Trophic Relationships Among Seabirds in Central California: Combined Stable Isotope and Conventional Dietary Approach ’

The Condor 99327-336 0 The Cooper Ornithological Society 1997

TROPHIC RELATIONSHIPS AMONG SEABIRDS IN CENTRAL CALIFORNIA: COMBINED STABLE ISOTOPE AND CONVENTIONAL DIETARY APPROACH ’

WILLIAM J. SYDEMAN Point ReyesBird Observatory,4990 ShorelineHighway, StinsonBeach, CA 94970

KEITH A. HOBSON Canadian Wildlife Service, 115 Perimeter Road, Saskatoon,Saskatchewan, Canada, S7N OX4

PETER PYLE AND ELIZABETHB. MCLAREN Point ReyesBird Observatory,4990 ShorelineHighway, StinsonBeach, CA 94970

Abstract. We used stable isotope analysis (SIA) and conventionaltechniques of diet assessmentto determinemarine trophic relationships in the Gulf of the Farallones,Califor- nia, with an emphasison marine birds. Stable-carbon(SW) and nitrogen(WN) isotopes were obtainedfrom 98 tissuesamples of 16 species representingprimary and secondary consumersin 1993-1994. The valies of iYT ranged from -2O.i%; in whole euphausiids (krill) to - 15.0%0in muscle of northern sea lions. Values of 815Nshowed steu-wisetronhic enrichment and ranged from 11.2%0in euphausiidsto 19.8%0 in sea lion;. SIA of-egg albumenfrom birdsindicated reliance on zooplanktonby Cassin’sAuklet, CommonMurre, and Western Gull, and on fish by Brandt’s and Pelagic Cormorants, Rhinoceros Auklets, and Pigeon Guillemots during egg formation (April-May). However, analysis of prey broughtto chicks during summerindicated the prevalenceof fish in the diet of most seabirds, except Cassin’s Auklet which fed primarily on krill. Results suggesta shift in trophic level and diet between spring and summer from krill to fish for Common Murres. ST analysis confirmed that Brandt’s Cormorantsand northern sea lions feed in neritic habitats,whereas Cassin’s and RhinocerosAuklets foraged in epipelagic offshore waters. Our approachdem- onstratesthe utility of combining both SIA and conventional dietary assessmentsto under- stand trophic relationshipsin dynamic marine ecosystems. Key words: central California, diet, food-web dynamics,marine birds and mammals, stable isotopeanalysis, trophic relationships.

INTRODUCTION meer 1981, Briggs et al. 1988, Ainley et al. Understanding the structure and functioning of 1990). marine ecosystemsrequires information on the However, evaluation of the structure and dy- trophic relationships of key species (Paine namics of marine food webs is a complex un- 1988). In coastal food-webs dependent upon up- dertaking because of logistic difficulties which welling and advection for nutrient input, such as restrict sampling in time and space(Paine 1988). boundary current systems of California, Peru- Moreover, indirect methods of dietary assess- Chile, Benguela, Canary Islands, and Somalia, ment, such as stomach, pellet, or fecal analyses, only a few taxa often provide key nutritional may provide less than satisfactory results if diet links between primary producers and secondary composition is biased by differential rates of di- or higher-level consumers, including predatory gestion (reviewed by Duffy and Jackson 1986, fish, birds and mammals (Cushing 1975). Key Erilcstad 1990). Fortunately, the ratios of natu- speciesof primary consumersinclude zooplank- rally-occurring isotopes of carbon (13C/12C)and ton (Calanus spp. and Euphausia spp.) and sev- nitrogen (15N/14N)in consumer tissues provide eral species of epipelagic fish (Engraulis spp., insight into food-web structurethat can comple- Sardinops/Sardinaxand Clupea spp.); theseprey ment conventional dietary assessments.Stable are then consumed by secondary predators, in- isotope analysis (SIA) can be a useful comple- cluding many marine birds and mammals (Ver- ment because stable isotope ratios in consumer tissues reflect those in their diet (DeNiro and Epstein 1978, 1981). Metabolic processes in- I Received4 June 1996. Accepted 18 December volved in the synthesisof proteins in consumers 1996. result in preferential loss of the lighter isotope

