Sampling Little Fish in Big Rivers: Larval Fish Detection Probabilities in Two Lake Erie Tributaries

Larval fish are sampled in Great Lakes tributaries to elucidate factors driving recruitment, evaluate spawning success, and estimate larval production.  Larval fish are small and unevenly distributed; therefore detection of most species is likely imperfect. We estimated detection probabilities of larval fish from several taxa in the Maumee River and Detroit River. We then determined how accounting for imperfect detection influences estimates of density for larval fish of three important Detroit River species and estimates of taxonomic richness of larval fish in the Maumee River and Detroit River.  We found that detection probabilities were highly variable among taxa but were always less than 1.0.  In general, taxa with small, abundant larvae had high detection probabilities (>0.5) whereas those with large, less abundant larvae had low detection probabilities (<0.5). Accounting for imperfect detection increased density estimates for lake whitefish Coregonus clupeaformis but had little effect on density estimates for walleye Sander vitreus and yellow perch Perca flavescens. When undetected taxa were accounted for, richness estimates were one to four taxa greater than observed richness. Our findings show that incorporating detection probabilities may improve the analysis and interpretation of larval fish data.