Ontogenetic, Temporal, and Spatial Variation of Feeding Niche in an Unexploited Population of Walleye Pollock (Theragra chalcogramma)

Stable isotope ratios of carbon and nitrogen were measured in an unexploited population of walleye pollock (Theragra chalcogramma) from Prince William Sound, Alaska (PWS) as a metric of feeding niche. Stable isotope data of pollock were compared with each other and with stable isotope data of age-0 herring (Clupea pallasii). There were significant species, location, and time period differences in synoptically sampled age-0 pollock and herring. Age-0 pollock were generally 13C-enriched relative to herring even as carbon isotope values varied with time. A positive relationship between 13C content and length of pollock from Zaikof Bay, PWS, was consistent with a pattern of low 13C being attributable to food subsidies, most likely oceanic zooplankton. The carbon isotope difference between age-0 pollock and age-0 herring increased when 13C decreased. This coincided with a relative increase in age-0 pollock food chain length compared to age-0 herring. This is posited to reflect reduced overlap in food supply and thus competition during a period of inferred high food subsidies. Subsidies are further posited to vary according to hypothesized oceanic zooplankton population cycles. A systematic increase in nitrogen isotope values with respect to size suggested an average ontogenetic food chain length increase of about 1.3 trophic levels over a pollock’s life span after age-0. Trophic level variability was consistent with facultative planktivory for all adult sized pollock, including those with lengths over 400 mm, above which there was a sharp trophic level increase. The ability for pollock to maintain 252 Kline—Variation of Feeding Niche a relatively low trophic level for most of their life is hypothesized to enhance the species’ ability to sequester energy. Stable isotopes may provide a useful metric for detecting trophic change in terms of subsidies as well as food chain length. Introduction An unexploited population of walleye pollock, Theragra chalcogramma, the dominant gadid of the subarctic Pacific Ocean, was discovered in Prince William Sound, Alaska (PWS), during the mid-1990s (Willette et al. 1999). Pollock, along with Pacific herring (Clupea pallasii) and juvenile pink salmon (Oncorhynchus gorbuscha) were found to be the dominant constituents of the PWS pelagic fish fauna (Cooney et al. 2001). Willette et al. (1999) found that as well as being facultative planktivores, immature and adult pollock preyed upon nekton in PWS. Squid were the most common nekton prey of adult pollock. The nekton prey of immature pollock consisted of juvenile gadids, mainly pollock, as well as salmonids, mainly pink and chum salmon (O. keta). The proportion of nekton in diets was inversely correlated with abundance of large calanoid copepods (Willette et al. 1999). Pollock switched to principally consuming large calanoids and other zooplankton above certain threshold levels. This prey-switching facility has also been observed in Bering Sea pollock (e.g., Dwyer et al. 1987). Age-0 pollock, which were consistently sampled alongside age-0 herring in protected bays of PWS (Paul and Paul 1998, Paul et al. 1998, Kline 1999), fed on food chains of approximately the same length, which were consistent with zooplanktivory (Kline 2001, 2007). Stable isotope ratios are useful for food web analysis because of predictable relationships in the isotopic compositions of consumers and their forage (reviewed by Michener and Schell 1994). Age-0 pollock and age-0 herring, as well as other “forage fish” taxa, reflect carbon subsidies from the adjacent Gulf of Alaska by concomitant shifts in stable isotope composition throughout PWS (Kline 1999, 2007). Continental slope zooplankton populations were posited to be the source of occurrences of low 13C/12C abundance occasionally measured in PWS fishes (Kline 1999). Low 13C/12C values measured in PWS fishes were thus posited to reflect oceanic carbon subsidies. The natural abundance of 13C/12C expressed as δ13C′ (see Materials and methods) of PWS zooplankton was more than –21‰ whereas that of zooplankton of the same species sampled in the Gulf of Alaska was less than –22‰ (Kline 1999). Because of trophic enrichment of 13C of ~1‰ per feeding step (reviewed by Michener and Schell 1994), zooplanktivores such as age-0 pollock feeding exclusively on one of these sources would have correspondingly higher values than these, which was reflected in actual observations (Kline 1999). 253 Resiliency of Gadid Stocks to Fishing and Climate Change Whereas oceanic subsidies were observed in all forage taxa in late 1995, the relative effect in terms of inferred dependency based on δ13C′ values varied by species (Kline 2007). Sand lance (Ammodytes hexapterus), a nearshore fish species, had the highest δ13C′ values prior to late 1995 and had the highest δ13C′ values during late 1995. However, their δ13C′ values during late 1995 were lower than for herring prior to the shift. Age-0 pollock were second to herring in terms of having the greatest qualitative response to oceanic subsidies. Interactions among species in PWS in their response to oceanic subsidies of potential food can thus be inferred from differences in 13C/12C composition. Age-0 pollock in PWS were found to increase in energetic content during the winter, whereas age-0 herring lost energy, which was hypothesized to reflect species differences in foraging strategies (Paul and Paul 1998, Paul et al. 1998). To test whether differences in foraging strategy could be detected with stable isotopes, in particular differences in response to oceanic subsidies, this paper examines new as well as previously published age-0 pollock and herring stable isotope data. This context required greater detail in the analysis than previous studies (Kline 1999, 2007). Kline 1999 was limited to qualitative comparison of juvenile pollock and herring from two time periods, autumn 1994 and autumn 1995. Additional data of age-0 pollock from 1994 and 1996 are presented here. Supplementing these data are age-0 herring data for those time periods for which age-0 pollock data are available. These data are examined quantitatively by species, sampling period, and sampling site to assess the relative response of oceanic subsidies in terms of δ13C′ value. This is a finer scale analysis than done previously (Kline 2007). Additionally, previously unpublished stable isotope data of larger size classes of pollock are presented to consider ontogenetic relationships, if any, to oceanic subsidies as well food chain length, or trophic level (TL), which is reflected by δ15N value. Materials and methods Fish sampling Pollock and herring were sampled primarily by trawling and seining (Kline 1999). Samples were collected for multiple projects. The juvenile pollock and herring samples analyzed for stable isotopes were also analyzed for whole-body energetic content reported by Paul et al. (1998). The scope of sampling for pollock varied in time (Tables 1 and 2). Sampling in 1994 was limited whereas broadscale surveys in OctoberNovember 1995 and March 1996 covered PWS in its entirety (Fig. 1). Subsequent sampling focused on four representative bays (Norcross et al. 2001). There are five sampling periods, A to E (Table 1), for which both age-0 pollock and age-0 herring stable isotope data are available. Sampling periods C and D, corresponding to the period when exten254 Kline—Variation of Feeding Niche sive surveys were conducted, also corresponded to when there was an inferred strong pulse of oceanic subsidies (Kline 2007). Sampling periods A, B, and E are assumed to be a period when oceanic subsidies were lower than that inferred during C and D and collectively are referred to as the low subsidy period. Pollock over 125 mm in length were sampled at the same times as age-0 fish, but were fewer in total number (Table 2). Accordingly, these larger size classes of pollock were stratified into two time-integrated temporal groups, one corresponding to the inferred high subsidy period, and one corresponding to the low subsidy period. Only two of the 209 pollock from the high subsidy period were sampled during D. Within these two temporal strata, high and low subsidy, pollock over 125 mm in length were stratified by sampling site in PWS (Table 2). Epaxial muscle was dissected from larger size class pollock, whereas whole age-0 pollock and herring were analyzed. Fish tissues were freeze-dried and then ground to a fine powder. Powdered samples were sent to the University of Alaska Fairbanks Stable Isotope Facility 12 2