Empirical modelling of the population dynamics of a small population of the threespine stickleback, Gasterosteus aculeatus

We estimated the abundance of a small population of threespine stickleback, Gasterosteus aculeatus, by mark-recapture over a 21 year period. Length-frequency analysis showed that the population in October consisted almost entirely of young-of-the-year. The per capita annual rate of increase was inversely related to abundance in October. Time series analysis suggested the presence of a cycle of abundance with a period of about 6 years. There was a significant inverse relationship between abundance in year t and in year t + 3. A simple, empirical, deterministic model based on this inverse relationship and run for 100 years predicted that population abundance showed damped oscillations leading to a stable abundance. When a stochastic component was added to the model, seven of 10 runs included a component with a period of about 6 years. These simulations suggest that the dynamics of this population are driven by an interaction between a deterministic (density-dependent) component and a stochastic component. We compare these results with time series of abundance of threespine stickleback obtained from the Thames Estuary in south-east England and Loch Lomond in Scotland.