Indicators of fish population responses to avian predation with focus on double-crested cormorants

Abstract

Double-crested cormorants (Nannopterum auritum) have been implicated as causes of fish population declines in many locations across their breeding range. Two challenges facing managers are identifying fisheries population metrics indicative of cormorant impacts and determining when this evidence becomes actionable. Building upon existing studies, we conducted a meta-analysis of eight data-rich systems across the Laurentian Great Lakes region of the United States for common fish population responses to changes in cormorant abundance. Specifically, we examined trends in mean total female length at age-3 (TL3), female mean length and age at 50 % maturity, and mean age evenness as indicated by Shannon’s Equitability Index. Annual observations for these metrics were independently regressed linearly against cormorant density by system for walleye (Sander vitreus), yellow perch (Perca flavescens), smallmouth bass (Micropterus dolomieu), and northern pike (Esox lucius) populations. TL3 was the most sensitive with 9 of the 14 datasets statistically significant (r2 range 0.29 to 0.86). Maturity metrics were moderately sensitive to trends in cormorant predation with mean total length at 50 % maturity significant in 4 out of 11 datasets (r2 range 0.27–0.41) and mean age at 50 % maturity significant in 3 out of 11 datasets (r2 range 0.12 – 0.51). Least sensitive was age evenness with the Shannon Index significant in 3 out of 12 datasets (r2 typically < 0.25). Of metrics tested, TL3 was the most reliable indicator of changes in cormorant effects despite varying system changes and management responses among locations.