Sea Lions of the World 117 Alaska Sea Grant College Program • AK-SG-06-01, 2006

Geographical Variation in Steller Sea Lion Prey Quality in Alaska

Lawrence Schaufler NOAA, NMFS, Alaska Fisheries Science Center, Juneau, Alaska

Elizabeth Logerwell NOAA, NMFS, Alaska Fisheries Science Center, Seattle, Washington

Johanna Vollenweider NOAA, NMFS, Alaska Fisheries Science Center, Juneau, Alaska

Abstract Nutritional stress is one of the leading hypotheses explaining the decline in Steller sea lion populations of the western stock. Central to this hypothesis is the possibility that western stock sea lions encounter prey of significantly lower quality than those from the eastern stock. We collected and analyzed over 1,200 whole fish representing species identified as sea lion prey items from the Aleutian Islands and southeastern Alaska, including species that reside in both regions. We present proximate composition and calculated mean energy densities based on the lipid and protein contents for the sampled fish. Initial comparisons of the proximate compositions and energy densities between the Aleu- tian Islands and southeastern Alaska fish on a species basis revealed significant differences in prey energetic content in the two regions for the sampled prey. Overall, the mean energy density for 22 forage species from southeastern Alaska (1.62 ± 0.02 kcal per g on a wet weight basis) was greater than that of 15 species from the Aleutians (1.44 ± 0.03 kcal per g), but these variations could be attributed to size differences among the fish sampled from the two regions as well as species composition and collection season differences. For example, Pacific cod sampled from the Aleutians were significantly larger p( < 0.001) than those from southeastern Alaska and had a higher energy density (p < 0.001). However, controlling for size revealed no difference in energy density between the two populations of cod (p > 0.5). Similarly accounting for size, no difference was found in the energy density of walleye pollock or arrowtooth 118 Schaufler et al.—Prey Variation in Alaska flounder from the two locations. In contrast, squid and sandfish from southeastern Alaska had higher energy densities (p < 0.01) while Aleu- tian rockfish had higher energy densities than those from southeastern Alaska (p < 0.001), though these may represent seasonal and species composition differences. These data reveal the importance of considering size, season, and species when making energy density comparisons of the available prey between geographical regions.

Introduction The nearly 80% population decline of the western stock of Steller sea lions over the last 30 years has been attributed to a number of factors, including the “junk food hypothesis.” This theory postulates that a shift in diet from higher energy forage fish such as Pacific herring, to a diet consisting mainly of lower energy fish such as walleye pollock led to a nutritional deficiency (Alverson 1992, Trites and Donnelly 2003). The lack of a significant decline in the eastern Steller sea lion stock could indicate a difference in the quality of available food between the Aleutian (western) and southeastern Alaska (eastern) regions during the Steller sea lion population decline. Because pinnipeds rely heavily on a piscivorous prey base, prey quality becomes paramount during lactation, molting, and other periods of increased energetic need. Relevant prey quality issues include the amount of fat and protein, vitamin and essential fatty acid contents, and caloric value or energy density. In this study we focused on the fat and protein contents and estimated energy density aspects of prey quality. The prey items available to and consumed by the western and eastern Steller sea lion stocks vary significantly both in species composition and average prey size (Merrick et al. 1997, Zeppelin et al. 2004). Atka mackerel (Pleurogrammus monopterygius) and walleye pollock (Theragra chalcogramma) dominate the diets of Aleutian Steller sea lions, while forage fish such as Pacific herringClupea ( pallasii ), eulachon (Thaleichthys pacificus), and capelin (Mallotus villosus) as well as walleye pollock form the majority of the southeastern Alaska Steller sea lion diet (Winship and Trites 2003). Other species are common to both regions, such as Pacific cod (Gadus macrocephalus), arrowtooth flounder Atheresthes( stomias), and certain squid (Berryteuthis) and skates (Bathyraja), but compose different percentages of the western and eastern stock Steller sea lion diets. Differences in prey sizes have also been reported between the two regions, likely from adaptations to environmental conditions such as water temperature, available food sources, and the particular variety of predators in each region. (Shuter and Post 1990, Tollit et al. 2004, Zeppelin et al. 2004). This study examines prey quality (i.e., nutritional value of prey) available to Steller sea lion in the Aleutian Islands and southeastern Sea Lions of the World 119

Lynn Canal Frederick Sound Attu

Buldir Sitka Sound

Amchitka Akun Adak

Western Region Eastern Region (Aleutians) (Southeast AK)

Figure 1. Prey collection sites in the Aleutian Islands and southeastern Alaska.

Alaska. Proximate analysis was used to determine protein, lipid, and moisture content, along with calculated energy densities based on these values for prey collected in the two regions. We present comparisons of proximate composition and energy density values between species common to the regions and energy density values for prey found primarily in only one of the two regions. We report energy density on a wet weight basis since we are considering the nutritional value of a whole fish as it is consumed by a predator such as a marine mammal.

Methods Sample collection and storage Opportunistic sample collections were performed on various National Oceanic and Atmospheric Administration (NOAA) cruises and charter vessels in the Aleutian Islands and southeastern Alaska. Specimens were collected in the proximity of (Aleutians) Adak, Akun, Amchitka, Attu, and Buldir Islands; (southeastern Alaska) Berners Bay, Frederick Sound, Lynn Canal, and Sitka Sound (Fig. 1). Aleutian trawls occurred primarily during summer months (April through July), while southeast Alaska trawls were performed year-round, mostly on a quarterly basis 120 Schaufler et al.—Prey Variation in Alaska

(March, May, September, and December). Typical trawls were mid-water, approximately 20 minutes in duration, performed during daylight hours using a 164 Nordic rope trawl with 1.5 m2 alloy doors, 7 m × 17 m (height × width) with a 19 mm mesh codend liner. Trawling depths ranged from approximately 75 to 225 meters. Samples were frozen whole immediately after morphometric measurements and gender were recorded. Gender was determined by direct examination of the gonad, and all gut contents were vacuum-sealed along with the fish in individual bags. When practical due to fish size, specimens were quick-frozen in liquid nitrogen. When this protocol could not be implemented, fish were vacuum-sealed after gender determination and placed in single layers in a commercial-grade –20ºC freezer. Upon returning to the laboratory, samples were stored in a –20ºC freezer for short-term (0-3 months) or a –80ºC freezer for longer-term (4+ months) storage.

