Additional samples from different production cycles were used to validate the model, and the predicted values were not statistically distinguishable from the observed values regardless of fish species (P<0.05).