A 5 year individual-based data set was used to estimate size-specific survival rates in a wild brook trout Salvelinus fontinalis population in a stream network encompassing a mainstem and three tributaries (1.5-6 m wetted width), western Massachusetts, U.S.A. The relationships between survival in summer and temperature and flow metrics derived from continuous monitoring data were then tested. Increased summer temperatures significantly reduced summer survival rates for S. fontinalis in almost all size classes in all four sites throughout the network. In contrast, extreme low summer flows reduced survival of large fish, but only in small tributaries, and had no significant effects on fish in smaller size classes in any location. These results provide direct evidence of a link between season-specific survival and environmental factors likely to be affected by climate change and have important consequences for the management of both habitats and populations.

1. Modelling the effects of climate change on freshwater fishes requires robust field-based estimates accounting for interactions among multiple factors. 2. We used data from an 8-year individual-based study of a wild brook trout (Salvelinus fontinalis) population to test the influence of water temperature on season-specific growth in the context of variation in other environmental (i.e. season, stream flow) or biotic factors (local brook trout biomass density and fish age and size) in West Brook, a third-order stream in western Massachusetts, U.S.A. 3. Changes in ambient temperature influenced individual growth rates. In general, higher temperatures were associated with higher growth rates in winter and spring and lower growth rates in summer and autumn. However, the effect of temperature on growth was strongly context-dependent, differing in both magnitude and direction as a function of season, stream flow and fish biomass density. 4. We found that stream flow and temperature had strong and complex interactive effects on trout growth. At the coldest temperatures (in winter), high stream flows were associated with reduced trout growth rates. During spring and autumn and in typical summers (when water temperatures were close to growth optima), higher flows were associated with increased growth rates. In addition, the effect of flow at a given temperature (the flow-temperature interaction) differed among seasons. 5. Trout density negatively affected growth rate and had strong interactions with temperature in two of four seasons (i.e. spring and summer) with greater negative effects at high temperatures. 6. Our study provided robust, integrative field-based estimates of the effects of temperature on growth rates for a species which serves as a model organism for cold-water adapted ectotherms facing the consequences of environmental change. Results of the study strongly suggest that failure to derive season-specific estimates, or to explicitly consider interactions with flow regime and fish density, will seriously compromise our ability to predict the effects of climate change on stream fish growth rates. Further, the concordance we found between empirical observations and likely energetic mechanisms suggests that our general results should be relevant at broader spatial and temporal scales.