Relating trends in walleye pollock (Theragra chalcogramma) abundance in the Bering Sea to environmental factors

Abstract

Generalized additive models (GAM), a nonparametric regression method with less restrictive statistical assumptions than traditional regression methods, were used to model the trend in mean abundance of Bering Sea walleye pollock (Theragra chalcogramma) as a function of ocean environmental conditions including water column depth, temperature at 50 m, and depth of the thermocline. Acoustic survey data collected in the summers of 1988 and 1991 were used to test these relationships. In both surveys, mean walleye pollock abundance was highest in areas having a 70–130 m depth range and where the 50-m temperature was close to 2.5 °C. Thermocline depth, while not itself significant, had a significant effect on walleye pollock abundance through interactions with both bottom depth and temperature at 50 m. Walleye pollock in the top 50 m of the water column (mostly juveniles) were influenced differently by temperature and thermocline depth than the adult walleye pollock, which were generally deeper in the water column. The depth, temperature, and thermocline preferences of walleye pollock are hypothesized to be linked to food availability which is, in turn, related to temperature regimes or fronts along the Bering Sea shelf slope.

Spatial and temporal trends in condition of Atlantic cod Gadus morhua on the Icelandic shelf

“…where deviance was analogous to the residual sums of squares (Swartzman et al 1995). The contribution of each effect to the total pR 2 was also calculated.…”

Critical timing for reproductive allocation in a capital breeder: evidence from sandeels

“…An additional measure of model fit was based on a pseudo-coefficient of determination, R 2 , which was taken to be the fraction of the total variation explained by the model (Swartzman et al 1995): (4) Fecundity versus size and condition relationships across the 2 tanks were then fitted within the framework of a general linear model (GLM) (Sokal & Rohlf 1995). The model fitted to the data included all explanatory variables before non-significant effects were removed by a process of step-wise deletion.…”

Essential relationships incorporating the influence of age, size and condition on variables required for estimation of reproductive potential in Atlantic cod Gadus morhua

“…Generalized linear models (GLM, binomial distribution) were then used to examine the influence of length, age and condition on maturity. An additional measure of model fit was based on a pseudo-coefficient of determination (R 2 ), which was the fraction of the total 'variation' explained by the model where deviance was analagous to the residual sums of squares (Swartzman et al 1995). Analysis of variance (ANOVA) on log e -transformed data was used to compare length and condition among age groups of each sex in each region.…”