Food web structure and climate effects on the dynamics of small mammals and owls in semi-arid Chile

The four-year cycle of microtine rodents in boreal and arctic regions was first described in 1924 (ref. 1). Competing hypotheses on the mechanisms underlying the small mammal cycle have been extensively tested, but so far the sustained rodent oscillations are unexplained. Here we use two mutually supportive approaches to investigate this question. First, building on studies of the interaction between rodents and their mustelid predators, we construct a predator-prey model with seasonality. Second, we use a new technique of nonlinear analysis to examine empirical time-series data, and compare them with the model dynamics. The model parameterized with field data predicts dynamics that closely resemble the observed dynamics of boreal rodent populations. Both the predicted and observed dynamics are chaotic, albeit with a statistically significant periodic component. Our results suggest that the multiannual oscillations of rodent populations in Fennoscandia are due to delayed density dependence imposed by mustelid predators, and are chaotic.

We present a model on plant-deer climate interactions developed for improving our understanding of the temporal dynamics of deer abundance and, in particular, how intrinsic (density-dependent) and extrinsic (plants, climate) factors influence these dynamics. The model was tested statistically by analysing the dynamics of five Norwegian red deer populations between 1964 and 1993. Direct and delayed density-dependence significantly influenced the development of the populations: delayed density-dependence primarily operated through female density, whereas direct density-dependence acted through both female and male densities. Furthermore, population dynamics of Norwegian red deer were significantly affected by climate (as measured by the global weather phenomenon, the North Atlantic Oscillation: NAO). Warm, snowy winters (high NAO) were associated with decreased deer abundance, whereas the delayed (two-year) effect of warm, snowy winters had a positive effect on deer abundance. Our analyses are argued to have profound implications for the general understanding of climate change and terrestrial ecosystem functioning.

Vole dynamics in northern Europe exhibit a well-defined geographical gradient, with oscillatory populations being confined to high latitudes. It has been proposed that oscillations in northern vole populations are driven by their interaction with specialist predators (weasels), while the more southern rodent populations are relatively stable because of regulation by generalist predators. We tested this generalist/specialist predation hypothesis by constructing an empirically based model for vole population dynamics, estimating its parameters, and making predictions about the quantitative pattern of the latitudinal shift in vole dynamics. Our results indicated that the model accurately predicted the latitudinal shift in the amplitude and periodicity of population fluctuations. Moreover, the model predicted that vole dynamics should shift from stable to chaotic as latitude is increased, a result in agreement with nonlinear time-series analysis of the data. The striking success of the model at predicting the shifts in amplitude and stability along the geographical gradient in northern Europe provides strong support for the key role of specialist and generalist predators in vole population dynamics.