Abstract

Climate can have complex effects on demographic rates and the endogenous feedback structure regulating mammal populations, and this can create problems for predictive modelling. In northern and Alpine environments, weather appears to influence ungulate population growth rates mainly during years of high population density, suggesting the possibility of nonlinear interactions between the 2 variables. Threshold models have been employed to account for these nonlinearities. For example, in the case of Alpine ibex Capra ibex in the Gran Paradiso National Park of Italy, stronger density dependence is assumed to occur after snowfall exceeds 1.54 m. In this paper we use more objective nonparametric methods to evaluate the form of the functional relationships governing the dynamics of this ibex population. No evidence was found for a threshold effect in the data. Instead we uncovered a non-additive and nonlinear interaction between climate and population density. The resulting models predict ibex numbers as well or better than previous threshold models despite requiring fewer parameters, and also conform well to traditional ecological concepts. We conclude with several lessons for those who wish to predict the effects of climate change on animal population dynamics. © Inter-Research 2006.