Biology, Assessment, and Management of North Pacific Rockfishes 21 Alaska Sea Grant College Program • AK-SG-07-01, 2006 Rockfish Trophic Relationships in Prince William Sound, Alaska, Based on Natural Abundance of Stable Isotopes Thomas C. Kline Jr. Prince William Sound Science Center, Cordova, Alaska Abstract About two dozen rockfish species coexist within Prince William Sound (PWS). According to ecological theory these species should reduce competition by minimizing diet overlap. This was verified by using eco- logical metrics based on the natural abundance of carbon and nitrogen stable isotopes. Carbon source dependency was based on a regional carbon isotope gradient whereas trophic level was based on trophic enrichment of 15N. There was a gradient in carbon source dependency among rockfish slope, pelagic, and demersal eco-groups. Pelagic rock- fishes had the greatest dependency on Gulf of Alaska (GOA) derived car- bon. Within demersal rockfishes, copper rockfish were most dependent on PWS derived carbon. Rockfishes did not respond to a strong pulse in GOA subsidies, like forage groups, confirming that their stable isotope composition reflected relatively longer time-integration. Yelloweye and quillback rockfish shifted concordantly in carbon source dependency but separated in terms of relative trophic level. Relative trophic level (TL), based on the nitrogen isotope value of a herbivorous copepod for reference TL = 2.0, ranged from 3.2 (juveniles) to 5.1 (shortspine thorny- head). Another slope species, the shortraker rockfish, had the second highest TL = 4.9. Dark rockfish was the lowest adult TL = 3.6. Alternative nonlethal sampling for isotope values using blood is possible given normalization of the data for lipid isotope effects. Introduction In the northeast Pacific Ocean, the scorpaenid fishes of the genera Sebastes and Sebastolobus, commonly called rockfishes, comprise more 22 Kline—Rockfish Trophic Relationships Table 1. Twenty-five rockfishes found in PWS from the literature and this study. Meyer Rosenthal This Species Common name 1992 1980 study Sebastes melaonostomus Blackgill rockfish × Sebastes melanops Black rockfish × × × Sebastes paucispinis Bocaccio × Sebastes auriculatus Brown rockfish × × Sebastes pinniger Canary rockfish × Sebastes nebulosus China rockfish × × × Sebastes caurinus Copper rockfish × × × Sebastes ciliatus Dark rockfisha × × × Sebastes elongatus Greenstripe rockfish × Sebastes variegatus Harlequin rockfish × Sebastes alutus Pacific ocean perch × Sebastes emphaeus Puget Sound rockfish × Sebastes maliger Quillback rockfish × × × Sebastes babcocki Redbanded rockfish × Sebastes proriger Redstripe rockfish × × Sebastes helvomaculatus Rosethorn rockfish × Sebastes aleutianus Rougheye rockfish × × Sebastes zacentrus Sharpchin rockfish × Sebastes borealis Shortraker rockfish × × Sebastes brevispinis Silvergray rockfish × × × Sebastes diploproa Splitnose rockfish × Sebastes nigrocinctus Tiger rockfish × × × Sebastes ruberrimus Yelloweye rockfish × × × Sebastes flavidus Yellowtail rockfish × × × Sebastolobus alascanus Shortspine thornyhead × × aFor nomenclature see Orr and Blackburn 2004. than 60 species. Twenty-five of these species have been found in Prince William Sound (PWS), Alaska, a subarctic marine ecosystem of about 100 by 100 km (Table 1). Because much of PWS is north of 60ºN, this is the northern limit for some rockfish species, while others are also found in the higher latitude portions of the Bering Sea. PWS has depths to about 800 m thus providing a good range of depths accommodating the varying depth range preferences among the many rockfish species (Love et al. 2002). Ecological theory, which suggests rockfish species coexisting within PWS reduce would compete by minimizing diet overlap, was tested Biology, Assessment, and Management of North Pacific Rockfishes 23 for adults of twelve rockfish Sebastes( and Sebastolobus) species and juvenile Sebastes using stable isotope analysis (SIA) as a metric of diet overlap (Welch and Parsons 1993, Murie 1995). Application of SIA methods was enhanced by an existing isotopic context of potential prey sampled at the same time as most of the rock- fishes (Fig. 1). A regional stable carbon isotope gradient can be used to differentiate the relative dependency of autochthonous production sources (from PWS) versus allochthonous sources assumed to be from the Gulf of Alaska (GOA; Kline 1999). Trophic level (TL) can be estimated based on nitrogen stable isotope values assuming a consistent nitrogen isotope trophic enrichment (Kline and Pauly 1998; Kline 2001, 2002; Wada et al. 1991). The mean δ13C′ value (see Materials and methods for definition) of whole individual Neocalanus copepods analyzed from GOA was –22.7 (SD = 2.6, SE <0.1, N = 1,590) during 1995-2004; whereas that from PWS was –19.8 (SD = 2.0, SE <0.1, N = 758) (Kline 1999, 2001, 2002, and unpubl. data for post-1996). Carbon δ13C′ values less than about –20.5 were hypothesized to reflect allochthonous carbon with respect to PWS, reflecting GOA production sources. In contrast, theδ 15N gradient is much weaker (Kline 1999). The