Redd superimposition mediates the accuracy, precision, and significance of redd counts for cutthroat trout

Abstract

Redd counts are commonly applied to estimate spawning population size for salmonids and allow for broad spatial and temporal coverage in monitoring efforts. However, the utility of redd counts may be compromised by observation error, particularly with respect to superimposition, where later arriving spawners construct redds overlapping existing redds. Here, we provide a mechanistic evaluation of the effects of superimposition on the error structure and biological significance of redd count data for Yellowstone cutthroat trout ( Oncorhynchus clarkii bouvieri) spawning within tributaries to the Snake River, Wyoming. We used a Bayesian framework to parse observation error into distinct components and found low detection of redd clusters (i.e., areas of superimposition) was offset by overestimates of the number of redds per cluster, such that observed counts accurately reflected census redd abundance. However, a saturating relationship between redd counts and spawner abundance indicated that counts is best interpreted as effective reproductive effort rather than spawner abundance. Our results provide a mechanistic understanding of redd count data that can be used to assess their application and interpretation for monitoring.