Home range relocation: How habitat quality, landscape connectivity and density affect movements in coral reef fish

Short-term immigration via home range relocation has important implications for metapopulation dynamics, sustainable harvest and pest control strategies, and conservation in populations experiencing high localized mortality. Despite its importance, no suitable theory is available to predict immigration in response to harvest near an adjacent protected area. There is also little information about the factors that influence the magnitude of immigration. I developed a compensatory immigration model to predict the effect of harvesting on immigration under different assumptions about the factors limiting immigration. The model predicts that immigration from protected areas can contribute importantly to total yield and population recovery in the harvest area and can strongly affect demography in the protected area. Immigration and total yield can show complex non-linear relationships with harvesting as the size of the protected area, initial rate of replacement, mobility and behavioral interactions vary. I carried out field experiments to validate the model and assess the relative influence of limiting factors to immigration (i.e. population size in the protected area, the relative habitat quality and functional connectivity between the harvest and protected area). I used longfin ( Stegastes dienaceus) and dusky damselfish (S. adustus) as a model system. I first examined functional connectivity by translocating damselfish to investigate which habitat or social features represent a barrier to movement during homing. Small sand gaps constitute a partial barrier to movement, but the effect of sand gap width varies with reef configuration, and fish minimize travel over conspecific territories. I carried out replicated, experimental, incremental reduction of damselfish populations to examine the mechanisms behind home range relocation at the scale of the territory and to evaluate immigration at the landscape scale. At the territory scale, the probability that a territory would be recolonized decreases with local density. Territories tend to be reoccupied by individuals of the same species, sex and size as the original occupant, but territories occupied by larger individuals are more likely to be recolonized. At the landscape scale, models assuming a constant but partial replacement of removed individuals predict immigration much better than models that include either no replacement or complete replacement. In several sites, the best fitting model also included parameters describing density dependence that changed in direction and magnitude with cumulative harvest. Total yield and the proportion of removed individuals replaced by immigrants were correlated with the combined effects of relative habitat quality and connectivity of replicate sites. Overall, my thesis proposes and tests a new compensatory immigration model to predict immigration from protected to harvested areas and suggests that variation in mobility, habitat quality, functional connectivity and behavioral interactions must be considered when predicting the effects of immigration in a harvesting context on metapopulation dynamics, sustainable harvest and conservation.