W71 328 WILLIAM J. SYDEMAN ET AL. and hence enrichment of the 15N/14Nratio in dividual pairs. We also obtained muscle tissue their tissues. In marine ecosystems a step-wise samples of northern sea lions (Eumetopius ju- enrichment of 15Ntypically occurswith each tro- butus) from dead pups collected at AN1 (n = 5, phic level (Wada et al. 1987, Rau et al. 1992, July-October) and SEFI (n = 2) and a few (see Hobson et al. 1994). This change in isotope ra- below) pectoral muscle samples from seabirds tios between trophic levels appears to be rela- on SEFI (Common Murre, Pigeon Guillemot, tively constant at -3%0 (%o = parts per thou- Rhinoceros and Cassin’s Auklets) salvaged dur- sand), so measurements of 15N/14Nratios in ing June-September, 1993. food-web components can be used to infer trophic relationships. Stable-carbon isotope values PREY SAMPLES provide insight into the source of feeding, in- Based on previous studies of bird diets in the cluding inshore vs. offshore foraging in marine region (Briggs et al. 1988, Ainley et al. 1990, habitats (Hobson et al. 1994). Croll 1990), we selectedthe following prey spe- Isotopic studies to date have used primarily cies for SIA: juvenile short-bellied rockfish (Se- whole bodies or specific tissues of consumers. bastes jordanii), northern anchovy (Engruulis For investigation of avian diets, bird eggs are mordux), juvenile Pacific sardine (Surdinops sa- another potential source of material, since they jux), adult market squid (Loligo opalescens), ju- are formed from nutrients derived from the diet venile sablefish (Anoplopomu jimbriu), juvenile of the laying female (Krapu 1981, Afton and lingcod (Ophiodon elongutus), juvenile king Ankney 1991, Schaffner and Swart 1991). The salmon (Oncorhynchus tshuwytschu), and adult use of eggs as source material is important be- krill (Euphuusiu pacifica and Thysunoessu spi- cause knowledge of marine bird diet during egg niferu). Prey samples (excluding krill) were ob- formation is scant, yet critical for evaluating hy- tained by capturing Rhinoceros Auklets as they pothesesconcerning timing of breeding and oth- returned with food at dusk to feed their chicks er life history traits. In this study, we examine (see below). Samples were identified and frozen species in the coastal marine food-web of the for later SIA analyses. We obtained krill samples Gulf of the Farallones, California, using a com- from the National Marine Fisheries Service bination of SIA to infer relative trophic position (NMFS) that were collected during rockfish during the early breeding period, and conven- stock recruitment surveys made in the Gulf of tional diet studiesto infer chick diets later in the the Farallones during February 1994 (n = 5), season. We compare and contrast these results and one sample from a fresh Western Gull re- and discuss the utility of using both approaches gurgitation collected on SEFI in June 1993. to understanding seabird tropho-dynamics in this coastal food-web. We also compare our results STABLE ISOTOPE MEASUREMENTS with those of Rau et al. (1983) who used SIA to investigate other portions of the California Cur- A minimum of 2 ml of albumen was extracted rent marine ecosystem. from each egg, stored frozen in glass scintilla- tion vials and later freeze-dried. Lateral muscle METHODS tissue from fish and marine mammals also was freeze-dried, powdered and subjected to lipid PREDATOR SAMPLES extraction using a Soxhlet apparatus with chlo- Eggs of Brandt’s Cormorants (Phalacrocorux roform solvent. Krill samples were similarly penicillutus), Pelagic Cormorants (P. pelagicus), treated, but also were soaked with 1 N HCL to Western Gulls (Larus occidentalis), Common remove carbonates before isotopic analysis. We Murres (Uriu aulge), Pigeon Guillemots (Cep- loaded samples for 13C/12Cand 15N/14Nanalysis phus columba), Rhinoceros Auklets (Cerorhincu in Vycor tubes with wire-form CuO, silver foil, monocerutu), and Cassin’s Auklets (Ptycho- and elemental Cu, then sealed tubes under vac- rumphus uleuticus) were collected from South- uum and cornbustedthe material at 800°C for at east Farallon Island (SEFI; 37”N 123”W) and least 6 hr. Carbon dioxide and nitrogen gas was Aiio Nuevo Island (ANI, 36”N 122”W) located separated cryogenically and analyzed using a in waters off central California. Eggs were col- VG OPTIMA isotope-ratio mass-spectrometerat lected early in the breeding season between the National Hydrology Research Institute, Sas- April and June 1993 to facilitate relaying by in- katoon, Saskatchewan.Stable isotope concentra- CALIFORNIA SEABIRD TROPHIC DYNAMICS 329 tions are expressedin delta (6) notation as parts identifications we used collections maintained at per thousand according to the following: PRBO, Southwest Fisheries Science Center (La Jolla, California) or Moss Landing Marine Lab- 6X = KRsamp,&an& - 11 X 1,000 oratory (Moss Landing, California). where X is 15Nor 13Cand R is the corresponding RESULTS ratio 15N/r4Nor r3C/12C.Rtandard for lJN and 13C are atmosphericN, (AIR) and the PDB standard, TROPHIC POSITION respectively. Based on numerous measurements S5N measurements. WN values (mean t SD) of organic standards,the analytical precision of ranged from 11.2 + OS%0 in krill to 19.8 2 these measurements is estimated to be ? 