Power of Revisit Monitoring Designs to Detect Forestwide Declines in Trout Populations

Abstract

Managers are often required to monitor networks of sites to make inferences about trends in fish populations over large geographic areas. We evaluated the statistical power of four revisit monitoring designs (always revisit design; two types of augmented serially alternating design; and serially alternating design) to detect declines in trout population biomass over a 30-year time horizon. The Medicine Bow National Forest, Wyoming, was used as a case study, and the designs were based on sampling 20 or 30 sites every other year but varied in the total number of sites (20–40 or 30–60) and the timing and frequency with which sites were monitored. The always revisit design, in which the same 20 or 30 sites were sampled during every monitoring period, often had slightly higher power than the other designs. However, any differences in power quickly diminished due to the rapid increase in power over time for all four designs. The similar levels of statistical power to detect population declines among the designs we evaluated suggest that managers have flexibility to choose a revisit design with increased spatial coverage and implementation complexity without sacrificing trend detection capability.