4.1 Estimated level of intrapopulation feeding diversity
We used cross-sectional sampling to calculate \(\text{PS}_{i}\), a specialization metric, as an estimate of intrapopulation feeding diversity in Salish Sea harbor seals. Our data confirmed intrapopulation feeding diversity across the spatial (hundreds of km) and temporal (years) scales that the scat samples represented (average\(\text{PS}_{i}\) = 0.399, 95% CI = 0.026, R = 100,000). These data demonstrate intrapopulation feed diversity but leave room for two alternative hypotheses (which cannot be separated in this case), regarding the absolute level of individual specialization: the occurrence of congruent long-term generalists and short-term specialists, or the occurrence of long-term specialists.
4.2 Importance of Time of year, Sex, Location, and Year on relative specialization ( \(\text{PS}_{i})\)
Month was an important predictor of relative specialization because removing it from the model caused a large drop in goodness-of-fit measurements (Table 2). This pattern makes intuitive sense as the type of prey eaten by harbor seals (Lance et al., 2012; Olesiuk et al., 1990) as well as their dive foraging behavior (Wilson et al., 2014) vary throughout the year. Therefore, changes in foraging behavior (both prey choice and dive type) were likely mechanisms behind the observed change in relative specialization throughout the year. However, there were likely other factors influencing relative specialization in addition to month.
Sex also had an impact on relative specialization, yet smaller than that of Month (Table 1). Differences in the level of relative specialization between female and male harbor seals were likely due to females and males in the region eating different prey items and having different foraging strategies (Bjorland et al., 2015; Schwarz et al., 2018; Wilson et al., 2014). For instance, females more often perform deeper foraging dives, eat benthic prey more commonly, and have smaller home ranges than males ((Peterson, Lance, Jeffries, & Acevedo-Gutiérrez, 2012; Schwarz et al., 2018; Wilson et al., 2014). Therefore, we propose the following theoretical resource distribution: males have more overlap between individuals while the females have less overlap between individuals; variations in this pattern appear to be associated with prey type (which will be addressed in the following section) (Figure 6).
Including an interaction term between Month and Sex increased the goodness-of-fit of the model (Table 2). This result indicates that differences between male and female seals likely varied throughout the year. Specifically, there were clear decreases in relative specialization in male harbor seals during the summer and fall months that were not reflected in females (Figure 2), indicating that the behavior of both sexes was similar in the spring but diverged in the summer and fall. This behavior was likely due to changes in feeding patterns of females and males throughout the year (Lance et al., 2012; Wilson et al., 2014). A possible reason for the different feeding patterns in the summer months is female change in behavior due to pupping (Ternte, Bigg, & Wigg, 1991). While nursing, females spend most of their time on the haul-out and make short foraging trips (Boness, Bowen, & Oftedal, 1994; D’Agnese, 2015). A similar difference was seen between sexes during the fall; however, both sexes were relatively least specialized during the fall. During the fall, there is a large influx of returning adult Salmoniformes (Quinn, 2005) that are preyed upon by both female and male harbor seals (Schwarz et al., 2018). In the Salish Sea, Salmoniformes can compose >50% of the population diet in the summer and fall (Lance et al., 2012). This resource could be rich enough that it is beneficial for a majority of seals, both males and females, resulting in less need for specialization. This explanation is further supported by the correlation between feeding on adult Salmonifomes and a relatively less specialized diet (Table 3), indicating it was a widely used resource in the region.
Our data also suggest that location explained a large amount of variance in relative specialization. The random factors of Year and Location increased the \(r^{2}\) by more than four times, indicating that both had a large influence on relative specialization. However, because Sample Size, Location, and Year explained 0.39, 0.36, and 0.002 of the variance (SD = 0.62, 0.597, 0.05), respectively, one can assume that Sample Size and Location were the random factors responsible for the increase in goodness of fit of the model, not Year. This result indicates that where the seals were foraging impacted the level of relative specialization in the population, without noticeable changes from year to year. Our results also indicate that there was likely some bias introduced by the number of samples in a group. For instance, there was a correlation between average \(\text{PS}_{i}\) and theoretical minimum \(\text{PS}_{i}\) (rho = -0.231, p = 0.03). However, this potential bias is unlikely to have had a substantial effect on the outcome of our study because we included sample size as a random variable in the model and variation in sample size does not appear to explain the seasonal changes in \(\text{PS}_{i}\) (Figures 2, 3).