Population dynamics of pampas mice (Akodon azarae): signatures of competition and predation exposed through time-series modeling

The ability of four mechanistic population models to mirror a 5-year time series of seasonal densities of the pampas mouse, Akodon azarae, from central Argentina, is evaluated in this paper. The different models included, singly or in combination, the effects of a putative competitor and a specialist predator on the dynamics of pampas mice. The most simple model, a logistic single-species stochastic differential equation, failed to fit the observed time series satisfactorily (R model ² = 0.36), while models including interspecific effects gave much better fits with increasing model complexity (competition R model ² = 0.71, predation R model ² = 0.62, competition and predation R model ² = 0.83). Using Akaike's information criterion, the competition model was selected as the best alternative with regard to model fit and model complexity. Synthetic time-series probes obtained by simulations from the parameterized stochastic mechanistic models were significantly different and indicated the competition plus predation model as a slightly better alternative than the competition model. The differences between the two methods for selecting models indicate that the incorporation of predation only slightly improves the match between the predicted and observed time series and that this slight improvement is not sufficient to override the increase in model complexity. Competition thus seems to play a more important role than predation in shaping the population dynamics of pampas mice.