0.1%0 0.6%0 in northern sea lions (Table 1, Fig. 1); and 2 0.3%0 for carbon and nitrogen, respec- overall, species differences were significant tively. (ANOVA: F,, 82 = 70.9, P < 0.001). Krill was significantly different from fish, except for ling- CONVENTIONAL TECHNIQUES OF DIET cod, sardine, sablefish, and squid (Fig. 1, Bon- ASSESSMENT ferroni inequality, P < 0.001). Among fish, ling- To quantify food habits of seabirds,we observed cod showed the lowest 615Nvalues and northern adults feeding chicks. Common Murres and Pi- anchovy the highest. We found the following geon Guillemots carry single prey items in their statistically significant differences among fish: beaks to provision chicks, and diet items can be (1) values for lingcod were lower than those for identified, often to species,by trained observers salmon, anchovy, and short-bellied rockfish, (2) situated in blinds near breeding colonies. We those for sablefishand squid were lower than for conducted observations on approximately 100 anchovy and short-bellied rockfish, and (3) those pairs of murres and 50 pairs of guillemots for sablefish were lower than salmon (Bonfer- throughout the chick-rearing period (murres: late roni inequality, all P < 0.02). Among birds, two May through early July; guillemots: late June distinct groups were evident (Table 1, Fig. 1). through mid-August). To quantify food habits of 615N values derived from eggs of Brandt’s and Rhinoceros Auklets, we collected food items Pelagic Cormorant, Pigeon Guillemot, and Rhi- from adults captured at SEFI in mist nets opened noceros Auklet were similar, but, in turn were near dusk, monitored continuously (-1.5 hr.), different from the second group composed of and closed just after night fall. We conducted at Cassin’s Auklet, Common Murre, and Western least three capture sessionsat each of three lo- Gull (Bonferonni inequality, all P < 0.001). The cations on the Farallones from 15 June to 30 6r5N value for northern sea lion muscle was sig- August. Brandt’s Cormorant diet was deter- nificantly greater than for birds (Bonferroni in- mined from analyses of remains in pellets (fish equality, all P < 0.001). The 615N values mea- otoliths, squid beaks) collected from nesting ar- sured for seabird pectoral muscle were: Cassin’s eas. A total of 133 pellets from 108 breeding Auklet: 14.0 ? 1.2%0 (n = 4), Common Murre: pairs (l-3 pellets per nest) was collected during 17.3 ? 0.7%0 (n = 3), Pigeon Guillemot: 15.2%0 August-September. Pellets were soakedin water, (n = 2), and Rhinoceros Auklet: 17.8%0(n = 2). otoliths were extracted from organic material, Trophic levels were approximated from our dried, and sorted. Otoliths were identified and nitrogen isotope data by applying a trophic en- measured using a dissection microscope and the richment factor throughout the food-web after reference collection of identified otoliths main- assuming that krill occupied trophic level 2.5 tained at Point Reyes Bird Observatory (PRBO) (Sanger 1987). The isotopic fractionation factor or the Museum of Vertebrate Zoology, Univer- of 3.1%0 was determined previously by Hobson sity of California, Berkeley. Diet of Cassin’s (1995) and correspondsto changesexpected be- Auklets was based on about 100 regurgitations tween diet and egg albumen. This value is sim- collected each year from parents returning to ilar to trophic enrichment factors derived for feed chicks at night. About 10 collections were other components of marine food-webs off Cal- made every 10 days throughout chick-rearing, ifornia and elsewhere (Rau et al. 1983, 1992, mid-May through July. Each regurgitation was Hobson and Welch 1992). We therefore applied weighed and frozen for analysis. During analy- it throughout the seabird food-web of the Gulf sis, otoliths, squid beaks, and whole bodies of of the Farallones (independent of tissue type) as crustaceanswere separatedand counted. To aid follows: 330 WILLIAM J. SYDEMAN ET AL.

