2.3 Age/sex class energy patterns
Individual body measurements collected at capture were used to estimate energetic metrics for each bear. Straight-line body length and axillary girth were used to estimate body mass using regression equations in Table 2 from Thiemann et al. (2011b) and then energy density (MJ kg-1) and storage energy (MJ) were calculated using equations 18 A-E from Molnár et al. (2009).
Energy density and storage energy trends over time for each demographic class were analyzed using linear regression models. In addition, multiple linear regression models (Table S1 in Supporting Information) were defined a priori based on ecological hypotheses and were used to assess the relationship between energy density or storage energy for each class and the environmental variables (sea ice breakup, length of the open water period, AOw, NAOw, and lagged effects). Environmental variables were assessed for collinearity and variables that were correlated (r > |0.6|) were not included in the same model (Table S2). Model selection was determined using Akaike’s Information Criterion (AIC).
As the energy density and storage energy values were non-normally distributed (Shapiro-Wilk test, P ≤ 0.05) and standard transformations did not improve normality, we used Kruskal-Wallis ANOVA and Dunn’s non-parametric tests to examine differences among age/sex classes.