Experimental evidence of gradual size-dependent shifts in body size and growth of fish in response to warming.

A challenge facing ecologists trying to predict responses to climate change is the few recent analogous conditions to use for comparison. For example, negative relationships between ectotherm body size and temperature are common both across natural thermal gradients and in small-scale experiments. However, it is unknown if short-term body size responses are representative of long-term responses. Moreover, to understand population responses to warming, we must recognize that individual responses to temperature may vary over ontogeny. To enable predictions of how climate warming may affect natural populations, we therefore ask how body size and growth may shift in response to increased temperature over life history, and whether short- and long-term growth responses differ. We addressed these questions using a unique setup with multidecadal artificial heating of an enclosed coastal bay in the Baltic Sea and an adjacent reference area (both with unexploited populations), using before-after control-impact paired time-series analyses. We assembled individual growth trajectories of ~13,000 unique individuals of Eurasian perch and found that body growth increased substantially after warming, but the extent depended on body size: Only among small-bodied perch did growth increase with temperature. Moreover, the strength of this response gradually increased over the 24 year warming period. Our study offers a unique example of how warming can affect fish populations over multiple generations, resulting in gradual changes in body growth, varying as organisms develop. Although increased juvenile growth rates are in line with predictions of the temperature-size rule, the fact that a larger body size at age was maintained over life history contrasts to that same rule. Because the artificially heated area is a contemporary system mimicking a warmer sea, our findings can aid predictions of fish responses to further warming, taking into account that growth responses may vary both over an individual's life history and over time.