TABLE 1. Stable-carbonand nitrogen isotope abundance(2 ? SD) of componentsof the Gulf of the Farallones food web. Source material includes egg albumen for birds, muscle tissue for fish and northern sea lions, and whole bodies for krill.

Species n S’C(%o) 6”N(%.) Trophic level’ Crustaceans Krill 5 -20.2 ? 0.3 11.2 5 0.5 2.5 Cephalopods Market squid -17.1 ” 0.2 12.3 C 0.4 2.9 Fishes Lingcod -18.3 ? 0.5 12.5 2 0.3 2.9 Sablefish - 17.3 2 0.2 12.8 ? 0.7 3.5 Pacific sardine - 17.0 t 0.3 12.9 2 0.1 3.1 Shortbelly rockfish -17.1 5 0.3 13.8 2 0.2 3.3 King salmon - 17.5 -c 0.2 13.8 ? 0.2 3.3 Northern anchovy - 16.8 ? 0.4 13.9 rfr0.8 3.4 PlanktivorousBirds Cassin’s Auklet - 18.2 r 0.5 13.9 2 1.0 3.4 Western Gull 5 - 16.4 2 0.3 14.4 * 1.1 3.5 Common Murre 3 - 16.9 2 0.5 14.8 2 0.9 3.7 PiscivorousBirds Pelagic Cormorant 5 - 17.7 2 0.2 16.7 2 0.5 4.3 Rhinoceros Auklet 6 - 17.7 2 0.7 16.9 ir 0.5 4.3 Pigeon Guillemot 13 - 17.7 lr 0.2 16.9 + 0.5 4.3 Brandt’s Cormorant 7 - 15.9 2 0.3 17.3 2 0.2 4.5 Mammals Northern sea lion 5 - 15.2 5 0.5 19.8 f 0.6 5.3

a See text for denvation of trophic level estimates.

