Well-functioning food webs are fundamental for sustaining rivers as ecosystems and maintaining associated aquatic and terrestrial communities. The current emphasis on restoring habitat structure-without explicitly considering food webs-has been less successful than hoped in terms of enhancing the status of targeted species and often overlooks important constraints on ecologically effective restoration. We identify three priority food web-related issues that potentially impede successful river restoration: uncertainty about habitat carrying capacity, proliferation of chemicals and contaminants, and emergence of hybrid food webs containing a mixture of native and invasive species. Additionally, there is the need to place these food web considerations in a broad temporal and spatial framework by understanding the consequences of altered nutrient, organic matter (energy), water, and thermal sources and flows, reconnecting critical habitats and their food webs, and restoring for changing environments. As an illustration, we discuss how the Columbia River Basin, site of one of the largest aquatic/riparian restoration programs in the United States, would benefit from implementing a food web perspective. A food web perspective for the Columbia River would complement ongoing approaches and enhance the ability to meet the vision and legal obligations of the US Endangered Species Act, the Northwest Power Act (Fish and Wildlife Program), and federal treaties with Northwest Indian Tribes while meeting fundamental needs for improved river management.

Estimated growth rates, condition, and stomach fullness of juvenile coho salmon (Oncorhynchus kisutch) caught in the ocean in early summer, when mortality was most variable, were as high in 1983 and 1984, years of very low survival and low early upwelling, as in 1981, 1982, and 1985, years of higher survival and higher early upwelling. Chronic food shortage leading to starvation, poor condition, or slow growth apparently was not the cause of the increased mortality of juvenile coho salmon in 1983 and 1984. Survival of juvenile coho salmon was positively correlated with purse seine catches of fish in June and with early summer upwelling, 1981–85. Hence, year-class success probably was determined early in the summer, soon after most juvenile coho salmon entered the ocean. Spacing of the first five ocean circuli, which was positively correlated with growth rate, was not significantly different for fish caught early in the summer and those caught late in the summer, suggesting that growth rate selective mortality in the ocean was not strong. The increase in mortality in 1983 and 1984 may have been caused by increased predation on juvenile coho salmon due to decreased numbers of alternative prey for predators.

The very strong 1982–1983 El Niño is implicated in extraordinary northern range extensions and shifts in population distributions of many species of marine organisms from Oregon to Alaska during 1983. Some species such as Pacific and jack mackerel maintained large populations in the Pacific Northwest during the summer of 1983, as well as the summer of 1984, a year after physical subsidence of the event. Although salmon catches were large in Alaskan waters, populations off Oregon were severely reduced during this El Niño period. Mortality of maturing coho was unprecedented; the sizes of both coho and Chinook salmon in fisheries were at record lows, and survival of juvenile coho that entered the ocean in 1983 and 1984 was low. Decreased reproductive success and increased mortality of seabirds were observed during 1983 from Oregon, Washington, and Alaska and were probably related to decreased availability and productivity of usual prey in the abnormally warm surface waters. The large numbers of independent observations on unusual biological events during 1983 and the coincidence of occurrences of some species with previous strong El Niños infer that some changes observed in 1983 were a consequence of the strong 1982–1983 El Niño.

The Columbia Basin once supported a diversity of native fishes and large runs of anadromous salmonids that sustained substantial fisheries and cultural values. Extensive land conversion, watershed disruptions, and subsequent fishery declines have led to one of the most ambitious restoration programs in the world. Progress has been made, but restoration is expensive (exceeding US$300M/year), and it remains unclear whether habitat actions, in particular, can be successful. A comprehensive approach is needed to guide cost‐effective habitat restoration. Four elements that must be addressed simultaneously are (1) a scientific foundation from landscape ecology and the concept of resilience, (2) broad public support, (3) governance for collaboration and integration, and (4) a capacity for learning and adaptation. Realizing these in the Columbia Basin will require actions to rebalance restoration goals to include diversity, strengthen linkages between science and management, increase public engagement, work across traditional ecological and social boundaries, and learn from experience.

The overall food consumption by juvenile coho (Oncorhychus kisutch) and chinook salmon (O. tshawytscha) along the continental shelf off Washington and Oregon was estimated for the summer months of 1981–84 using both a direct field and indirect bioenergetic approach. Both models yielded similar estimates but the consumption estimated by the bioenergetic model was 14.8 and 3.2% higher for coho and chinook salmon, respectively. In a given year, overall consumption increased throughout the summer, despite decreasing abundances of both salmon species, but there were substantial differences among the four years within the same month. A preliminary comparison of the consumption estimates for important prey taxa early in the summer with estimated zooplankton biomass suggested that sufficient food levels may have been available for salmon during 1981. During the El Niño of 1983, however, several prey groups were consumed at high levels relative to their estimated availability and the biomass of these taxa may not have been sufficient to sustain the total population of salmon.