Comparison of Mechanistic and Empirical Methods for Modeling Embryo and Alevin Development in Chinook Salmon

Embryo and alevin development in chinook salmon Oncorhynchus tshawytscha were analyzed with mechanistic and empirical models. Output from a mechanistic model (model BA) based on differential equations that generates a trajectory of embryo, alevin, and yolk sac mass development between fertilization and first feeding was compared with similar output from empirical models specifically designed to make one of four predictions: time and weight at maximum alevin wet weight (MAWW) and time and weight at fry emergence (models Mt, Mw, Et and Ew, respectively). A coefficient of determination (R 2) was used to evaluate the effectiveness of each individual model and was the basis for comparison between the two different types of predictions. For constant temperatures, the empirical models were comparable, based on the coefficient of determination (R 2 Mt = 0.991, R 2 Mw = 0.967, R 2 Et = 0.969, and R 2 Ew = 0.839; the equivalents for model BA were 0.987, 0.935, 0.984, and 0.825). Under varying temperatures, model BA had a slight advantage for emergence weight (R 2 BA = 0.766 versus R 2 Ew = 0.733); for emergence time, the difference between the R 2 values was the greatest of all the cases (R 2 BA = 0.951 versus R 2 Et = 0.812). If intraspecific differences in development rates can be attributed entirely to the local temperature profile and egg size, which are inputs to these models, then model BA has the potential to predict the growth of any chinook salmon egg.