Inclusion of water temperature in a fuzzy logic Atlantic salmon (Salmo salar) parr habitat model

As water temperature is projected to increase in the next decades and its rise is clearly identified as a threat for cold water fish species, it is necessary to adapt and optimize the tools allowing to assess the quantity and quality of habitats with the inclusion of temperature. In this paper, a fuzzy logic habitat model was improved by adding water temperature as a key determinant of juvenile Atlantic salmon parr habitat quality. First, salmon experts were consulted to gather their knowledge of salmon parr habitat, then the model was validated with juvenile salmon electrofishing data collected on the Sainte-Marguerite, Matapedia and Petite-Cascapedia rivers (Québec, Canada). The model indicates that when thermal contrasts exist at a site, cooler temperature offered better quality of habitat. Our field data show that when offered the choice, salmon parr significantly preferred to avoid both cold areas (<15 °C) and warm areas (>20.5 °C). Because such thermal contrasts were not consistently present among the sites sampled, the model was only validated for less than 60% of the sites. The results nevertheless indicate a significant correlation between median Habitat Quality Index and parr density for the Sainte-Marguerite River (R² = 0.38). A less important, albeit significant (F-test; p = 0.036) relationship was observed for the Petite-Cascapedia river (R² = 0.14). In all instances, the four-variable (depth, velocity, substrate size and temperature) model provided a better explanation of parr density than a similar model excluding water temperature.