Modelling feeding, growth, and habitat selection in larval Atlantic cod (Gadus morhua): observations and model predictions in a macrocosm environment

Individual-based models (IBMs) integrate behavioural, physiological, and developmental features and differences among individuals. Building on previous process-based models, we developed an IBM of larval Atlantic cod (Gadus morhua) that included foraging, size-, temperature-, and food-limited growth, and environmental factors such as prey-field, turbulence, and light. Direct comparison between larval fish IBMs and experimental studies is lacking. Using data from a macrocosm study on growth and feeding of larval cod, we forced the model with observed temperature and prey-field and compared model predictions with observed distribution, diet, size-at-age, and specific growth rates. We explored implications of habitat selection rules on predicted growth rates. We analyze the sensitivity of model predictions by the Latin Hypercube Sampling method and individual parameter perturbation. Food limitation prevented larvae from growing at their physiological maximum, especially in the period 5-17 days post hatch (DPH). Active habitat selection had the potential to enhance larval growth rates. The model predicted temperature-limited growth rates for first-feeding larvae (5-20 DPH) when prey density is > 5 nauplii-L-1. After age 20 DPH, maximum modelled growth required a diet of copepodites. Simulated growth rates were.close to observed values except for the period just after the start of exogenous feeding when prey density was low.