What We Don’t Know About the Effects of Temperature on Salmonid Growth

Abstract

AbstractSalmonid biologists use both mental models—simple assumptions—and bioenergetics‐based mathematical models to understand and predict effects of temperature regime on growth. Bioenergetics model results, however, sometimes conflict with common assumptions. Previous studies plus “Wisconsin model” bioenergetics simulations lead to four conclusions that conflict with some management assumptions. The first conclusion is that food consumption is at least as important as temperature in explaining growth; we cannot understand temperature effects without understanding food consumption. Second, at natural food consumption rates, there is no “optimal temperature for growth”; growth peaks in model results are artifacts of food consumption assumptions, and growth peaks in laboratory studies are (apparently) artifacts of ad libitum feeding. Third, effects of temperature on growth can be stronger during cooler seasons than in summer; traditional temperature criteria are not useful for managing such effects. Fourth, salmonid populations that are adapted to survive higher temperatures may be more, not less, vulnerable to temperature effects on growth due to their higher metabolic rates. Temperature–growth relations observed under ad libitum feeding seem risky for managing wild populations. Model predictions of growth need to carefully consider assumptions about food consumption. For predicting effects of increased temperature, the traditional assumption that consumption is a constant fraction of maximum consumption rate appears especially uncertain and incautious, with its hidden assumption that consumption increases with temperature; assuming a constant ration is simpler and more cautious. Growth can be predicted more reliably with feeding models and individual‐based population models that consider how consumption and energetic costs depend on processes such as habitat selection, competition, and adaptive behaviors involving tradeoffs between food intake and predation risk. Two research needs are clear: empirical observations for parameterizing and testing the Wisconsin model comprehensively under natural conditions (which we lack despite the extensive energetics literature), and methods for predicting salmonid food production responses to temperature and flow regimes.