Ecosystem-based management of predator-prey relationships: piscivorous birds and salmonids.

Predator-prey relationships are often altered as a result of human activities. Where prey are legally protected, conservation action may include lethal predator control. In the Columbia River basin (Pacific Northwest, USA and Canada), piscivorous predators have been implicated in contributing to a lack of recovery of several endangered anadromous salmonids (Oncorhynchus spp.), and lethal and nonlethal control programs have been instituted against both piscine and avian species. To determine the consequences of avian predation, we used a bioenergetics approach to estimate the consumption of salmonid smolts by waterbirds (Common Merganser, California and Ring-billed Gull, Caspian Tern, Double-crested Cormorant) found in the mid-Columbia River from April through August, 2002-2004. We used our model to explore several predator-prey scenarios, including the impact of historical bird abundance, and the effect of preserving vs. removing birds, on smolt abundance. Each year, <1% of the estimated available salmonid smolts (interannual range: 44,830-109,209; 95% CI = 38,000-137,000) were consumed, 85-98% away from dams. Current diet data combined with historical gull abundance at dams suggests that past smolt consumption may have been 1.5-3 times current numbers, depending on the assumed distribution of gulls along the reaches. After the majority (80%) of salmonid smolts have left the study area, birds switch their diet to predominantly juvenile northern pikeminnow (Ptychocheilus oregonensis), which as adults are significant native salmonid predators in the Columbia River. Our models suggest that one consequence of removing birds from the system may be increased pikeminnow abundance, which--even assuming 80% compensatory mortality in juvenile pikeminnow survival--would theoretically result in an annual average savings of just over 180,000 smolts, calculated over a decade. Practically, this suggests that smolt survival could be maximized by deterring birds from the river when smolts are present, allowing bird presence after the diet switch to act as a tool for salmonid-predator control, and conducting adult-pikeminnow control throughout. Our analysis demonstrates that identifying the strength of ecosystem interactions represents a top priority when attempting to manage the abundance of a particular ecosystem constituent, and that the consequences of a single-species view may be counterintuitive, and potentially counterproductive.