Watershed-scale land use activities influence the physiological condition of stream fish

Land use changes within watersheds can have large effects on stream ecosystems, but the mechanistic basis of those effects remains poorly understood. While changes to population size presumably reflect underlying variation in organismal health and condition, such individual-level metrics are rarely evaluated in the context of ecosystem disturbance. To address this deficiency, we combined physiological sampling with geographic information systems to quantify the effects of land use on physiological indicators of health in largemouth bass. More specifically, we first quantified blood metrics relating to nutrition, oxidative stress, and the glucocorticoid stress response from largemouth bass residing in eight watersheds. We then used Akaike’s information criterion to define relationships between these blood metrics and land cover, including forests, agricultural areas, urban areas, and wetlands. The proportion of forest cover in a watershed was the best predictor of blood metrics representing recent feeding and resistance to oxidative stress, whereas the proportion of wetlands was the best predictor of glucocorticoid function; however, further investigation is needed, as the explanatory power of the models was relatively low. Patterns in energy reserves were not influenced by any land use practices. Interestingly, anthropogenic land use categories, such as urban and agricultural areas, were not the best predictor for any blood metrics. Together, our results indicate that fish health is most related to natural features of a landscape rather than anthropogenic land uses. Furthermore, these findings suggest that physiological methods could supplement traditional population and community assessments to develop a more comprehensive understanding of ecosystem interactions and improve stream management.