Population and Community Dynamics in Variable Environments

All natural populations fluctuate in space and time. These fluctuations are a result of internal dynamic processes, and of a variable environment. To understand how and why population fluctuate, one has to understand the nature of the internal demographic processes as well as how a variable environment affects these processes and how they act in concert to generate population fluctuations. This thesis concentrates on the last step - how a variable environment and internal dynamic processes act together to produce population dynamics. In particular, the dynamical consequences of temporal or spatial structure of the environmental fluctuations are investigated, but also how external disturbances propagate through ecological systems. It is shown that a positively autocorrelated environment is disadvantageous to populations with undercompensating dynamics, and vice versa for overcompensating populations. It is further shown that overcompensating dynamics leads to extinctions from high population densities, above carrying capacity. A two-species system, where only one of the populations is subject to environmental variability, is investigated and a general rule for the dynamics is derived: the sign of a single element of the Jacobian matrix determines whether the dynamics of one of the interacting populations is more or less dominated by low frequency fluctuations than the other. Also, the causes behind spatial correlation of populations (synchrony) is studied. It is shown that only unstable or close to unstable populations can be efficiently synchronised by dispersal between habitat patches. A correlated environment always generates synchronous populations (the Moran effect). The strengths and benefits of using a linear approach to analyse stochastic population dynamics is emphasised and illustrated with several examples.