|dc.description.abstract||One challenge in ecology is to predict long-term effects of high herbivore densities on the vegetation, and its feedback consequences. I examined plant-herbivore dynamics during a 26-year period following reintroduction of Svalbard reindeer to Brøggerhalvøya, and how contrasting resource levels influence reindeer foraging behaviour. After one century with reindeer absence, the vegetation had reached some climax state dominated by lichens. Irruptive dynamics are expected following occupation of new ranges. Accordingly, the population size increased exponentially for then to crash and subsequently fluctuate around low densities. The irruption caused a general reduction in vegetation cover, from which only mosses recovered following the relief in herbivore pressure. Lichen depletion and a decline in dominant vascular forage plants suggest overgrazing and reduced carrying capacity K. This provides support for the "exploitation ecosystem hypothesis", predicting top-down control in low-productive environments without predators.
Some reindeer dispersed during the crash and founded a new population at Sarsøyra, providing higher forage availability. Combined with radiotelemetry, I applied two approaches to variation in reindeer resource selection within and between the two populations, based on; (1) vegetation measurements and (2) data from a satellite image and a terrain model. These approaches can provide complementary information about plant-herbivore interactions at different spatial scales. Resource selection was analysed during two summers and one winter. In late winter, i.e. the most critical period for survival and reproduction, reindeer at Brøggerhalvøya selected high-productive habitat and cratered for forage of high quantity but low quality. Individuals that migrated to Sarsøyra for the winter (but returned prior to calving) applied similar habitat selection. Probably due to contrasting experience, the immigrants were spatially and ecologically segregated from the residents at Sarsøyra. The latter group selected particularly for high-quality lichens and habitat with low plant cover but, most likely, favourable snow and ice conditions (i.e. relatively high elevations and low summer moisture), which reduce costs of foraging. As a consequence of different habitat selection, migrants used patches with the highest biomass and plant cover, followed by reindeer at Brøggerhalvøya and residents at Sarsøyra. Winter home range size was inversely related to individual selectivity for the selected habitat variable, suggesting a close relationship between resource level, habitat selection and space use.
Similarly, reindeer at Brøggerhalvøya were more selective for biomass during summer than reindeer at Sarsøyra. High-productive habitats were selected by the former population, whereas habitat selection by reindeer at Sarsøyra was mainly influenced by variables related to abiotic conditions, such as moisture, which may possibly be negatively related to plant quality. The average individual at Brøggerhalvøya thus used patches with higher biomass compared with reindeer at Sarsøyra, whereas the latter were more selective against the satellite variable reflecting moisture. At the home range level, the Sarsøyra population traded high-quantity plants for those that were less common but of higher quality. As in winter, summer home range size was inversely related to individual selectivity for the preferred resource. Thus, reindeer at Brøggerhalvøya, but not Sarsøyra, reduced their home range size in response to high availability of productive habitat and high abundance of forage plants. This provides support for the “habitat productivity hypothesis”, predicting a need for larger areas to fulfil the energy requirements at low productivity. The hypothesis is however only valid when biomass is a selected currency. Despite their increased space use, individuals with low resource availability used patches with lower values of the respective resource than individuals with higher resource availability. In terms of forage types, the positive useavailability relationship was present for high-quality plants but not common plants of lower quality, suggesting that quality was traded for quantity at low resource levels within the populations. This may be particularly important in light of the strong home range fidelity recorded between years.
In conclusion, reindeer suppress their food base by top-down control in this predatorfree system. Contrasting resource availability generates a change in habitat preferences (i.e. functional response in habitat selection) that is closely related to a change in the relative importance of foraging niche components. Several results indicate that forage quantity and quality are the most selected components at low and high resource levels, respectively. This is consistent with optimal foraging theory.||nb_NO