Linking growth of juvenile sockeye salmon to habitat temperature in Alaskan lakes

We examined the influence of temperature on the size of age-1 smolts of soc keye salmon Oncorhynchus nerka relative to that of food availability and de nsity dependence across a variety of Alaskan lakes. We analyzed data from 3 6 lakes providing 134 lake years (annual means) of data. These lakes repres ent clear, organically stained, and glacially turbid lake types, They produ ce age-1 sockeye salmon smelts exhibiting a threefold variation in average length (52-145 mm) and a 30-fold variation in average weight (1-30 g). We u sed simple regression analysis to test for linkages between age-1 smolt siz e and the independent variables zooplankton biomass (mg/m(2)), smolt densit y (number/km(2)), length of the growing season, and mean water column tempe rature. Zooplankton biomass (ZB) was the strongest single predictor of smol t size, accounting for 52% of the variation in mean age-1 smolt fork length (FL) and weight (W). Individually, smolt density (SMD) accounted for only 10% of the variance in age-1 smolt size: however, mean water column tempera ture (T-s) explained 24% of the variation in FL and 19% of the variation in W. Smolt size was unrelated to the length of the growing season (S); howev er, growth rate indexed by length varied inversely with S while growth in t erms of weight did not. Subsequently, we combined all factors into a multiv ariate model of age-1 smolt size and the population's biological and therma l environment. Taken together. the factors ZB, SMD, and T-s accounted for 7 0% of the variation in FL and W. The influence of temperature on smolt size appeared to be more direct (through its effect on metabolic rates) than in direct (through limiting food availability [plankton]), Our results may inf luence fishery management objectives and serve as a template to project fut ure trends in juvenile sockeye salmon growth under different climatic condi tions.