Search

It has long been recognized that, as populations increase in density, ecological processes affecting growth and survival reduce per capita recruitment in the next generation. In contrast to the evidence for such "compensatory" density dependence, the alternative "depensatory" process (reduced per capita recruitment at low density) has proven more difficult to demonstrate in the field. To test for such depensation, we measured the spawner-recruit relationship over five decades for a sockeye salmon (Oncorhynchus nerka) population in Alaska breeding in high-quality, unaltered habitat. Twenty-five years of detailed estimates of predation by brown bears, Ursus arctos, revealed strong density dependence in predation rate; the bears killed ca. 80% of the salmon in years of low salmon spawning abundance. Nevertheless, the reconstructed spawner-recruit relationship, adjusted to include salmon intercepted in the commercial fishery, provided no evidence of demographic depensation. That is, in years when few salmon returned and the great majority were killed by bears, the few that spawned were successful enough that the population remained highly productive, even when those killed by bears were included as potential spawners. We conclude that the high quality of breeding habitat at this site and the productive nature of semelparous Pacific salmon allowed this population to avoid the hypothesized depressed recruitment from depensatory processes expected at low density. The observed lack of demographic depensation is encouraging from a conservation standpoint because it implies that depleted populations may have the potential to rebound successfully given suitable spawning and rearing habitat, even in the presence of strong predation pressure.

The metacommunity concept has recently been described to account for the roles of dispersal in regulating community structure. Despite its strong theoretical basis, there exist few large-scale and long-term examples of its applicability in aquatic ecosystems. In this study we used a long-term dataset (1961-2007) on the relative abundances of the dominant limnetic fishes from two interconnected lakes to investigate the synergistic effects of naturally declining lake volume (approximately 50% in 50 years), climate variation, fishery management, and dispersal on community composition. We found a marked shift in fish community composition and variability during a period of rapid natural habitat change; however, the change was most apparent in the downstream, more stable lake of the system rather than at the site of disturbance. Multivariate analysis suggested significant shifts in community composition and variability in the downstream lake. Results indicated that the community composition in both lakes was best explained by habitat loss in the upper watershed and the number of spawning adult sockeye salmon the previous year (reflecting both natural processes and commercial fishing). Furthermore, communities exhibited site-specific responses to climatic conditions (e.g., index of the Pacific Decadal Oscillation), whereby the upper lake responded to climate within a given year and with a 1-year time lag, whereas the downstream community responded only with a 1-year lag. We attribute this difference largely to downstream dispersal and recruitment of fish from the upper lake. Thus, we suggest that the interconnected nature of the communities in this system provides a useful and large-scale example of the metacommunity concept, whereby the effects of environmental disturbance on community structure ultimately depend on the effects of these disturbances on dispersal among ecosystems.