| Summary: | Within the Arctic food web, Arctic cod (Boreogadus saida) is the dominant trophic link between zooplankton and higher predators such as seals and sea-birds. An ice-dependent lifestyle and a specialized physiology make this species particularly vulnerable to the predicted reduction in ice-cover and the potential invasion of its ecological niche by boreal generalists. With the unfolding of climate warming which is particularly intense in the Arctic, it is essential to document the factors that control the fate of this key species. The primary focus of this thesis is the identification of the necessary and sufficient processes required in a mathematical model to realistically simulate the growth and survival of the larval stage of Arctic cod. An individual-based model was developed and validated with observations collected in the North Water (Baffin Bay) and the Northeast Water (Greenland Sea). Temperature, as well as the concentration and quality of prey were necessary to reproduce observed growth and its regional differences. Predation pressure was simulated using a length- and growth-dependent mortality which allowed the selective elimination of small and slow-growing individuals. The model took into account the hatch-date of individuals to reproduce differences in life trajectory among individuals hatched before and after the polynya opened. For instance, early hatchers grew slower than late hatchers due to lower temperatures, but this trend was reversed if the yolk exhaustion of early hatchers coincided with high prey concentrations. A novel method of model output sub-sampling accounting for the age, hatch-date and geographical position of individuals captured at sea, reproduced the inevitable discontinuities that plague field observations. This method improved model validation, allowed weaknesses to be identified and highlighted new research directions. The model output sub-sampling has wide-ranging applications in individual-based modeling and is applicable to any model attempting to simulate organisms that ...
|
|---|