Population estimation in a salmonid removal project on a coaster brook trout stream in Michigan: Variation in population estimation

Abundance estimates are a key element needed for evaluating the status of a fishery.   Often various population estimation methods are performed over time; however, results may vary by methodology.   This study focuses on variability in three common methods used in stream systems: three pass depletion, mark/recapture, and catch per unit effort (CPUE). This study is part of a restoration project focusing on removing exotic salmonids from a coaster brook trout (Salvelinus fontinalis) stream in Michigan.  All salmonid population sizes were estimated via the three pass depletion estimator in the fall and tracked throughout the ice off months via CPUE.  Brook trout greater than 100mm were PIT tagged facilitating long term mark/recapture abundance calculations.  Abundances varied among estimation methods allowing us to assess the ramifications of using less labor intensive methods of estimation.   Initial results showed depletion estimates differing from mark/recapture estimates by 0% to 440%; standard error was high in the mark/recapture estimates.  CPUE tended to deviate from both the mark/recapture and depletion estimators; if abundances were tracked over a longer period with CPUE estimates should improve.  Due to the high standard error generated from the long term mark/recapture estimate the depletion estimator appears to perform better in these conditions.