TL = 2.5 + (615N - 11.2)/3.1 feronni inequality, all P < 0.01). The 613Cvalue for northern sea lion was significantly greater where TL is the trophic level of an organism and than for birds, except Brandt’s Cormorant (Bon- 615Nis its stable-nitrogen value in %o. Estimates ferroni inequality, all P < 0.003). Muscle tissue of trophic level for each speciesare presentedin 613C values for seabirds were: Cassin’s Auklet: Table 1. -18.3 -+ 0.4%0 (n = 4), Common Murre: -16.6 g3C measurements. V3C values ranged from 2 0.5%0 (n = 3), Pigeon Guillemot: -16.2%0 (n -20.2 ? 0.3%0 in krill to 15.2 ? OS%0in north- = 2), and Rhinoceros Auklet: - 17.1%0 (II = 2). em sea lions (Table 1, Fig. 1); overall, species differences were significant (ANOVA: F,,,,, = CHICK DIET COMPOSITION 52.2, P < 0.001). The 613C value for krill was significantly less than that for fish (Bonferroni Brandt’s Cormorant. Fifty-six of 108 samples inequality, all P < 0.001). Among fish, lingcod contained a total of 1,108 otoliths from at least showed the lowest 613C value and northern an- 16 identified species.A small number (1.6%) of chovy the highest. The value for lingcod was the otoliths could not be identified, primarily due significantly lower than those for other fish. The to wear or fragmentation. The number of otoliths value for salmon was significantly lower than for per sample ranged from 1 to 96, and number of anchovy and short-bellied rockfish (Bonferroni speciesper sample ranged from 1 to 6. The diet inequality, all P < 0.015). Among birds, two composition was diverse, including 13 families distinct results emerged: (1) Brandt’s Cormorant of fish (Table 2). Scorpaeniids (rockfish; pre- showed significantly higher ?Y3C values than sumably mostly Sebastes jordani) comprised other birds, except Western Gull, and (2) Pigeon over 57% of the diet by number; bothids and Guillemot, Rhinoceros and Cassin’s Auklet, and pleuronectids (flatfish) comprised -20%. Batra- Pelagic Cormorant were all similar, yet different choidids (midshipman) and engraulids (ancho- from Common Murre and Western Gull (Bon- vies) comprised an additional 16%. Carangids, CALIFORNIA SEABIRD ‘I’ROPHIC DYNAMICS 331

OFFSHORE INSHORE < >

20 . EIMU

PIGWRHAU . . PECO 3 16 0‘

.g COMU . . WEGU g 14 CAAU . . . ANCH SALM l SBRF

SABLE. . SARD LING . . SQUI 12

l EUPH

10 I I I I 1 -21 -20 -19 -18 -17 -16 -15 -14

FIGURE 1. Trophic structurein the Gulf of the Farallones based on seabird egg albumen, sea lion and fish muscle tissue, and whole bodies of krill (see Table 1 for statistical details). Abbreviations: EUPH (krill = euphausiids),LING (lingcod), SABLE (sablefish),SQUI (squid), SARD (sardine), SALM (salmon), SBRF (short- bellied rockfish), ANCH (anchovy), CAAU (Cassin’s Auklet), WEGU (WesternGull), COMU (Common Murre), PIGU (Pigeon Guillemot), RHAU (Rhinoceros Auklet), PECO (Pelagic Cormorant), BRCO (Brandt’s Cormo- rant), EJMU (northern sea lion)

cottids, and ophiids also were taken, but in relatively low numbers. Adult murres did not feed atively low numbers. crustaceansto their chicks. Cassins’ Auklet. The diet of Cassin’s Auklet Pigeon Guillemot. We were unable to docu- chicks was composedof crustaceans,principally ment the use of prey to the species-level; thus, two species,Euphausia paci$ca and Thysanoes- much of the compositional analysis of Pigeon sa spinifera (Table 3). Mysids (Acanthomysis Guillemot chick diet is restricted taxonomically columbiae) and another euphausiid crustacean, to the family-level. The diet composition of Pi- Nyctyphanes simplex, comprised -20% of the geon Guillemot chicks was more diverse than diet. Amphipods, decapods, and fish comprised that of Common Murre chicks (Table 5). Rock- small amounts of the diet. fish (scorpaenids) and sculpins (cottids) were Common Murre. Principle diet items of Com- eaten most frequently, together comprising mon Murre chicks included juvenile rockfish, -90% of the diet. Flatfish (bothids and pleuro- mainly Sebastesjordani, and northern anchovy nectids) comprised -5% of the chick diet. Other (Engraulis mordax) (Table 4). Salmon (Oncho- prey eaten included pholids and stichaeids(gun- rhynchus tshawytscha), butterfish (Peprillus nels and warbonnets), clinids and blenniids simillimus), night smelt (Spirinchus starksi), and (kelpfish and blennies), ophidiidsfliparids (cus- cephalopods (Z&go opalescens) each made up keels and snailfish), and cephalopods(squid, but < 5% of the diet. Flatfish (bothids and pleuro- especially octopus). nectids) and other specieswere consumed in rel- Rhinoceros Auklet. The diet of Rhinoceros 332 WILLIAM J. SYDEMAN ET AL.