decadal mean δ15N for PWS was 8.9 (SD = 1.9, SE <0.1, N = 752); whereas GOA was 7.3 (SD = 2.5, SE <0.1, N = 1,588). Because of the gradient, though small, when a δ15N value of 8.4 is used as the reference for calculating TL, the uncertainty is about 0.3 TL (Kline 2001). Nevertheless, there was good agreement between TL based on δ15N and TL based on predictions made by the Ecopath model (Kline and Pauly 1998). Furthermore, TL of juvenile herring, which is possibly the most important forage fish in PWS, based onδ 15N was consistent, about 3.2 ± <0.2 over a four-year period (Fig. 1). A goal of this study was to determine whether PWS rockfishes parti- tion food resources in terms of carbon source and relative trophic level with respect to species as well as the three depth range eco-groups: de- mersal (shelf species associated with substrates), pelagic (shelf species associated with the upper water column), and slope (deep-dwelling spe- cies, generally found deeper than 200 m). Accordingly, synoptic mean 13 15 δ C′TL and mean δ N values of eco-groups or species should differ. Rockfishes are long-lived with relatively slow growth. The stable isotope values of adults were thus expected to reflect or integrate longer time spans than fishes with short life spans (cf. the 2 to 6 years of Pacific salmon species). Accordingly, their isotope values were not expected to vary much over times of less than ~1 year (Hesslein et al. 1993). For example, the large shifts in δ13C′ value were not expected, in response to a hypothesized pulse of GOA carbon during late 1995 (Kline 1999) observed in several taxa of fast-growing, lower trophic level organisms (Fig. 1). 24 Kline—Rockfish Trophic Relationships −19.5 −20.0 −20.5 −21.0 −21.5 −22.0 −22.5 13 E CaTL −23.0 1994B 1995B 1996B 1997B 1994A 1995A 1996A 1997A 1998A 4.2 TL 4.0 3.8 3.6 3.4 3.2 3.0 2.8 2.6 1994B 1995B 1996B 1997B 1994A 1995A 1996A 1997A 1998A Capelin Northern lampsh Eulachon Northern smoothtongue Glass shrimp Sandlance Juvenile herring Squid Juvenile pollock Figure 1. Trophic level and lipid normalized stable carbon isotope values 13 (δ C′TL, upper panel) and nitrogen stable isotope value–based trophic levels (TL, lower panel) of forage taxa sample in PWS, 1994-1998. See text for explanation of how years were split. Error bars indicate standard error. Biology, Assessment, and Management of North Pacific Rockfishes 25 Rockfishes prefer rocky, high relief habitats. Western Prince William Sound (W-PWS) is characterized by deep fjords with depths reaching 800 m within 2 km of shore. Depths are <300 m in eastern Prince William Sound (E-PWS), the portion of PWS east of the tanker lanes exclusive of the area near (<30 km) this boundary. Nevertheless, there are loca- tions with rocky habitats in E-PWS with populations of shallow-dwelling rockfish species. A secondary objective was to ascertain if carbon or TL differed qualitatively across PWS. A long-term goal is to recover tagged rockfishes and resample over time for SIA to detect ecosystem shifts. A nonlethal sample would thus be required. The potential for this was tested by tagging and recover- ing rockfishes and by comparing SIA of blood with SIA of muscle tissue. Muscle is typically sampled in most SIA studies after killing the organ- ism being sampled. Materials and methods Rockfishes and potential prey taxa were sampled incidentally during the Sound Ecosystem Assessment (SEA) project that focused on pink salmon, Pacific herring, and walleye pollock during 1994-1999 (Cooney et al. 2001). Samples were obtained on SEA cruises by seines, trawls, gillnets, and hook and line. Additional rockfish samples from the PWS longline fishery were obtained from Cordova fish processors during SEA. Samples for blood were collected by hook and line or by hand-nets during scuba diving as part of the Oil Spill Recovery Institute Sentinel project in 1999. Rockfish samples were organized according to the fol- lowing three depth-based ecological categories: demersal, pelagic, and slope favored by each species based on their distribution in Alaska (Kramer and O’Connell 1995; Table 3). Rockfishes newly recruited from the plankton have a different appearance from adults and were simply classified as juveniles for the purposes of this study. Rockfish samples were opportunistic and thus not systematic in terms of sampling date or location. Samples came from a range of sites in PWS, which were divided into E-PWS and W-PWS. Data were temporally aggregated to match that of forage data described below. Approximately 1 g of anterior epaxial muscle tissue was collected from each adult rockfish (except for a single ~20 cm long copper rock- fish sampled only for blood), frozen, freeze-dried, then ground to a fine powder. Juvenile rockfish and forage taxa were treated similarly except that the entire organism was ground due to their size (length ≤15 cm). Blood samples taken of fishes collected during July 1999 via caudal vein puncture using non-heparin 3.0 ml syringes were centrifuged, then plasma and cell fractions were frozen and later freeze-dried and agitated to a fine power with a dental amalgamator.