Proximate analysis Entire frozen fish were cut into cross-sections with a Bizerba FK23 in- dustrial meat saw, then homogenized in a Fleetwood M12S meat grinder using a 4.5 mm die. Three to five gram subsamples of the homogenate were randomly chosen for analysis and further liquified using an Oster 4134 blender with food processor attachment. Lipid content was determined gravimetrically after a modified Folch extraction employing 0.1% BHT as an antioxidant (Christie 2003, Vollenweider 2004) using a Dionex 200 accelerated solvent extractor (ASE). Protein content was determined with the Dumas method on a Leco FP-528 nitrogen analyzer and a 6.25 nitrogen-to-protein conversion factor was used (AOAC 1995). Moisture and ash contents were measured gravimetrically using a Leco TGA-601 ther- mogravimetric analyzer, heating at 135ºC for 2-3 hours and then 600ºC for 3-4 hours for moisture and ash, respectively. Carbohydrate contents were estimated by subtraction. All proximate contents (lipid, protein, moisture, and ash) are reported as percentages of total wet weight. All analyses implemented quality control procedures, including a sample replicate, a method blank with no sample material, and a reference standard with each group of 15-20 fish. For lipid analysis, an in- house herring composite reference sample was used to ensure sample group comparability. National Institute of Standards and Technology (NIST) Standard Reference Materials (SRM) 1946 and 2974 were analyzed for protein, moisture, and ash content to verify analytical accuracy. Protein analyses were performed on dried material, in duplicate, with samples reanalyzed if their deviation was more than 1.5 standard devia- tions from the mean. Sea Lions of the World 121

Data analysis For mean comparisons, Levene’s test was first performed to confirm homogeneity of variance, then a two-sample, two-tailed Student’s t-test was applied to determine significance if the variances were equal. In cases where the variances for the two data sets were not equal, Welch’s approximate t was determined, with the calculated degrees of freedom (DF) indicated. A general linear model was employed to identify covari- ates, and significance levels ofα = 0.05 were used for all tests.

Results Over 1,200 whole fish specimens were collected opportunistically from five locations in the Aleutian Islands and three sites in southeastern Alaska (Fig. 1). A total of 316 (26%) fish representing 15 species from the Aleutians and 915 (74%) fish representing 22 species from southeastern Alaska were analyzed. Five species were collected from both regions: arrowtooth flounder, Pacific cod, walleye pollock, Pacific sandfish, and Commander squid (Table 1). Species targeted for collection were based on known prey items for Steller sea lions (Sinclair and Zeppelin 2002, Winship and Trites 2003). Proximate analysis was performed on all fish samples and the results are presented in Table 1. Lipid and protein contents were used to estimate energy density values using conversion factors: 9.45 kCal per g for lipid (Brody 1945) and 5.65 kCal per g for protein (Van Pelt et al. 1997, Payne et al. 1999). Carbohydrates were typically <1.5% of the total weight, and did not significantly contribute to caloric value. Overall, the prey items examined contained an average of 7.8% lipid, 14.9% protein, 75.4% moisture, and 2.5% ash (mineral). The average energy density for all southeast Alaska prey sampled was 1.62 ± 0.02 kCal per g while the sampled Aleutian prey averaged 1.44 ± 0.03 kCal per g. This significant difference in overall energy densities for prey items from the two regions

(t633 = –5.22, p < 0.001) was primarily due to the lower average lipid content of the Aleutian prey. Aleutian items had a higher average protein content, but protein has a significantly lower caloric value than lipid. Of the species collected only in the Aleutians, Atka mackerel had the highest energy density of 1.93 ± 0.05 kCal per g. Of the species collected exclusively from southeast Alaska, lampfish (2.64 ± 0.02 kCal per g), eulachon (2.22 ± 0.03 kCal per g), and herring (2.04 ± 0.05 kCal per g) had the highest energy densities. Atka mackerel, collected during the summer spawning season, also differed in energy density between males (2.20 ± 0.09 kCal per g) and females (1.77 ± 0.07 kCal per g), as has been previously observed (Logerwell and Schaufler 2005). None of the species collected from southeastern Alaska displayed a significant gender difference in energy density. 122 Table 1. Average proximate composition and energy densities for sampled prey items.

Energy Collection density Speciesa Region monthb N (kCal/g) % Lipidc % Proteinc % Moisture Length (mm)d Weight (g) All Aleutians 316 1.44 ± 0.03 5.94 ± 0.29 15.51 ± 0.07 76.56 ± 0.28 All SE AK 915 1.62 ± 0.02 8.40 ± 0.21 14.62 ± 0.07 74.99 ± 0.20 All Combined 1,231 1.57 ± 0.02 7.77 ± 0.17 14.85 ± 0.05 75.39 ± 0.17 Arrowtooth Aleutians May-Jun 34 1.38 ± 0.08 5.33 ± 0.80 15.42 ± 0.13 77.23 ± 0.84 411 ± 19 740 ± 130 flounder SE AK Sep 5 1.80 ± 0.04 9.62 ± 0.56 15.76 ± 0.35 72.68 ± 0.86 590 ± 49 2365 ± 660 Pacific cod Aleutians Apr-Jul 28 1.14 ± 0.02 2.34 ± 0.23 16.32 ± 0.16 79.30 ± 0.29 608 ± 26 2982 ± 408