TABLE 2. Brandt’s Cormorant diet composition in TABLE 4. Common Murre diet compositionin 1993 1993 based on numericaloccurrence. The percent con- based on numerical occurrence.Percent of total iden- tribution of total identified individuals (n = 1,090) is tified individuals (n = 1,985) is shown. shown.

%Numerical Prey* occurrence Osteichthyes Osteichthyes Bothidae/Pleuronectidae Batrachoididae Engraulidae Poricthys notatus 9.2 Engraulis mordax 62.8 Bothidae Osmeridae Citharichthyssordidus 6.5 Spirinchusstarksi 3.0 Citharichthysstigmaeus 4.3 Salmonidae 1.0 Carnangidae Scorpaenidae(subtotal) 31.2 Trachurussymmetricus 2.0 Sebastesjordani 30.0 Cottidae 1.5 Sebastesspp. 1.2 Embiotocidae Stomateidae Cymatogasteraggregata 0.1 Peprillus simillimus 1.0 Engraulidae OtheP 1.2 Engraulis mordax 7.1 Cephalopoda Gobiidae 0.1 Merluccidae 0.1 Loligo opalescens 1.0

Ophiidae 1 Includes: Clupea harengus, Cololobris saira, Oqjulis califomica, and Chilara taylori 1.1 ocropus rufescens. Osmeridae (unidentified) 0.1 Pleuronectidae(unidentified) 2.8 Glyptocephaluszachirus 1.7 Auklet chicks was dominated by northern an- Lepidopsettabilineata 2.9 Sciaenidae chovy (Engraulis mordax), short-belly rockfish Genyonemuslineatus 1.4 (Sebastesjordani), and sardines (Sardinops sa- Scorpaenidae jax); together, these species comprised -80% of Sebastesspp. 57.2 the diet (Table 6). An additional 7 families of aKnown prey not represented in 1993 include Microgadus proximus fish and cephalopods were taken in low frequen- (Gadidae), Hexagrammidae, Parophyrs vetulus (Pleuronectidae), Stoma- teidae, Zamolepididae, and Lo/&o opalescrns and Octopus rubescens cies. Predation on sardines by this species in (Cephalopoda). 1992-1993 represents the first report of this prey in seabird diets since the recent recovery of the

TABLE 5. Pigeon Guillemot diet composition in TABLE 3. Cassin’s Auklet diet composition in 1993 1993 based on numerical occurrence.Percent of total based on numerical occurrence.Percent of total iden- identified individuals (n = 4,268) is shown. tified individuals (n = 3,487) is shown. Sample is based upon regurgitations(n = 62) collected through- % Numerical out the chick-rearing season,May-July. Preya Occ”*e”ce Osteichthyes Bothidae/Pleuronectidae 5.3 Brotulidae 0.1 Crustaceans(subtotal) 99.2 ClinidaeiBlennidae 1.1 Amphipods (subtotal) >o. 1 Cottidae 28.7 Decapods 1.1 Ophidiidae/Liparidae 1.2 Euphausiids(subtotal) 83.4 Pholidae/Stichaeidae 0.9 Euphausiapaci$ca 48.3 Scorpaenidae(subtotal) 61.5 Nyctyphanessimplex 7.6 Sebastesjordani 38.7 Thysanoessaspinifera 27.5 Sebastesspp. 22.8 Mysids 14.6 Cephalopoda Osteichthyes(subtotal) co.2 Bothidae co. 1 Loligo opalescens 0.1 Citharichthyssordidus

‘Known prey not represented in 1993 include Cephalopoda, Gastropo- * Known prey not represented in 1993 mclude Ammodytidae, Batra- da, Gammarids, Hypenids, and Cypnids, and Nematocelis dijjkdis (Eu- choididae, Engraulidae, Gobiidae. Labridae, Merluccidae, and Scomber- phausiidae). esocidae. CALIFORNIA SEABIRD TROPHIC DYNAMICS 333

