A Simple Mixed-Effects Model to Smooth and Extrapolate Weights-at-Age for 3Ps Cod

Abstract

Good estimates of weight-at-age for fishery catches and the stock are necessary for more reliable stock assessment and projections. I use a simple model to filter out “noise” in the 3Ps cod weight-at-age estimates, and to fill in missing values, especially for older ages and the age 14+ group that will be used in assessment models for the stock. The weight model is applied to estimated weight-at-age from fishery monitoring activities and also weight-at-age from the Fisheries and Oceans Canada (DFO) Spring research vessel (RV) survey which are assumed to represent weight-at-age in the stock. The other important model inputs are information about the precision of sampling estimates of weight-at-age which the model uses to help distinguish between population variability and measurement error. The model fits the fishery weights-at-age closely for those ages with low measurement error coefficients of variation (CVs). The survey weights-at-age have more between-year variability and presumably higher CVs and therefore the model did not fit these data as well but did capture the overall trends in the weights-at-age over time.