Comparison of various potential fecundity models for north-east Arctic cod Gadus morhua, L. using oocyte diameter as a standardizing factor

To build a better data foundation for recruitment models of north-east Arctic cod Gadus morhua the construction of fecundity models reflecting variation in the nutritional status of the fish was attempted. The models were based on fecundity time series covering 9 years within the period 1986-2004 and included both general and year-specific approaches. Initial data analysis revealed that the potential fecundity (F-P) (standing stock of vitellogenic oocytes) was significantly reduced as the vitellogenic oocytes increased in size towards the start of spawning. Histological examination strongly indicated that this seasonal reduction was caused by atresia. Regression analysis showed that the F-P was positively correlated to fish total length (L-T) and the Fulton's condition factor (K). A multiple regression including data for all years using fish L-T, K and mean oocyte diameter (D-O) as independent predictors described the F-P with an r(2) = 0.94. This was considerably higher than comparable univariate L-T or mass-based regressions. These univariate regressions had fairly high r(2) values when split by years, but not as high as found for year-specific multiple regressions. An important application for individual-based fecundity models may be to generate outputs that can be fed into stock level fecundity and recruitment models. Overall, the multivariate models seemed to be the most accurate. The multivariate model including mean D-O, however, also had the potential to correct for maturity and thus provide unbiased fecundity comparisons between years, stocks and locations. (c) 2006 The Fisheries Society of the British Isles.