TABLE 6. Rhinoceros Auklet diet composition in and our sampling included many food-web com- 1993 based on numerical occurrence.Percent of total ponents, our conclusions are limited to late identified individuals (n = 348) is shown. spring and early summer. It is unknown how well this food-web structure would apply to other seasons, but we suspect many similarities Osteichthyes would be found. Ammodytidae Rau et al. (1983) studied the pelagic food web Ammodyteshexapterus 1.0 of the southern California Bight and concluded Anoplopomatidae that this food-web contained five trophic levels Anoplopomafimbria 1.7 Clupeidae between “netted plankton” (composedprimarily Sardinopssajax 18.1 of zooplankton) and an apex predator, the great Cottidae white shark (Carcharodon carcharius). Rau et Scorpaenichthysmarmoratus 1.0 al. (1983) based their food web structureon Q3C Engraulidae values, whereas we based our food web on U5N Engraulis mordax 37.4 Hexagrammidae values. In central California, great white sharks Ophiodonelongatus 6.0 are known to extensively feed upon seals and Merluccidae sea lions (Ainley et al. 1985) indicating an Merlucciusproductus 1.0 “apex” trophic level in the Gulf of the Faral- Salmonidae Oncorhynchustshawytscha 3.2 lones food-web. The Gulf of the Farallones and Scorpaenidae(subtotal) 29.6 southern California Bight are separatedby ap- Sebastesjordani 25.0 proximately 300 km of coastline and are influ- Sebastesspp. 4.6 enced, to a certain extent, by different oceano- Stromateidae graphic processes.However, many components Peprillus simillimus Cl.0 of these food-webs are similar, including large Cephalopoda populations of Western Gulls, Brandt’s and Pe- Loligo opalescens Cl.0 lagic Cormorants, Pigeon Guillemots, and Cas- Octopusrufescens Cl.0 sin’s Auklets, and a reliance on zooplankton Otherb 1.0 (krill), anchovies and rockfish by breeding sea- A Known prey not represented m 1993 mcludes Cololabris saira (Scom- birds in both systems(Anderson et al. 1982, Sy- beresocidae). h Includes: Ciupea harmgus, Oxyj’ulis californica, Lamperramdenrara. deman et al. 1991, Ainley et al. 1995). Lesridium ringens, and Myctopbiidae. TEMPORAL CHANGES IN DIET AND FORAGING DISTRIBUTION sardine population in central California (Barnes With the exception of Common Murres, chick et al. 1992). diet and trophic position of adults based on the isotopic analysis of egg albumen were concor- DISCUSSION dant. SIA indicated that Common Murres, West- TROPHIC STRUCTURE em Gulls and Cassin’s Auklets were planktivo- Using our isotopic model to calculate average rous during the period of egg formation, whereas trophic level, we determined that the marine Brandt’s and Pelagic Cormorants, Rhinoceros food-web in the Gulf of the Farallones is com- Auklets, and Pigeon Guillemots were piscivo- posed of approximately five trophic levels with rous. The diet of seabird chicks in the summer marine birds occupying levels 3.4 to 4.5, and provides corroborating evidence for the pisciv- northern sea lions occupying the fifth trophic orous diets of both cormorants,Rhinoceros Auk- level (Table 1, Fig. 1). However, isotopic anal- lets and Pigeon Guillemots, and the planktivo- yses of egg albumen reflects a narrow window rous diet of Cassin’s Auklet. In addition, Ainley of dietary integration when compared with mus- et al. (1990) found a varied diet of Western Gull cle or other hard tissues(Hobson 1995). For Far- chicks, including marine arthropods mostly allon Island seabirdsin 1993, the period of egg- composed of euphausiids. laying included the months of April-June (Sy- Common Murres occupied a lower trophic deman and Eddy 1995). Although this represents level (TL = 3.7) consistent with a diet of krill typical timing of breeding for these birds (Ain- during the egg formation period. Other evidence ley and Boekelheide 1990, Sydeman et al. 1991) from the colony on the Farallon Islands and Gulf 334 WILLIAM J. SYDEMAN ET AL. of the Farallones also suggeststhat these birds waters of the Gulf of the Farallones were more feed on invertebrates during spring (Boekelheide enriched in 6°C than offshore or pelagic feeders, et al. 1990; PRBO, unpubl. data). Isotopic anal- a pattern found in other marine food webs (Hob- yses of our small sample of muscle tissue from son and Welch 1992, Hobson et al. 1994). No- murres indicates a higher trophic level (TL = tably, the two species of krill are primarily pe- 4.5) later in the season corresponding to a pi- lagic in their distribution and showed the lowest scivorous diet. Although our sample size of Q3C values. Among birds, Cassin’s Auklet, a pe- murre muscle tissue is small, these data indicate lagic feeder, showed the most depleted 6r3C val- a switch in trophic level between spring and ues, whereas Brandt’s Cormorant, which feeds summer for Common Murre adults in the Gulf in nearshorewaters of the Gulf of the Farallones of the Farallones. A switch in the diet and tro- and in coastal estuaries (Ainley et al. 1981, phic level for Western Gulls also may occur as 1990) was most enriched. Common Murres and this speciesfeeds extensively on fish later during Western Gulls, which feed in both pelagic and chick-rearing (Sydeman et al. 1991); thus, fur- neritic habitats (Ainley et al. 1990, Allen 1994), ther isotopic investigations on these species had intermediate IY3C values. Thus, SIA also ac- would be informative. curately reflected the at-sea foraging habitat of Our findings suggest that some marine birds seabirds and their prey in the Gulf of the Far- may have the ability to track the most abundant allones. and/or energetically profitable prey on relatively short time scales. In marine ecosystemscharac- CONCLUSIONS terized by extensive changes in physical ocean- Our study of egg albumen and pectoral muscle ographic processes,such as upwelling and coast- tissue illustrates the dynamic nature of seabird al advection in the California system, food-web trophic relationships in the Gulf of the Farallo- development may be temporarily delayed and nes, California. In doing so, we have demon- spatially variable. In turn, this could have sub- strated the power of conducting SIA on several stantial effects on seabird demography. Species tissues from a single species to infer temporal able to respond to rapidly changing foraging and spatial variation in trophic relationships and conditions, or those able to use prey occupying the utility of the stable-isotope approach in de- different trophic levels, may have an advantage lineating trophic relationships. This study adds in being able to maintain relatively constant en- to the growing number of reports (Hobson and ergy budgets. In variable marine environments, Welch 1992, Rau et al. 1992, Hobson et al. seabirds capable of trophic-level switching may 1994) affirming the importance of this technique therefore have a greater chance of successfulre- in determining marine food-web structure and production under unfavorable conditions. In- foraging dynamics. However, conventional diet deed, if we consider interammal variation in the studiesare the only means of establishing details reproductive performance of Common Murres, of the types and amounts of prey taken. Thus, Cassin’s Auklets, and Western Gulls with that of whereas SIA provides information on trophic re- Brandt’s and Pelagic Cormorants and Pigeon lationships and structure, and conventional di- Guillemots on SEFI, the latter group shows con- etary assessmentsprovide details of prey utili- siderably more annual variability (Ainley et al. zation patterns, both techniques, when used to- 1995). Although differences in reproductive ef- gether, provide a powerful means of detecting fort (clutch size) and foraging range (nearshore patterns in marine ecosystems. Notably, both vs. pelagic) explain much of these patterns, techniques indicate that the seabird food-web of questions remain concerning which foraging the central California Current ecosystem is dy- characteristicsallow species to follow different namic. This fundamental result suggests that reproductive options. Trophic-level switching food-web models based on static trophic rela- may provide a mechanism for opportunistic spe- tionships would inadequately describe the Cali- cies to cope with limited or uncertain forage re- fornia Current marine ecosystem. serves. Results of SIA also are compatible with in- ACKNOWLEDGMENTS formation on the at-sea foraging habitat of sea- We thank the Farallon National Wildlife Refuge, U.S. birds in the Gulf of the Farallones. In general, Fish and Wildlife Service, and Aiio Nuevo State Re- organisms which foraged in benthic or nearshore serve, California Department of Parks and Recreation CALIFORNIA SEABIRD TROPHIC DYNAMICS 335