SE AK Jul 11 1.01 ± 0.02 1.31 ± 0.23 15.72 ± 0.11 80.38 ± 0.25 304 ± 27 425 ± 173 Schaufler etal.—Prey Variation inAlaska Pollock Aleutians May-Jun 70 1.14 ± 0.01 2.90 ± 0.13 15.38 ± 0.11 79.46 ± 0.20 373 ± 15 482 ± 39 SE AK Quarterly 280 1.15 ± 0.01 3.19 ± 0.10 15.05 ± 0.07 79.16 ± 0.13 320 ± 10 515 ± 31 Rockfish Aleutians May 16 1.87 ± 0.06 9.39 ± 0.64 17.46 ± 0.17 70.54 ± 0.64 349 ± 17 714 ± 99 SE AK Jul, Dec 23 1.56 ± 0.04 6.21 ± 0.41 17.21 ± 0.16 73.48 ± 0.44 283 ± 12 449 ± 62 Sandfish Aleutians May 6 1.10 ± 0.04 3.02 ± 0.32 14.41 ± 0.27 80.09 ± 0.27 157 ± 22 57 ± 24 SE AK Dec 3 1.44 ± 0.10 5.41 ± 0.95 16.42 ± 0.33 75.82 ± 0.93 170 ± 39 81 ± 55 Squid Aleutians May 15 1.02 ± 0.04 3.06 ± 0.33 12.91 ± 0.14 81.75 ± 0.39 219 ± 8 347 ± 30 SE AK Sep 19 1.45 ± 0.04 6.61 ± 0.42 14.60 ± 0.27 77.21 ± 0.45 138 ± 19 238 ± 97 Atka Aleutians May-Jul 100 1.93 ± 0.05 10.92 ± 0.55 15.83 ± 0.07 71.93 ± 0.43 370 ± 5 662 ± 26 mackerel Table 1. (continued.) Sea LionsoftheWorld Energy Collection density Speciesa Region monthb N (kCal/g) % Lipidc % Proteinc % Moisture Length (mm)d Weight (g) Rock sole Aleutians May 18 1.04 ± 0.03 1.74 ± 0.25 15.57 ± 0.25 79.13 ± 0.39 300 ± 10 320 ± 38 Skates Aleutians May 14 1.07 ± 0.04 2.49 ± 0.30 14.75 ± 0.31 82.79 ± 0.47 391 ± 27 399 ± 63 Yellow Irish Aleutians May 15 1.16 ± 0.03 3.73 ± 0.33 14.34 ± 0.11 78.22 ± 0.42 379 ± 20 690 ± 69 lord Capelin SE AK Quarterly 84 1.26 ± 0.03 4.67 ± 0.28 14.40 ± 0.11 78.97 ± 0.36 107 ± 2 8.2 ± 0.4 Eulachon SE AK Quarterly 154 2.22 ± 0.03 15.85 ± 0.31 12.70 ± 0.10 70.24 ± 0.36 166 ± 2 33 ± 1 Hake SE AK Quarterly 87 1.41 ± 0.02 6.64 ± 0.24 13.87 ± 0.07 77.69 ± 0.23 491 ± 7 878 ± 44 Herring SE AK Quarterly 158 2.04 ± 0.05 11.93 ± 0.47 16.19 ± 0.16 69.84 ± 0.48 195 ± 4 90 ± 5 Lampfish SE AK Apr, Dec 44 2.64 ± 0.02 18.99 ± 0.16 15.04 ± 0.11 65.15 ± 0.19 101 ± 1 12.0 ± 0.4 Lump- SE AK Dec 11 0.69 ± 0.06 3.61 ± 0.47 6.24 ± 0.31 88.97 ± 0.50 270 ± 15 1530 ± 140 sucker Sand lance SE AK Jul, Dec 16 1.61 ± 0.06 6.51 ± 0.55 17.53 ± 0.27 73.66 ± 0.68 114 ± 5 8.6 ± 2.2 Smooth- SE AK Dec 14 1.54 ± 0.10 10.20 ± 0.94 10.28 ± 0.26 77.63 ± 0.89 120 ± 5 13.5 ± 1.9 tongue aSpecies examined are arrowtooth flounder Atheresthes stomias; Atka mackerel Pleurogrammus monopterygius; capelin Mallotus villosus; eulachon Thaleichthys pacificus; hake Merluccius productus; herring Clupea pallasii; Pacific codGadus macrocephalus; walleye pollock Theragra chalcogramma; rockfish includeSebastes alutus, Sebastes ciliatus, Sebastes maliger, Sebastes polyspinis, and Sebastes variegatus; rock sole Lepidopsetta bilineata; sandfishTrichodon trichodon; skates include Bathyraja interrupta, Bathyraja minispinosa, Bathyraja parmifera, and Bathyraja taranetzi; squid include Berryteuthis magister and Loligo opalescens; yellow Irish lord Hemilepidotus jordani. bCollection months are indicated, mostly from 2002. “Quarterly” is a minimum of quarterly coverage throughout the year, typically March, May/June, September, and December. cPercentages are given on a wet weight basis. dLengths for most species were measured as distances from snout tip to the tail fork (fork lengths), or tail center for rounded tails. Mantle lengths (excluding tentacles) were used for squid. Overall body lengths, including tail, were used for skates. 123 124 Schaufler et al.—Prey Variation in Alaska

2.5 Aleutians Southeast AK

1.5

1 Ave Energy Density (kCal/g) Ave

0.5

34 5 28 11 70 280 16 23 6 3 15 19 0 ATF P.Cod Pollock Rockfish Sandfish Squid

Prey Species

Figure 2. Comparisons of the average energy densities for species collected in the Aleutian Islands and southeastern Alaska. Error bars indicate standard error, and sample sizes are shown at the base of each bar. ATF = arrowtooth flounder; P. cod = Pacific cod.

Comparison of the energy densities for species found in both regions revealed a number of significant differences (Fig. 2). Arrowtooth flounder had a significantly higher average energy density in southeastern Alaska than in the Aleutians (t37 = –4.87, p < 0.001), as did sandfish t( 7 = –4.09, p = 0.005) and squid (t32 = –5.72, p < 0.001), mainly due to differences in lipid content. Pacific cod had a significantly higher energy density in the

Aleutians than in southeastern Alaska (t31 = –4.32, p < 0.001), as did the

Aleutian rockfish t( 37 = –4.44, p < 0.001), also due to higher lipid content. Interestingly, Pacific cod showed a significant linear correlationr ( 2 = 0.498) between energy density and fish weight that was not observed for the other species with significant numbers of observations N( > 20) (Fig. 3). Walleye pollock from both regions had similar estimated energy densities (1.14 ± 0.01 kCal per g and 1.15 ± 0.01 kCal per g for the Aleu- tians and southeastern Alaska, respectively), even though there was a significant difference in fish lengths collected from the two geographical areas (t132 = 2.95, p = 0.004). Overall, age classes sampled included young- of-the-year through adult, with lengths from 7 cm to 67 cm and weights ranging from 2 g to over 2.5 kg. Length was observed to be only weakly correlated with energy density (r2 = 0.113). The lack of substantial dependence of walleye pollock lipid content (and energy density) on fish size Sea Lions of the World 125

1.4

1.2

1.0 Energy Density (kCal/g)

0.8

y = 4e-5x + 1.015 R2 = 0.498 0.6 0 2000 4000 6000 8000

Weight (g)

Figure 3. Size correlation of energy density for Pacific cod.

further suggests that this relationship may be species-dependent, as has been previously observed (Anthony et al. 2000, Vollenweider 2004).

