Since the seminal work of Charles Elton (1927), population cycles of predators and their prey have been a focal topic of animal ecology and population dynamics (Southern 1970; Hanski, Hansson & Henttonen 1991; Hanski & Korpimäki 1995; Krebs et al. 1995; Lambin, Petty & MacKinnon 2000; Lindström et al. 2001; Gilg, Hanski & Sittler 2003; Sundell et al. 2004; Korpimäki et al. 2005a, b). Cyclic fluctuations in the abundance of herbivores are commonly found in populations on high latitudes and/or high altitudes (Lindström et al. 2001). Because these herbivores typically are basal to the ecosystem, the effect of the cycles in their abundance reverberates across the food web (Ims, Henden & Killengreen 2008). The consequences of herbivore cycles are thus apparent also on higher trophic levels, even when these predators do not directly drive the cycle. This perspective is in contrast to the classic view of predator–prey dynamics as a Lotka–Volterra type of dynamics, where the predator drives the fluctuations in the abundance of the prey and shows similar cycles as its prey but lagging in time. A classic example of predator-prey population dynamics where the predator’s population size does not track the fluctuations in the abundance of its main prey, voles, is provided by Southern’s (1970) study of tawny owls Strix aluco Lin. in southern England.
In Northern boreal environments in Fennoscandia, tawny owls and other birds of prey occur in such low densities that they do not have the potential to impose sufficient predation pressure to make a serious impact on the vole dynamics and hence are, themselves, not driving the cyclic fluctuations in their main prey (Korpimäki et al. 2002; Norrdahl et al. 2004). Resident owl species (Ural owl Strix uralensis Pall. and tawny owl) respond to fluctuations in food abundance by adjusting their reproduction, but – once they have occupied a territory – do not disperse to breed where there are plenty of voles as other (semi-) nomadic species do (Andersson 1980).
By refraining from breeding when food is scarce, the proportion of breeding site-tenacious owls can increase rapidly with increasing numbers of voles, without any delay (Southern 1970; Brommer, Pietiäinen & Kolunen 2002; see also Korpimäki & Norrdahl 1989, 1991; Rohner 1996). On the other hand, the mortality of territorials (and their offspring) is drastically increased when the voles crash in abundance every third year (Brommer et al. 2002). This recurring ‘bottleneck’ creates opportunities for prebreeding individuals (floaters; Rohner 1996) to start breeding when food abundance increases again. As a consequence, fluctuations in food abundance generate changes in the population’s age distribution, as the proportion of young, first-breeding individuals in the population increases when food abundance becomes more favourable (Brommer, Pietiäinen & Kolunen 1998).
For a variety of reasons, young and/or inexperienced individuals may respond differently to environmental fluctuations than older and experienced ones (Metcalf & Pavard 2007). The change in age structure over a cycle therefore potentially creates marked variation in the population’s reproductive output and survival. One powerful way to incorporate such individual differences in performance is to group the individuals in relevant stages. In general, such grouping has important consequences for the understanding of population growth and dynamics (Caswell 2001). In case of a population living in a cyclic environment, changes in population structure across the cycle need to be incorporated and the consequences of a variable population structure for reproduction and survival need to be understood when modelling cyclic population dynamics.
In many places, and particularly in Fennoscandia, herbivore cycles are fading out (Ims et al. 2008), which is expected to present a major change in the environment for many other species that are (partly) dependent on these herbivores. Avian predators of voles are prime candidates for species likely to be negatively affected by changes in the vole dynamics (Hörnfeldt, Hipkiss & Eklund 2005). Although the tawny owl is a generalist predator in Northerly populations, it almost non-exclusively uses voles as a prey when vole abundance is high (Petty 1999) and it is highly dependent on voles for reproduction (Kekkonen et al. 2008). In this study, we determine how reproduction and survival, which together define population growth, of different stages of tawny owls respond to variation in food supply during 15 years of persistent cycles in vole abundance. Our aim was to provide a benchmark for understanding the cyclic tawny owl – vole system to evaluate changes in this system when vole cycles fade out. In particular, we aim to assess the relative importance of variation in reproduction and survival for the dynamics of a predator population subsisting on prey that shows periodic fluctuations in its abundance. Previous studies have shown that reproduction and survival vary in such an environment (e.g. Brommer et al. 1998), but no study has – to our knowledge – quantitatively compared the importance of such variation for the population dynamics. It is not obvious how variation in reproduction can be compared with variation in survival without formal consideration in a population dynamical model. We therefore construct a matrix population model based on our study of fecundity and survival to perform an elasticity analysis, in which we quantitatively resolve how a change in survival and fecundity rate of different stages across the vole cycle would alter population growth rate.