Migratory shearwaters integrate oceanic resources across the Pacific Ocean in an endless summer
- Scott A. Shaffer*,†,
- Yann Tremblay*,
- Henri Weimerskirch‡,
- Darren Scott§,
- David R. Thompson¶,
- Paul M. Sagar‖,
- Henrik Moller§,
- Graeme A. Taylor**,
- David G. Foley††,
- Barbara A. Block‡‡, and
- Daniel P. Costa*
- *Ecology and Evolutionary Biology, University of California, Santa Cruz, CA 95064;
- ‡Centre d’Etudes Biologiques de Chizé, Centre National de la Recherche Scientifique, 79360 Villiers en Bois, France;
- §Department of Zoology, University of Otago, Dunedin, New Zealand,
- ¶National Institute of Water and Atmospheric Research Ltd., Wellington, New Zealand;
- ‖National Institute of Water and Atmospheric Research Ltd., Christchurch, New Zealand;
- **Department of Conservation, Wellington, New Zealand;
- ††Joint Institute for Marine and Atmospheric Research, University of Hawaii, Honolulu, HI 96822; and
- ‡‡Hopkins Marine Station, Stanford University, Pacific Grove, CA 93950
-
Edited by Gordon H. Orians, University of Washington, Seattle, WA, and approved July 5, 2006 (received for review May 8, 2006)
Abstract
Electronic tracking tags have revolutionized our understanding of broad-scale movements and habitat use of highly mobile marine animals, but a large gap in our knowledge still remains for a wide range of small species. Here, we report the extraordinary transequatorial postbreeding migrations of a small seabird, the sooty shearwater, obtained with miniature archival tags that log data for estimating position, dive depth, and ambient temperature. Tracks (262 ± 23 days) reveal that shearwaters fly across the entire Pacific Ocean in a figure-eight pattern while traveling 64,037 ± 9,779 km roundtrip, the longest animal migration ever recorded electronically. Each shearwater made a prolonged stopover in one of three discrete regions off Japan, Alaska, or California before returning to New Zealand through a relatively narrow corridor in the central Pacific Ocean. Transit rates as high as 910 ± 186 km·day−1 were recorded, and shearwaters accessed prey resources in both the Northern and Southern Hemisphere’s most productive waters from the surface to 68.2 m depth. Our results indicate that sooty shearwaters integrate oceanic resources throughout the Pacific Basin on a yearly scale. Sooty shearwater populations today are declining, and because they operate on a global scale, they may serve as an important indicator of climate change and ocean health.
Animal migration is regarded widely as a biological phenomenon driven by seasonal cycles in food supply and weather patterns (1). Although migration patterns of a few marine predator species recently have been established by using electronic tags (2–5), the spatial and temporal dynamics of individuals is unknown for most species. This fact is particularly true of oceanic seabirds that spend up to 90% of their lives in the open sea (6).
Sooty shearwaters (Puffinus griseus) are small seabirds (≈800 g) that are one of the most abundant species in the world (7). Observations of migrating sooty shearwaters at sea suggest that the population conducts a transequatorial pan-Pacific flight (8, 9). However, without tracking individual birds from a known provenance, it is impossible to determine the exact timing and direction of migratory flyways or the individual spatial and temporal patterns of habitat use once at the designated wintering grounds (3). Sooty shearwaters are major consumers of marine resources (10), and given their large population size, which is estimated to be several million birds (11), their potential impact on resources at the wintering grounds is likely to be significant. However, we know little regarding their specific foraging and activity patterns. Recent evidence also shows that sooty shearwater populations are declining in the Northern and Southern Hemisphere, purportedly as a result of climate change (12, 13) and fisheries interactions (14). Studying the at-sea behavior and distributions of shearwaters therefore may be critical for understanding the putative causes of their decline.
Here, we report the extraordinary migratory journeys and broad-scale habitat use of sooty shearwaters by using miniature archival geolocating tags. The data reveal that shearwaters traverse the entire Pacific Ocean while integrating oceanic productivity over a global scale.
Results and Discussion
Sooty shearwaters (n = 19) were tracked with geolocating archival tags from two breeding colonies in New Zealand (see Materials and Methods) for an average of 262 days (±23 days, 222–313 days). The Pacific Ocean migration cycle had a figure-eight pattern (Fig. 1) that lasted an average of 198 ± 17 days. Shearwaters made directed movements between both ends of the Pacific basin with prolonged aggregations on the wintering grounds in the North Pacific. All shearwaters (10 females, 4 males, and 5 undetermined) used the entire Pacific Ocean (Fig. 1) with latitudinal coverage ranging from Antarctic waters to the Bering Sea and longitudinal coverage ranging from Japan to Chile. No sex-specific behavioral tendencies were observed, and each individual from three different breeding pairs dispersed independently of each other (Fig. 1 b–d). These extraordinary migration routes (64,037 ± 9,779 km excluding the breeding period) represent the longest recorded of any animal tracked to date. Although new information has been shed on the remarkable flight capabilities and distribution of albatrosses during sabbatical years by using electronic tags (2, 3, 16), until now, almost nothing comparable was known regarding smaller seabird species (<1 kg) outside their breeding periods. Arctic terns (Sterna paradisaea) undertake a similar transequatorial migration as sooty shearwaters (17, 18) but the movements of individual terns have not been studied, so the true ranges, distances and rates of travel, habitat usage, and migratory flyways remain equivocal. Cory’s shearwaters (Calonectris diomedea), another small seabird (≈700 g), also have been tracked outside the breeding period by using satellite telemetry (19), but the migration tracks were incomplete, so the ranges and extent of migration are unknown.