Discussion Knowledge of the species composition of two regional diets as well as the nutritional differences between prey items are critical steps in evaluat- ing the prey quality aspect of the junk food hypothesis. Opportunistic collections of prey species in the western and eastern regions provided a sampling of potential Steller sea lion prey for nutritional analysis from species primarily found in only one region and species common to both areas. Proximate analysis revealed that on an overall basis, the Aleutian prey items collected had a lower average lipid content, and consequently a lower estimated average energy density, than did the potential prey items collected from southeastern Alaska. Therefore, if trawl-based sampling can serve as a proxy for opportunistic foraging by sea lions, and the prey collections used in this study serve as an unbiased representa- tive of the available prey, the difference in overall energy density would suggest that sea lions in the western region encounter, on average, lower energy density prey than those in the eastern region. However, trawl- based sampling tends to be biased toward pelagic species and researcher prey sampling is unavoidably influenced by factors such as daylight, weather and ocean conditions, shore proximity, and other logistical issues (Stoner 2004). Furthermore, while systematic prey collections in 126 Schaufler et al.—Prey Variation in Alaska southeast Alaska are possible on a year-round basis, they are extremely difficult in the harsh winter conditions of the Aleutian Islands, limiting the available data. The main objective of this study was to determine the proximate compositions of common Steller sea lion prey items in the Aleutians and southeastern Alaska, and provide comparisons for species found in both regions. However, sampling conditions and whether a given species was located in both areas limited the available comparisons. Some species were not found in one location during a given season, and average sizes of the fish collected sometimes varied by location. Comparing the available data for fish collected from both areas indicated differences overall, but many of these may be attributable to factors such as size and season. For the species collected from the Aleutians and southeastern Alaska, we wished to address whether they had the same average energy densities in both areas. Energy density estimates were calculated based on the proximate composition, particularly lipid and fat content, of the sampled fish. Comparison of published values obtained using bomb calorimetry with our calculated energy densities showed excellent consistency for species with available data such as walleye pollock (1.15 ± 0.01 kCal per g [this study] vs. 1.11 ± 0.03 kCal per g [Perez 1994]), and Pacific herring (2.04 ± 0.05 kCal per g [this study] vs. 2.05 ± 0.18 kCal per g [Perez 1994]). Proximate values observed for species with published data were also consistent with observed ranges (Payne et al. 1999, Iverson et al. 2002). Of particular interest was the comparison of average energy densities for walleye pollock from both regions, given that some of the sharp- est declines in Steller sea lion populations have occurred in areas where walleye pollock dominates the diet, and the fact that pollock is a major food component to both Steller sea lion stocks (Rosen and Trites 2000, Winship and Trites 2003). The energy densities observed for Aleutian and southeastern Alaska walleye pollock were equivalent (p > 0.5), suggesting that the quality of pollock in both regions is similar and is not likely to be a factor in the decline. Comparisons of average energy densities for other species collected from both regions revealed differences that could be attributed to factors other than geographical region. For instance, arrowtooth flounder collected in southeastern Alaska were significantly larger than those sampled from the Aleutians (t37 = –3.37, p = 0.002), while Aleutian Pacific cod were larger than those from southeastern Alaska (t28 = 8.08, p < 0.001), and in both cases prey size accounted for the observed differences in energy densities (Table 1). For arrowtooth flounder, lipid content, which often dominates the energy density value, varied between locations (t37 = –4.40, p < 0.001) primarily due to size, though protein content did not

(t37 = –0.91, p = 0.40). In general, protein content did not appear to vary significantly with size for any of the species we examined. Sea Lions of the World 127

For sandfish and squid collected from both regions, the observed differences in energy densities were in part attributable to length differences (p = 0.04 and p = 0.01, respectively), but length did not completely account for the higher energy densities of the samples collected in southeastern Alaska. The season in which these species were encountered in the two regions differed, and is likely to be another source of variation; they were collected in May in the Aleutians, and in the fall in southeastern Alaska. Differences in the spawning state for these species may have caused them to have significantly different lipid content. Observed differences in the energy densities for rockfish from the Aleutians and southeastern Alaska are likely due to differences in the particular species collected in the two regions. Though all of the rockfish studied were in the same genus, and were therefore grouped together for comparison, the mean lipid content varied significantly per species. For instance, the sampled Aleutian rockfish were Pacific ocean perch and northern rockfish, which had average lipid values of 8.18 ± 1.04% and 10.60 ± 0.51%, respectively. From southeastern Alaska, collections included dusky rockfish, which had an average lipid value of 6.64 ± 0.37%. Furthermore, no single species of rockfish was collected in both regions, and the months in which they were encountered differed, confounding the comparison. The effect of fish size on energy density makes geographical comparisons of prey quality difficult, particularly when the size of similarly aged fish differs significantly between two regions due to differences in environmental conditions. This effect appears to be specific to certain species. For some pelagic species, such as herring and sandfish, there is a known relationship between size and lipid content (and hence energy density) due to changes in energy allocation with age (Anthony et al. 2000). However, in walleye pollock there is no apparent relationship between energy density and specimen length (Vollenweider 2004). Differences in the mean size of fish captured in the Aleutians versus southeastern Alaska may be due to sampling-related issues, or may represent real differences in the sizes of prey encountered by Steller sea lions. Little is known regarding the size structures of various prey species in the Aleutians versus southeastern Alaska, but there are several physiological advantages to being a larger size in a harsh environment, and prey size has been observed to vary by location and environmental conditions (Shuter and Post 1990, Zeppelin et al. 2004). Another aspect of diet quality is the diversity of available quality prey items. Population declines previously have been correlated to diets of low diversity (Merrick et al. 1997), and it has been suggested that aquatic carnivores have increased digestibility when consuming mixed diets (Trumble and Castellini 2005). In addition, prey diversity likely increases foraging efficiency and reduces the dependence of predator consumption on the seasonal energetic cycles of a particular prey spe- 128 Schaufler et al.—Prey Variation in Alaska cies. In comparing Aleutian versus southeastern Alaska prey items, four quality prey species, with energy densities of 1.8 kCal per g or higher, were identified in southeastern Alaska, whereas only two were found in the Aleutian prey subset. Admittedly, we have not yet sampled all of the known prey items from either region, but a significant difference in the number of quality prey items available in the two regions could have a noticeable impact on sea lion foraging success and nutritional state in those areas. Opportunistic sampling of potential prey items is problematic due to issues such as fish availability, choice of fishing gear, fishing conditions including prior fishing pressure, and particular geographical region sampled. Furthermore, comparison of prey quality between areas is difficult due to regional differences related to environmental conditions such as species assemblages, population structures, and differences in growth rate, as well as the fact that a given species may spawn at different times in the two regions. Comparisons are hindered by these multiple factors and limit interpretation of the data. Even with these limitations, however, proximate composition data and energy density estimates for species from remote regions such as the Aleutian Islands provide a snapshot of prey quality for a given species at a specific time, which is a critical tool to assess the nutritional state of predators in the area such as Steller sea lion. As more data are collected from these regions, more specific comparisons can be made with fewer confounding factors.

