes of life history variation using principal component analysis. I simulate population models under the assumption that density dependence constrains growth rates as a result of a carrying capacity: a non-adaptive constraint. I found density dependent constraints explained much of the covariance patterns in life history metrics. In chapter three, I use this simulated population model framework to explore the link between life history and transient dynamics, quantified as responses to perturbations. Indices of transient response derived from population models also exhibited non-adaptively constrained covariance patterns. Transient response was characterised on two axes; magnitude of transient response and tendency to attenuated as opposed to amplify. In chapter four, I show that how we model the interannual fluctuations in vital rates affects our model's resulting population dynamics, life history metrics and responses to perturbations, using data from Soay Sheep population on the island of St Kilda. I show this by modelling vital rates with generalised linear mixed-effects models (GLMMs) and hierarchical generalised additive models (HGAMs) and comparing the resulting integral projection models. In chapter five, I discuss the need to be precise in interpretating results of demographic approaches. How we model populations, and the resulting non-adaptive constraints, play an important role in shaping these results. I outline future uses of this simulated population model framework for time-varying population models and single species study systems.