Cereal aphid populations in non-crop habitats show strong density dependence.
Few studies have addressed how density-dependent and density-independent regulation of population growth and abundance varies among habitats for a species that requires multiple habitat types to complete its life cycle. Understanding such relationships, however, are of direct relevance to the control of crop pest insects that regularly move between crop and non-crop habitats. We used autoregressive models to analyse a series of seasonal catches of the cereal aphid Rhopalosiphum padi. The data were collected from regional suction traps and egg counts on the overwintering host, over a period of 14-31 years, at four locations in Sweden. R. padi is an obligate host-alternating species in Sweden and seasonal catches reflect habitat use in a year: the primary woody host in winter and spring, the cereal crops in summer, and the perennial grasslands in the autumn. Strong direct density dependence acting within the year was found, but the strength varied between seasons depending on habitat use by the aphids during the year. Only a weak indication or no indication at all of density dependence was found during the period of residency on the primary host in the winter and spring periods. Density dependence occurred when R. padi utilized summer cereals (42% of the variation was explained), and even stronger density dependence occurred in the perennial grasslands in the autumn (70% variation explained). Stochastic fluctuations in the winter and spring were balanced by a strong density dependence in the cereal and grassland habitats in the summer and autumn periods, which reduced variability in population fluctuations. Weather, measured as seasonal average temperature and accumulated precipitation, did not affect aphid abundance fluctuations much, explaining only 1-9% of the variability. Synthesis and applications. Our results suggest that density-dependent regulation of R. padi occurs in late summer grasslands and early summer cereals. The mechanisms causing these patterns are not understood, making it difficult to provide specific pest management recommendations at this stage. The results do indicate, however, that pest management needs to involve a landscape-level approach, taking into account mechanistic information about the plant, herbivore and predator interactions in multiple habitats visited by the herbivorous pest.