| Summary: | Population dynamics in space and time are manifested as changes in the distribution and abundance of organisms. To couple such patterns to the underlying processes is a central question in ecology and also key to successful management. In this thesis, I use theoretical models as well as time series data to analyze population dynamics in environments that are variable in space or time. An example of spatial heterogeneity is when we establish a reserve - where individuals are protected from exploitation - and let individuals distribute themselves between the reserve and the surrounding exploited area. I show that if a population conforms to the ideal free distribution (IFD), the harvest rate resulting in the maximum sustainable yield is unaffected by the size as well as the quality of the reserve. Source-sink systems, where there is a net flow from "good" to "bad" habitats, complicate population management, and optimal harvesting decisions are contingent on dispersal rates and quality differences among habitats. Populations also experience temporal variation within a year (seasonality) as well as between years. By incorporating seasonality in a population harvesting model, I give an explanation to the observation that pre-harvest population densities are sometimes unaffected by harvesting. Between-year variability is studied by building stochastic population models that can be approximated by statistical time series models and applied to real data. In exploited populations, the harvesting process itself is another stochastic factor that influences the dynamics. I demonstrate that under many circumstances, variable harvest can explain a considerable proportion of the variation in population density, sometimes even more than explained by environmental stochasticity. The eastern Baltic cod ( Gadus morhua ) seems to be an example of that. Finally, I show that if the environmental stochasticity is temporally autocorrelated, any attempt to disentangle demographic and environmental impact on population dynamics will be ...
|
|---|