PLoS ONE (Jan 2014)
A modeled comparison of direct and food web-mediated impacts of common pesticides on Pacific salmon.
Abstract
In the western United States, pesticides used in agricultural and urban areas are often detected in streams and rivers that support threatened and endangered Pacific salmon. Although concentrations are rarely high enough to cause direct salmon mortality, they can reach levels sufficient to impair juvenile feeding behavior and limit macroinvertebrate prey abundance. This raises the possibility of direct adverse effects on juvenile salmon health in tandem with indirect effects on salmon growth as a consequence of reduced prey abundance. We modeled the growth of ocean-type Chinook salmon (Oncorhynchus tshawytscha) at the individual and population scales, investigating insecticides that differ in how long they impair salmon feeding behavior and in how toxic they are to salmon compared to macroinvertebrates. The relative importance of these direct vs. indirect effects depends both on how quickly salmon can recover and on the relative toxicity of an insecticide to salmon and their prey. Model simulations indicate that when exposed to a long-acting organophosphate insecticide that is highly toxic to salmon and invertebrates (e.g., chlorpyrifos), the long-lasting effect on salmon feeding behavior drives the reduction in salmon population growth with reductions in prey abundance having little additional impact. When exposed to short-acting carbamate insecticides at concentrations that salmon recover from quickly but are lethal to invertebrates (e.g., carbaryl), the impacts on salmon populations are due primarily to reductions in their prey. For pesticides like carbaryl, prey sensitivity and how quickly the prey community can recover are particularly important in determining the magnitude of impact on their predators. In considering both indirect and direct effects, we develop a better understanding of potential impacts of a chemical stressor on an endangered species and identify data gaps (e.g., prey recovery rates) that contribute uncertainty to these assessments.