for permission and encouragement to conduct this I? WOLF P 1992. Recent populationstrends and study on wildlife reservesunder their management.Fi- abundanceestimates for the Pacific sardine (Sar- nancial support for SIA and analysis of conventional dinops sujax). California Coop. Ocean. Fish. In- dietary assessmentwas provided by contractCX-8140- vest. Rep. 33:60-76. 93-008 from NOAA/Gulf of the Farallones National BOEKELHEIDE,R. J., D. G. AINLEY,S. H. MORRELL,H. Marine Sanctuary to WJS and Walter M. Jarman of R. HUBER, AND T. J. LEWIS. 1990. Common the University of California at SantaCruz (UCSC). We Murre, p. 245-275. In D. G. Ainley and R. J. Boe- also thank the Bradford and Homeland Foundations, kelheide [eds.], Seabirds of the Farallon Islands: Friends of the Farallones, U.S. Fish and Wildlife Ser- ecology, dynamics, and structureof an upwelling- vice. CanadianWildlife Service. and membersand do- system community. Stanford Univ. Press, Palo nors of PRBO for financial assistance.Stable isotope Alto, CA. analyseswere performed with the help of L. Wassenaar BRIGGS,K. T, D. G. AINLEY,L. B. SPEAR,I? B. ADAMS, and R. George of the National Hydrology Institute AND S. E. SMITH. 1988. Distribution and diet of (Saskatoon).Some samples were collected by K. Ono Cassin’s Auklet and Common Murre in relation to of UCSC and M. Hester of PRBO at Afio Nuevo Is- central California upwellings. Proc. Int. Ornithol. land, and by S. Ralston of National Marine Fisheries Congr. 18:982-990. Service (NMFS) aboard the RV David Starr Jordan. CROLL,D. A. 1990. Physical and biological determi- Identifications of fish prey were confirmed by K. Sec- nants of the abundance,distribution, and diet of uma and T Laidig of the NMFS, Tiburon Laboratory, the Common Murre in Monterey Bay, California. Tiburon, California. D. J. Long helped with the iden- Stud. Avian Biol. 14:139-148. tification and measurementof cormorant otoliths. L. GUSHING,D. H. 1975. Marine ecology and fisheries. B. Spear of PRBO, S. Smith of Southwest Fisheries Cambridge Univ. Press, Cambridge. Center in La Jolla, CA and I? Slattery of Moss Landing DENIRO,M. J., AND S. EPSTEIN.1978. Influence of diet Marine Laboratoriesaided in the identificationof crus- on the distribution of carbon isotopesin animals. taceans. D. Croll, L. Spear and D. Ainley provided Geochem. Cosmochim. Acta 42:495-506. helpful comments on earlier versions of this manu- DENIRO,M. J., AND S. EPSTEIN.1981. Influence of diet script. This is PRBO contributionno. 668. on the distributionof nitrogen isotopesin animals. Geochem. Cosmochim. Acta. 45:341-351. LITERATURE CITED DUFFY,D. C., AND S. JACKSON.1986. Diet studiesof seabirds: a review of methods. Colonial Water- AFTON,A. D., AND C. D. ANKNEY. 1991. Nutrient re- birds 9:1-17. serve dynamics of breeding Lesser Scaup: a test of competing hypotheses.Condor 93:89-97. ERIKSTAD,K. E. 1990. Winter diets of four seabird species in the Barents Sea after a crash in the AINLEY,D. G., D. A. ANDERSON,AND l? KELLY. 1981, Feeding ecology of marine cormorants in south- capelin stock. Polar Biol. 10:619-627. western North America. Condor 83: 120-131. HOBSON,K. A. 1995. Reconstructingavian diets using AINLEY, D. G., AND R. J. BOEKELHEIDE[EDS.]. 1990. stable-carbonand nitrogen analysis of egg com- Seabirdsof the Farallon Islands: ecology, dynam- ponents: patterns of isotopic fractionation and ics, and structure of an upwelling-system com- turnover. Condor 97~752-762. munity. Stanford Univ. Press, Palo Alto, CA. HOBSON.K. A.. AND H. E. WELCH. 1992. Determina- AINLEY,D. G., R. I? HENDERSON,H. R. HUBER,R. J. tion of trophic relationshipswithin a high Arctic BOEKELHEIDE,S. G. ALLEN, AND T. L. MCELROY. marine food web using ‘C and 15Nanalysis. Mar. 1985. Dynamics of white shark/pinnipedinterac- Ecol. Progr. Ser. 84:9-18. tions in the Gulf of the Farallones. Mem. South. HOBSON,K. A., J. E PIATT,AND J. PITOCHELLI.1994. California Acad. Sci. 9:109-122. Using stable isotopesto determine seabirdtrophic AINLEY,D. G., C. S. STRONG, T M. FENNIMAN,’ AND R. relationships.J. Anim. Ecol. 63:786798. J. BOEKELHEIDE.1990. The feeding ecology of KRAPU,G. L. 1981. The role of nutrient reserves in Farallon seabirds,p. 51-127. In D. G. Ainley and Mallard reproduction.Auk 98:29-38. R. J. Boekelheide [eds.], Seabirdsof the Farallon PAINE,R. T 1988. On food webs: roadmapsof inter- Islands: ecology, dynamics, and structure of an actions or grist for theoretical development.Ecol- upwelling-system community. Stanford Univ. ogy 69:1648-1654. Press, Palo Alto, CA. RAU, G. H., A. J. MEARNS,D. R. YOUNG,R. J. OLSEN, AINLEY,D. G., W. J. SYDEMAN,AND J. NORTON. 1995. H. A. SCHAFER,AND I. R. KAPLAN.1983. Animal Upper trophic level predatorsindicate interannual ‘3C/1zC correlates with trophic level in pelagic negative and positive anomalies in the California food webs. Ecology 64:1314-1318. Current food web. Mar. Ecol. Progr. Ser. 118:69- RAU, G. H., D. G. AINLEY,J. L. BENGSTON,J. L. TOR- 79. RES,AND T. L. HOPKINS.1992. rSN114Nand ‘C/i2C ALLEN, S. G. 1994. The distribution and abundance in Weddel Sea birds, seals, and fish: implications of marine birds and mammals in the Gulf of the for diet and trophic structure. Mar. Ecol. Progr. Farallonesand adjacentwaters, 1985-1992. Ph.D. Ser. 84: l-8. diss.. Univ. California. Berkelev. CA. SANGER,G. A. 1987. Trophic levels and trophic re- ANDERSON,D. A., E GRESS,AND- K. MAIS. 1982. lationships of seabirds in the Gulf of Alaska, p. Brown Pelicans: influence of food supply on re- 229-257. In J. I? Croxall [ed.], Seabirds:feeding production. Oikos 39:23-3 1. ecology and role in marine ecosystems. Cam- BARNES,J. T, L. D. JACOBSON,A. D. MACCALL,AND bridge Univ. Press, Cambridge. 336 WILLIAM J. SYDEMAN ET AL.

SCHAFFNER,E C., AND F? K. SWART. 1991. Influence in laying date and its relationship to individual of diet and environmental water on carbon and quality in Common Murres. Condor 97:1048- oxygen isotopic signaturesof seabirdegg carbon- 1052. ate. Bull. Mar. Sci. 48:23-38. VERMEER,K. 1981. The importance of plankton to SYDEMAN,W. J., J. E PENNIMAN,T M. PENNIMAN,l? Cassin’s Auklet during breeding. J. Plank. Res. 3: 3 15-329. PYLE,AND D. G. AINLEY. 1991. Breeding perfor- WADA, E., M. TERAZAKI,Y. KABAYA,AND T. MEMOTO. mance in the Western Gull: effects of parentalage, 1987. r5N and r3C abundancesin the Antarctic timing of breeding, and year in relation to food Ocean with emphasis on the biogeochemical availability. J. Anim. Ecol. 60: 135-149. structureof the food web. Deep-Sea Res. 34:829- SYDEMAN,W. J., AND J. 0. EDDY. 199.5. Repeatability 841.