Embeddedness: A context dependent driver of fish habitat preference
Resources within streams can be patchy in their distribution at multiple spatial scales from within local habitat sections (i.e. pools, riffles, runs), to reaches and segments, to entire watersheds. It has been hypothesized that fish may move between patches in order to maximize their fitness, and it has been suggested that movement into and out of patches (i.e. immigration rate, emigration rate, and turnover) may be indicative of perceived habitat quality. We measured movement parameters of juvenile coaster brook trout (Salvelinus fontinalis), and sculpin (Cottus sp.) in two reaches of the Salmon Trout River, Marquette County, MI. We hypothesized that local habitat sections (~10m channel geomorphic units) containing patches of exposed larger substrates (e.g., boulders, cobbles, pebbles) would be more preferred over local habitat sections with greater fine sediment (sand) aggradation. We tested this hypothesis in two reaches: one heavily aggraded with sand (mean sand depth 0.232m ± 0.211m) and one less aggraded reach (mean sand depth 0.036m ± 0.087m). We used the dynamic turnover model developed by Bélanger and Rodriguez (2002) to estimate movement parameters in local habitat sections to infer habitat preferences of stream fish.
We found that in reaches heavily aggraded with fine sediments embeddedness seemed to be a strong predictor of immigration rate (i.e. habitat preference) of brook trout (r2=0.96, p=0.004) and sculpin (r2=0.93, p=0.008). We used Mantel’s test to assess the impact of other habitat characteristics and spatial autocorrelation. We found that within the heavily aggraded reach, local habitat sections with similar embeddedness had similar brook trout immigration rates, taking into account the indirect effects of physical habitat and local habitat position on embeddedness (Mantel’s r = 0.99, p=0.0001). Interestingly, we did not detect a significant response of immigration rate to embeddedness in the less aggraded site. This suggests that the importance of exposed large substrates such as cobbles appears to be context or scale dependent. When larger substrates are exposed in sandy reaches, their relative rarity may increase their attractiveness as habitats. However, where the prevalence of large substrates is greater, the relative importance of large substrate to habitat selection is less. This research suggests that restoration of heavily aggraded streams that exposes patches of large substrates (i.e., preferred habitats) may be of great benefit to local brook trout and sculpin populations within anthropogenic sand affected rivers.