Can reduced salmonid population abundance be detected in time to limit management impacts?

Abstract

We evaluated eight populations of native salmonids to determine if rapid, sensitive detection of a reduction in abundance is possible in the Yakima River basin, Washington, where a large-scale test of hatchery supplementation is being conducted. Prospective power to detect impacts to abundance was estimated from 3-16 annual baseline surveys conducted by electrofishing, trapping, or snorkeling. High interannual variation in abundance estimates (CV = 26-94%) prevented detection of small impacts for most taxa. For three taxa, models of environmental and biological influences accounted for between 42 and 49% of temporal variation, increasing our ability to detect impacts of other influences. Detectable impacts for a t test with alpha = 0.1 and beta = 0.1 were >18% for all eight taxa and >54% for four of eight taxa. We suggest that population abundance monitoring may not provide feedback sufficiently sensitive or rapid enough to implement corrective actions that prevent impacts from causing harm or exceeding an acceptable level, especially for rare or highly valued taxa with small acceptable impacts.