The best age-based and length-based models (see Table 2) evaluated using an equivalent dataset to allow for model comparisons.

Where age = the numeric fish age, length = individual fish length, lon = longitude in decimal degrees, lat = latitude in decimal degrees, year = the year sampling occurred as a factor, haul = individual hauls that samples originated from, incorporated as a random effect, (e)df = model estimated degrees of freedom, AIC = Akaike information criterion, Δi = AIC difference, AIC weights is the relative likelihood of a model, f = smooth terms. Age and length models were fit on different datasets as more length data are available.