Acknowledgments The authors would like to thank Robert Bradshaw and Darcie Neff for their focused efforts and careful attention to detail in sample analysis and data archiving, as well as Ron Heintz for his helpful comments on the data and manuscript. The authors are especially grateful to the re- viewers for their comments and suggestions. Special acknowledgments also go to the vessel crews who assisted in sample collection.

References Alverson, D.L. 1992. A review of commercial fisheries and the Steller sea lion (Eumetopias jubatus): The conflict arena. Rev. Aquat. Sci. 6:203-256. AOAC. 1995. Official methods of analysis, 16th edn. Association of Official- Ana lytical Chemists, Arlington, Virginia. Anthony, J.A., D.D. Roby, and K.R. Turco. 2000. Lipid content and energy density of forage fishes from the northern Gulf of Alaska. J. Exp. Mar. Biol. 248:53-78. Brody, S. 1945. Bioenergetics and growth. Hafner Publishing Co., New York. 1023 pp. Christie, W.W. 2003. Lipid analysis, 3rd edn. Oily Press, England, pp. 101, 208. Sea Lions of the World 129

Iverson, S.J., K.J. Frost, and S.L.C. Lang. 2002. Fat content and fatty acid composition of forage fish and invertebrates in Prince William Sound, Alaska. Mar. Ecol. Prog. Ser. 241:161-181. Logerwell, E.A., and L.E. Schaufler. 2005. New data on proximate composition and energy density of Steller sea lion (Eumetopias jubatus) prey fills seasonal and geographic gaps in existing information. Aquat. Mamm. 31:62-82. Merrick, R.L., M.K. Chumbley, and G.V. Byrd. 1997. Diet diversity of Steller sea lions (Eumetopias jubatus) and their population decline in Alaska: A potential relationship. Can. J. Fish. Aquat. Sci. 54:1342-1348. Payne, S.A., A.B. Johnson, and R.S. Otto. 1999. Proximate composition of some northeastern Pacific forage fish species. Fish. Oceanogr. 8:159-177. Perez, M.A. 1994. Calorimetry measurements of energy value of some Alaskan fishes and squids. NOAA Tech. Memo NMFS-AFSC-32. 32 pp. Rosen, D.A., and A.W. Trites. 2000. Pollock and the decline of Steller sea lions: Testing the junk-food hypothesis. Can. J. Zool. 78:1243-50. Shuter, B.J., and J.R. Post. 1990. Climate, population viability, and the zoogeog- raphy of temperate fishes. Trans. Am. Fish. Soc. 119:314-336. Sinclair, E.H., and T.K. Zeppelin. 2002. Seasonal and spatial differences in diet in the western stock of Steller sea lions (Eumetopias jubatus). J. Mamm. 83:973-990. Stoner, A.W. 2004. Effects of environmental variables on fish feeding ecology: Implications for the performance of baited fishing gear and stock assessment. J. Fish Biol. 65:1445-1471. Tollit, D.J., S.G. Heaslip, and A.W. Trites. 2004. Sizes of walleye pollock (Theragra chalcogramma) consumed by the eastern stock of Steller sea lions (Eumetopias jubatus) in Southeast Alaska from 1994 to 1999. Fish. Bull. U.S. 102:522-532. Trites, A.W., and C.P. Donnelly. 2003. The decline of Steller sea lions Eumetopias jubatus in Alaska: A review of the nutritional stress hypothesis. Mamm. Rev. 33:3-28. Trumble, S.J., and M.A. Castellini. 2005. Diet mixing in an aquatic carnivore, the harbour seal. Can. J. Zool. 83:851-859. Van Pelt, T.I., J.F. Piatt, B.K. Lance, and D.D. Roby. 1997. Proximate composition and energy density of some north Pacific forage fishes. Comp. Biochem. Phys. 118A:1393-1398. Vollenweider, J.J. 2004. Variability in Steller sea lion (Eumetopias jubatus) prey quality in southeastern Alaska. Master’s thesis, University of Alaska Fair- banks, Juneau. Winship, A.J., and A.W. Trites. 2003. Prey consumption of Steller sea lions (Eume- topias jubatus) off Alaska: How much prey do they require? Fish. Bull. U.S. 101:147-167. Zeppelin, T.K., D.J. Tollit, K.A. Call, T.J. Orchard, and C.J. Gudmundson. 2004. Sizes of walleye pollock and Atka mackerel (Pleurogrammus monopterygius) consumed by the western stock of Steller sea lions (Eumetopias jubatus) in Alaska from 1998 to 2000. Fish. Bull. U.S. 102:509-521.

