Abstract
Long-term, concurrent measurement of population dynamics and associated top-down and bottom-up processes are rare for unmanipulated, terrestrial systems. Here, we analyse populations of moose, their predators (wolves, Canis lupus), their primary winter forage (balsam fir, Abies balsamea) and several climatic variables that were monitored for 40 consecutive years in Isle Royale National Park (544 km2), Lake Superior, USA. We judged the relative importance of top-down, bottom-up and abiotic factors on moose population growth rate by constructing multiple linear regression models, and calculating the proportion of interannual variation in moose population growth rate explained by each factor. Our analysis indicates that more variation in population growth rate is explained by bottom-up than top-down processes, and abiotic factors explain more variation than do bottom-up processes. Surprisingly, winter precipitation did not explain any significant variation in population growth rate. Like that detected for two Norwegian ungulate populations, the relationship between population growth rate and the North Atlantic Oscillation was nonlinear. Although this analysis provides significant insight, much remains unknown: of the models examined, the most parsimonious explain little more than half the variation in moose population growth rate.