Local movement as a measure of habitat quality in stream salmonids

Habitat assessments are often based on the premise that spatial variation in population density arises from, and accurately reflects, underlying differences in quality among habitats. Nonetheless, this premise has been criticized on both theoretical and empirical grounds. Habitat quality perhaps is best evaluated by examining behavioural processes which directly influence habitat use. We present an approach based on the assumption that measures of local movement, such as habitat-specific immigration and loss rates, provide useful indicators of habitat quality. A dynamic turnover model was used in conjunction with mark-recapture techniques to estimate movement parameters for brook charr, Salvelinus fontinalis, and Atlantic salmon, Salmo salar, in different stream habitats during the summer. Immigration and loss rates were derived from mark-recapture experiments covering short periods of time (6 days). Movement-based rankings of habitat quality for both charr (pools greater than or equal to glides > riffles) and salmon (riffles > glides > pools) were in agreement with results from earlier studies. Over evaluation periods of up to 65 days, observed abundances were highly variable in time and fluctuated about the equilibrium abundances calculated from movement parameters in the short-term experiments, suggesting that movement-based parameters may be more stable than measures of abundance for evaluating salmonid habitat. Because the approach based on movement behaviour focuses on immigration and loss, two processes that directly generate density differentials between habitats, it provides a more reliable mechanistic basis for understanding habitat selection than do traditional approaches based on density-quality relationships.