Sea Lions of the World 131 Alaska Sea Grant College Program • AK-SG-06-01, 2006

Relationship between Steller Sea Lion Diets and Fish Distributions in the Eastern North Pacific

Emma L. Bredesen, Andrea P. Coombs, and Andrew W. Trites University of British Columbia, Marine Mammal Research Unit, Fisheries Centre, Aquatic Ecosystems Research Laboratory, Vancouver, British Columbia, Canada

Abstract Distributions of fish species were compared with diet information for Steller sea lions (Eumetopias jubatus) to assess the level of correspon- dence between potential prey availability and sea lion feeding habits. Fish distributions were compiled as part of the Sea Around Us Project at the UBC Fisheries Centre, and were based on published distributions and habitat preferences (e.g., latitude, depth). Sea lion scat samples were collected during the 1990s from seven geographic regions from Oregon to the western and central Aleutian Islands. The frequencies of occurrence of four prevalent species (walleye pollock, Theragra chalcogramma; Pacific herring, Clupea pallasii; Pacific cod, Gadus macrocephalus; and North Pacific hake,Merluccius productus) in the Steller sea lion diet were compared to their distributions in the North Pacific Ocean. The data suggest that Steller sea lion diets broadly reflect the distributions of these major prey species. However, some of the fish species that were region- ally predicted to be present in high abundance were not proportionally reflected in the Steller sea lion diet, suggesting that other factors in addition to fish abundance influence their diets.

Introduction The Steller sea lion population declined by more than 80% in western Alaska between the mid-1970s and the early 1990s (Trites and Larkin 1996, Loughlin 1998, NAS 2003) while the smaller eastern population increased. Accordingly, the western and eastern populations of Steller sea lions were listed as “endangered” and “threatened,” respectively, un- 132 Bredesen et al.—Steller Sea Lion Diets and Fish Distributions der the U.S. Endangered Species Act. In an attempt to better understand their role in the ecosystem and the differences between the decreasing western population and the increasing eastern population, considerable research has focused on determining the diet of Steller sea lions (Merrick et al. 1997, Sinclair and Zeppelin 2002). Comprehensive changes have occurred to the biomass and composition of the marine community off the Alaska coast since the oceanic regime shift of 1976. Increases were noted following the regime shift in the abundances of flatfish, gadids, and salmonids (Hare and Francis 1995, Hollowed et al. 2001, Benson and Trites 2002, Wilderbuer et al. 2002). A small mesh survey around Kodiak Island (1953-1997) also noted increases in groundfish such as cod and pollock, as well as a decline in the abundance of forage species such as capelin and shrimp (Anderson and Piatt 1999). Such changes in the abundances and distributions of key prey species composition may be related to the decline of Steller sea lions in western Alaska. One of the leading hypotheses for the decline in western Alaska is nutritional stress caused by a shift in ocean climate that favored the abundance of less nutritious fishes over those that had higher fat content (Alverson 1992, Rosen and Trites 2000, Trites and Donnelly 2003, Trites et al. 2006). Like other pinnipeds, Steller sea lions have often been classi- fied as generalist feeders. However, it is not clear whether Steller sea lions merely eat what is available to them, or whether other factors influence which prey they consume. We therefore sought to assess the level of cor- respondence between prey distribution and sea lion diets.

Methods Predicted fish distribution maps were obtained from the Sea Around Us Project (Fisheries Centre, University of British Columbia, www.seaaroundus.org) and were based on published distributions and habitat preferences (e.g., latitude, depth) (Watson et al. 2004). These maps represent the expected percent of world distribution of individual fish species and are an indication of relative abundance of each species across the North Pacific Ocean. Scat samples were collected during the 1990s (Riemer and Brown 1997, Sinclair and Zeppelin 2002, Trites et al. unpubl. data) from seven geographic regions from Oregon to the Central and Western Aleutian Islands (Oregon [OR], British Columbia [BC], Southeast Alaska [SEA], Gulf of Alaska 1 [GOA1], Gulf of Alaska 2 [GOA2], Eastern Aleutian Islands [EAI], and Western Central Aleutian Islands [WCAI]). Frequently occurring species (walleye pollock, Theragra chalcogramma; Pacific herring, Clu- pea pallasii; Pacific cod, Gadus macrocephalus; and North Pacific hake, Merluccius productus) in the Steller sea lion diet were compared to their predicted distributions. The importance of each prey species was deter- Sea Lions of the World 133 mined by the percent frequency of occurrence (%FO) in scat samples from each region (Croxall 1993). Proportionally sized circles, representing the %FO, were plotted for comparison with fish distributions in respective regions. Percentages of both diet and distribution data were arcsine transformed to satisfy the assumptions for statistical analysis.

Results and discussion The data suggest that there is a relationship between fish distributions and Steller sea lion diets (Figs. 1 and 2). For example, in the northern part of the sea lion’s range, walleye pollock has a high relative abundance and is an important part of the sea lion diet (Fig. 1A). Although not statistically significant r( = 0.63, P > 0.05), the correlation is positive and consistent with the relationship found for the other prey species. In the southern part of the sea lion range (e.g., Oregon), the predicted relative abundance of walleye pollock is lower and North Pacific hake is predominant both in relative abundance and in the sea lion diet (r = 0.74, P = 0.05) (Figs. 1A, 1D, and 2). A similar pattern can be seen with Pacific herring r( = 0.80, P = 0.03), arrowtooth flounder r( = 0.83, P = 0.02), and, to a lesser extent, Pacific cod r( = 0.28, P > 0.05), where the frequency of occurrence of the prey species in the diet of Steller sea lions is higher in regions that also have high predicted relative fish abundance (Figs. 1B, 1C, and 2). While the relationship between relative fish abundances and Steller sea lion diets seems strong, it is not as straightforward as Figs. 1 and 2 might suggest. For example, in the Gulf of Alaska the biomass of arrowtooth flounder (Turnock et al. 2001) is estimated to be approximately six times the biomass of walleye pollock (Dorn et al. 2001). However, pollock is two to eight times more prominent in the Steller sea lion diet (%FO) than arrowtooth flounder (Sinclair and Zeppelin 2002) (Tables 1 and 2). Atka mackerel is also 35 times more prominent than rockfish in the diet of Steller sea lions in the Aleutian Islands region (Sinclair and Zeppelin 2002) despite the fact that both have a similar biomass estimate (Lowe et al. 2002, Spencer and Ianelli 2002) (Tables 1 and 2). Although not statistically significant, Atka mackerel r( = –0.18, P > 0.05) and Pacific ocean perch/Rockfish r( = –0.29, P > 0.05) show a weak negative correlation between the percent of world distribution and percent frequency of occurrence in Steller sea lion diet (Fig. 3). Many factors likely influence the prey that sea lions choose to eat. These include the presence of spines, and the vertical and horizontal distribution of prey in the water column. Such factors may account for some of the apparent discrepancies between diets and relative prey abundances. For example, arrowtooth flounder may make up a large portion of the relative biomass, but may be harder for sea lions to locate and capture because they tend to be solitary and not school in easily 134 Bredesen et al.—Steller Sea Lion Diets and Fish Distributions

