Observations of gulls foraging on beach-stranded plankton in Otago Harbor, New Zealand

Red-billed gulls (Lams scupzdinus) and southern black-backed gulls (Lams dominicams) exploited a large and highly visible food source provided by beaching of euphausiids (Nyctiphanes australis) and galatheid crab larvae (Munida gregaria) in Otago Harbor. The plankton strandings occurred predominantly in summer, and provided a food source of -270 t dry weigh1 over 170 d along 12 km of shoreline. The largest strands (formed by M. gregaria) covered > 10 ha each, and attracted flocks of > 1,500 gulls for up to 2 d. Gull density was positively correlated with the biomass of strandings, and gull population size in the harbor increased during the peak period of stranding events, The spatial distribution of seabirds is influenced by different processes at differing scales. Density gradients, frontal zones, and water mass properties are correlated with distributions of some bird species over tens Acknowledgments to hundreds of kilometers (Pingree et al. We are grateful to K. Robb and B. Dickson for their assistance in the field. R. Vennel was helpful with pre1984; Hunt and Schneider 1987; Schneider liminary analyses. D. Winkler provided comments on in press). Small-scale physical processes, an early draft of the manuscript. We appreciate imsuch as tidally forced or bathymetrically provements suggested by K. Vermeer and an anonycontrolled upwelling may be sites of bird mous reviewer. aggregations (Braune and Gaskin 1982; Support was provided by the Department of Zoology, University of Otago, a New Zealand UGC PostBrown and Gaskin 1988; Vermeer et al. doctoral Fellowship to S.McC., and an American Ant1987). The attractant for birds at such loarctic Marine Living Resources Program grant (NOAA cations is concentrations of prey created by NA90AA-h-AF022) to C.H. Greene. the interaction of zooplankton behavior and physical forces acting on scales of tens to pelagic larvae and postlarvae of the galathehundreds of meters (Braune and Gaskin id crab, Munida gregaria, provide an abun1982; Obst 1985; Vermeer et al. 1987). dant zooplankton food source for gulls and Plankton-feeding seabirds often aggregate shearwaters in Otago coastal waters (Mcover zooplankton patches. The density of Clatchie et al. 1989). The surface swarms birds is nevertheless often poorly correlated which the birds exploit at sea are mainly a with the density of plankton (Heinemann et summer phenomenon. Swarms of M. greal. 1989). garia larvae appear in the plankton of Otago Plankton are available to seabirds only Harbor in November and persist until about when their vertical movements bring the April (Zeldis 1985). N. awtralis swarms first plankton within the diving range of the birds. appear at the mouth of the harbor in OcSpecies such as gulls, when feeding by surtober (Murdoch 1985). Swarms of adult N. face seizing, take prey from the near-surface australis are relatively common until March while sitting on the water (Ainley 1977). or April and occasionally occur into July Other species can reach plankton within 3and August (McClatchie unpubl.). Beaching 4 m of the surface, and dives to 30 m have of both N. australis and M. gregaria combeen recorded (Ainley 1977). In contrast, monly occurs in summer months, coincimurres were described feeding on near-bendent with surface swarms offshore, but the thic euphausiids at depths >30 m (Hunt et timing of strandings is unpredictable. It is al. 1988) and are known to dive to 180 m uncertain what causes this behavior (but see (Piatt and Nettleship 1985). Whatever the O’Brien et al. 1986). We report here on the foraging range of the predators, birds must occurrence and exploitation by gulls and expend energy to find prey which generally other birds of massive shore strandings of have an extremely patchy spatial distribuN. australis and M. gregaria in Otago Hartion. The persistence of swarms in surface bor. waters may be of short duration and unDaily records of strandings were collected predictable, resulting in food distribution along 12 km of shore in the harbor from 15 patterns with a highly stochastic compoOctober 1989 to 30 January 1990 (Fig. 1). nent. Peruvian gulls and terns, for example, The abundance of seabird species was fed on patches of crustacean plankton that counted at 13 coves or embayments during were available at the surface for as short as low water. The area censused at each lo5 min at a time (Due 1983). In contrast to cation (Fig. 1) was the same for birds and plankton swarms at sea, beachings of plankplankton strandings, but depended on the ton in Tasmania (O’Brien et al. 1986) and size of the beach. The approximate area and New Zealand provide very large, easily dedepth of each patch of stranded zooplanktected, and persistent food sources for seaton was estimated for calculation of strand birds. The strands of zooplankton are exvolume. Dry weight of each strand (B, mg) ploited by birds with a range of foraging was estimated from its volume (V, ml) with modes, but are especially utilized by gulls the regression of Wiebe et al. (1975): that feed by scavenging and surface seizing. The gull species in Otago Harbor, New ZeaB = lO(log~oC’1.048)/0.821 I land, are almost entirely red-billed gulls As strands aged, they rapidly changed color, (Laws scopulinus) and southern blacktexture, and odor, so it was unlikely that the backed gulls (Larus dominicanus). Other same strand was recorded twice at one lobirds which were sighted near strandings of cation on successive days. Correlation beplankton were pied oystercatchers (Haetween seabird numbers at strand sites and matopus$nschi), mallards (Anus platyrhynbiomass of freshly stranded plankton in the chos), white-faced herons (Ardea novaeholharbor was tested with Spearman’s rank landiae), and spotted shags (Stictocarbo method for gulls and birds other than gulls punctatus). Although plankton strands have (Conover 197 1). occasionally been reported (see O’Brien et Zooplankton strandings were almost inal. 1986), their value as a food resource to variably monospecific. The average beach birds has not been examined. area (*SD) covered by strandings was Adult krill (Nyctiphanes australis) and the 80 + 18 m2, and the median area was 22 m2. 1196 Notes