Modeling Juvenile Salmonid Hatchery Growth Using a Local Equilibrium Assumption and Measured Water Fraction to Parameterize Fish Energy Density

• Statistical correlations using water content are used to calculate proximate composition.
• The proximate composition is a bioenergetic model result rather than a model input.
• The model closely matches the measured coho growth and proximate composition.

Previous reports have described bioenergetic models for juvenile salmonid growth that were validated using data from laboratory and production-scale feeding studies. Time variable functions and power correlations with body weight based on measurement of proximate composition were used to delineate the fish energy density in these models. Although model simulations for growth corresponded closely with measured data using either characterization of energy density, there is significant laboratory burden associated with the measurement of proximate composition throughout the simulation period, and this approach may therefore be impractical for many hatchery applications. Here measurements of fish water fraction are used to estimate the mass of protein and ash using correlations developed in a companion paper. Lipid is subsequently determined by difference, and energy content is calculated using the proximate composition. Model simulations of juvenile coho (Oncorhynchus kisutch) growth and proximate composition using this approach closely matched measured data when provided steady satiation and maintenance rations. The model was less successful predicting detailed protein and lipid dynamics following large abrupt decreases in the food supply rate. However, model simulation errors associated with fish growth were small because of the off-setting nature of the proximate composition model errors. It is concluded that coho water fraction measurements and correlations for protein and ash offer a practical and convenient alternative to direct measurement of fish energy density for many bioenergetic model hatchery applications.