A) B)

GOA1 GOA1 SEA SEA GOA2 EAI GOA2 WCAI EAI BC WCAI BC

OR OR

Walleye pollock Pacific herring

C) D)

GOA1 SEA SEA GOA1 GOA2 GOA2 EAI EAI WCAI WCAI BC BC

OR OR

Pacific cod North Pacific hake

LEGEND % of world distribution

Country FAO

> 0.0005 < 0.0002 < 0.0005 < 0.0001 < 0.0004 < 0.00005 < 0.0003 Not found

% Frequency of Occurence in Steller sea lion scat

Figure 1. Predicted distribution and relative abundance of four species of fish in the eastern North Pacific (Watson et al. 2004). Grayscale shad- ing indicates relative abundance of each species. Proportionally sized circles, plotted in respective regions, represent the percent frequencies of occurrence (%FO) of fish species in scat samples from Steller sea lions during the 1990s (from Riemer and Brown 1997, Sinclair and Zeppelin 2002, and Trites et al. unpubl. data). Regions, from left to right, are Western Central Aleutian Islands (WCAI), Eastern Aleutian Islands (EAI), Gulf of Alaska 2 (GOA2), Gulf of Alaska 1 (GOA1), Southeast Alaska (SEA), British Columbia (BC), and Oregon (OR). Sea Lions of the World 135

60 Pollock 50 Herring Cod 40

%FO Hake 30 Arrowtooth

0 0 5 10 15 20 25 30 35

Percent of world distribution

Figure 2. Relationship between percent frequency of occurrence in the diet of Steller sea lions in the 1990s and percent of world distribution for five prominent prey species in the North Pacific Ocean. Statistically significant positive correlations occur between world distribution and frequency of occurrence for herring, hake, and arrowtooth flounder. Correlations for pollock and cod were positive but not significant.

80 Atka 70 POP/Rockfish 60

40 %FO

0 0 5 10 15 20 25 30 35

Percent of world distribution

Figure 3. Percent frequency of occurrence versus percent of world distribution for two Steller sea lion prey species (Atka mackerel and Pacific ocean perch). Neither of the relationships were statistically significant. 136 Bredesen et al.—Steller Sea Lion Diets and Fish Distributions

Table 1. Percent frequency of occurrence (%FO) for prominent prey species in Steller sea lion scat.

%FOa

N. Arrow- Walleye Pacific Pacific Pacific tooth Atka Region Nb pollock herring cod hake flounder mackerel Rockfish

WCAI 1,370 8.85 0.1c 7.62 0.1c 0.29 89.61 2.51

EAI 889 57.25 11.59 14.75 0.1c 3.93 25.32 4.02

GOA2 929 83.56 5.93 27.40 0.1c 8.49 3.12 3.13

GOA1 574 59.43 16.21 20.03 0.50 27.18 1.22 2.07

SEA 1,438 73.02 34.56 2.02 2.23 16.48 0.14 17.59

BC 1,077 13.28 47.45 5.29 7.80 28.51 0.09 37.60

OR 256 0 19.92 0.75 83.60 7.03 4.67 9.77 aPercentages were arcsine transformed prior to statistical analysis. bNumber of scat samples containing identifiable prey used to calculate annual average of percent frequency of occurrence (%FO). cSpecies present but <1 (Sinclair and Zeppelin 2002).

Table 2. Percent of world distribution for prominent prey species in Steller sea lion scat.

% of World distributiona N. Arrow- Walleye Pacific Pacific Pacific tooth Atka POP/ Region pollock herring cod hake flounder mackerel rockfish WCAI 0.09 0 1.81 0 1.44 0.06 0.33 EAI 12.50 2.08 6.12 0 3.40 11.09 32.45 GOA2 6.91 4.10 4.52 0 1.77 6.14 26.03 GOA1 2.99 7.22 7.08 0 4.60 2.53 11.34 SEA 0.71 5.00 4.36 0 3.38 0.54 1.78 BC 1.68 7.34 9.25 26.69 5.37 1.66 6.61 OR 0.22 2.77 2.93 22.30 1.73 0.27 1.30 aProportions of world distribution were summed by region for each fish species. Percentages were arcsine transformed prior to statistical analysis. Sea Lions of the World 137 exploitable densities for much of the year. Similarly, not all fish may be equally available to sea lions if they occur at depths or in areas that are difficult for sea lions to access. Thus, to fully understand the association between Steller sea lion feeding habits and their prey, consideration needs to be given to factors other than the simple distribution of prey species. The available data suggest that the diets of Steller sea lions broadly reflect the distributions of their major prey species. However, discrepancies suggest that other factors such as nutritional value, relative foraging costs, prey preference, etc., should also be considered to better understand the feeding habits of Steller sea lions. Nonetheless, given the general relationship between fish distributions and Steller sea lion diets, factors that affect fish assemblages (such as climatic change) may also have implications for sea lion populations. Additional analysis is therefore required to achieve a better understanding of Steller sea lion diet and how it is related to the distribution of their prey throughout their range. An analysis at finer spatial and temporal scales, incorporating seasonal or monthly sea lion diet data and fish abundance data would further help to elucidate factors that affect Steller sea lion feeding habits. Consider- ation should also be given to the different depth ranges that adults and juvenile fish species inhabit, and how it relates to the ability of sea lions to successfully forage.

Acknowledgments We are grateful to the Sea Around Us Project, particularly Dr. Reg Watson and Mr. Adrian Kitchingman; NOAA; and the North Pacific Marine Science Foundation for providing funding to the North Pacific Universities Marine Mammal Research Consortium.

References Alverson, D.L. 1992. A review of commercial fisheries and the Steller sea lion Eu( - metopias jubatus): The conflict arena. Rev. Aquat. Sci. 6:203-256. Anderson, P.J., and J.F. Piatt. 1999. Community reorganization in the Gulf of Alaska following ocean climate regime shift. Mar. Ecol. Prog. Ser. 189:117-123. Benson, A.J., and A.W. Trites. 2002. Ecological effects of regime shifts in the Bering Sea and eastern North Pacific Ocean. Fish Fish. 3:95-113. Croxall, J.P. 1993. Diet. In: R.M. Laws (ed.), Antarctic seals: Research methods and techniques. Cambridge University Press, Cambridge, pp. 268-290. Dorn, M.W., A.B. Hollowed, E. Brown, B.A. Megrey, C. Wilson, and J. Blackburn. 2001. Assessment of the walleye pollock stock in the Gulf of Alaska. In: Stock assessment and fishery evaluation report for the groundfish resources of the Gulf of Alaska. North Pacific Fishery Management Council, www.afsc.noaa. gov/refm/docs/2001/GOA pollock.pdf, pp. 1-88. 138 Bredesen et al.—Steller Sea Lion Diets and Fish Distributions

Hare, S.R., and R.C. Francis. 1995. Climate change and salmon production in the northeast Pacific Ocean. Can. J. Fish. Aquat. Sci. 121:357-372. Hollowed, A.B., S.R. Hare, and W.S. Wooster. 2001. Pacific-basin climate variability and patterns of Northeast Pacific marine fish production. Prog. Oceanogr. 49:257-282. Loughlin, T.R. 1998. The Steller sea lion: A declining species. Biosph. Conserv. 1:91-98. Lowe, S., J. Ianelli, H. Zenger, and R. Reuter. 2002. Stock assessment of Aleutian Islands Atka mackerel. In: Stock assessment and fishery evaluation report for the groundfish resources of the Bering Sea/Aleutian Islands regions. North Pacific Fishery Management Council, www.afsc.noaa.gov/refm/docs/2002/ BSatka.pdf, pp. 609-668. Merrick, R.L., M.K. Chumbley, and G.V. Byrd. 1997. Diet diversity of Steller sea lions (Eumetopias jubatus) and their population decline in Alaska: A potential relationship. Can. J. Fish. Aquat. Sci. 54:1342-1348. NAS (ed.). 2003. Decline of the Steller sea lion in Alaskan waters: Untangling food webs and fishing nets. National Research Council, National Academy of Sci- ence, Washington, D.C.: National Academies Press. 204 pp. Riemer, S.D., and R.F. Brown. 1997. Prey of pinnipeds at selected sites in Oregon identified by scat (fecal) analysis, 1983-1996. Oregon Department of Fish and Wildlife report no. 97-6-02. Rosen, D.A.S., and A.W. Trites. 2000. Pollock and the decline of Steller sea lions: Testing the junk-food hypothesis. Can. J. Zool. 78:1243-1250. Sinclair, E.H., and T.K. Zeppelin. 2002. Seasonal and spatial differences in diet in the western stock of Steller sea lions (Eumetopias jubatus). J. Mammal. 83:973-990. Spencer, P.D., and J.N. Ianelli. 2002. Pacific ocean perch. In: Stock assessment and fishery evaluation report for the groundfish resources of the Bering Sea/Aleu- tian Islands regions. North Pacific Fishery Management Council, www.afsc. noaa.gov/refm/docs/ 2002/BSpop.pdf, pp. 515-558. Trites, A.W., and C.P. Donnelly. 2003. The decline of Steller sea lions Eumetopias jubatus in Alaska: A review of the nutritional stress hypothesis. Mamm. Rev. 33:3-28. Trites, A.W., and P.A. Larkin. 1996. Changes in the abundance of Steller sea lions (Eumetopias jubatus) in Alaska from 1956 to 1992: How many were there? Aquat. Mamm. 22:153-166. Trites, A.W., A.J. Miller, H.D.G. Maschner, M.A. Alexander, S.J. Bograd, J.A. Calder, A. Capotondi, K.O. Coyle, E. Di Lorenzo, B.P. Finney, E.J. Gregr, C.E. Grosch, S.R. Hare, G.L. Hunt, Jr., J. Jahncke, N.B. Kachel, H.-J. Kim, C. Ladd, N.J. Mantua, C. Marzban, W. Maslowski, R. Mendelssohn, D.J. Neilson, S.R. Okkonen, J.E. Overland, K.L. Reedy-Maschner, T.C. Royer, F.B. Schwing, J.X.L. Wang, and A.J. Winship. 2006. Bottom-up forcing and the decline of Steller sea lions in Alaska: Assessing the ocean climate hypothesis. Fish. Oceanogr. Sea Lions of the World 139

Turnock, B.J., T.K. Wilderbuer, and E.S. Brown. 2001. Arrowtooth flounder. In: Stock assessment and fishery evaluation report for the groundfish resources of the Gulf of Alaska. North Pacific Fishery Management Council, www.afsc.noaa. gov/refm/docs/ 2001/goaatf.pdf, pp. 4-1 to 4-29. Watson, R., A. Kitchingman, A. Gelchu, and D. Pauly. 2004. Mapping global fisheries: Sharpening our focus. Fish Fish. 5:168-177. (www.seaaroundus.org/distribution/ search.aspx). Wilderbuer, T.K., A.B. Hollowed, W.J. Ingraham Jr., P.D. Spencer, M.E. Connors, N.A. Bond, and G.E. Walters. 2002. Flatfish recruitment response to decadal climatic variability and ocean conditions in the eastern Bering Sea. Prog. Oceanogr. 55:235-247.