Shearwater migrations originating from breeding colonies in New Zealand. (a) Interpolated geolocation tracks of 19 sooty shearwaters during breeding (light blue) and subsequent migration pathways (yellow, start of migration and northward transit; orange, wintering grounds and southward transit). The 30° parallels, equator, and international dateline are indicted by dashed lines. (b–d) Representative figure-eight movement patterns of individual shearwaters traveling to one of three “winter” destinations in the North Pacific. These tracks also represent those of three breeding pairs to reveal the dispersion and extent of each pair. The image was created by using the Blue Marble data set (15).
We hypothesize that the figure-eight migration pattern of sooty shearwaters is facilitated by global wind circulation patterns and the Coriolis effect, which influence the long-range trajectories of the birds as they fly rapidly between hemispheres (9). At the start of migration in early April, shearwaters traveled eastward between 40°S and 60°S in the direction of prevailing westerly winds at rates of 536 ± 159 km·day−1. Once shearwaters began heading north from the Southern Pacific Ocean, they appeared to use easterly trade winds to travel northwest across the Pacific. The highest transit rates were achieved during this period (910 ± 186 km·day−1), and birds were displaced an average of 38° (±20°) west of their starting longitude in the Southern Hemisphere (Fig. 1). Upon arrival into cool oceanic waters above 30°N (14.2 ± 3.0°C, tag temperatures before diving, see Fig. 2), shearwaters traveled to one of three wintering grounds, where they resided for the remainder of the migration cycle. On the wintering grounds, shearwater transit rates were significantly reduced (220 ± 37 km·day−1), and movements were confined to discrete regions of the North Pacific with no movement between regions. The figure-eight pattern is completed when birds return to New Zealand waters, again purportedly by using easterly trade winds in a southwesterly direction, at rapid travel speeds (837 ± 134 km·day−1). Remarkably, travel of all birds across the equator is highly synchronized (October 7 ± 5 days), and passage is through a narrow corridor (157°W ± 9°).
Sooty shearwater diving depths and frequency (a), sea surface temperatures experienced (b), and primary productivity (c) at each dive location, in relation to latitude. Sea surface temperatures (SST) were recorded by the archival tags on each bird just before a dive. Primary productivity (PP) was measured remotely by satellite and overlaid onto the locations of each dive (see Materials and Methods). Note the paucity of dives, warm sea surface temperatures, and low primary productivity at low latitudes of the South and North Pacific when shearwaters cross the equatorial region.
Very little information exists on the population structure of migrating sooty shearwaters (9, 10). However, we determined that shearwaters from both our study populations, separated by 850 km, overlapped completely within their distributions in both the South and North Pacific Oceans (Fig. 1). Four birds also traveled to the coast of Chile, where a sizeable population of sooty shearwaters breed (20), before migrating northward. This observation suggests that populations of sooty shearwaters from either side of the Pacific easily could mix within any of the major migration destinations of the North Pacific. Both members of a breeding pair also travel to either different or similar regions of the North Pacific during migration (Fig. 1), suggesting that shearwaters from the same provenance, let alone the same nest, are not geographically constrained. Alternatively each individual could have a specific sector where it winters, because no shearwaters moved between “over-wintering” regions. It is even possible that some individuals return to the same areas from one year to the next, as it has been observed for albatrosses during sabbatical years (2, 3).
The tracking data reveal that sooty shearwaters experience a perpetual cycle of spring, summer, and autumn from year to year. When shearwaters breed in the austral spring/summer, they predominantly travel to cold Antarctic waters (4.4 ± 2.9°C, tag temperatures before diving), where they engage in intense diving activity (Fig. 2), indicative of foraging, to mean depths of 14.2 ± 11.1 m (max 68.2 m). Upon completion of breeding in the austral autumn, shearwaters undertake their migration to one of three areas in the western, central, or eastern North Pacific: (i) Kuroshio and Oyashio Currents region off Japan and Kamchatka Peninsula, (ii) eastern Aleutian Islands and Gulf of Alaska region, or (iii) the California Current region (Fig. 1). The timing of their arrival on the wintering grounds (May 4 ± 13.4 days) occurs when oceanic productivity in the North Pacific exceeds that found in the South Pacific (Fig. 3). The North Pacific wintering grounds are biologically productive because of physical forcing, converging water masses, or coastal upwelling, which promote primary and secondary production attracting fish, squid, and krill consumed by sooty shearwaters (8, 10). Shearwater foraging activity is evident from the propensity of diving activity (Fig. 2; 10.7 ± 8.2 m, max 51.2 m) in highly productive waters during prolonged residency periods (150.9 ± 12.4 days). Thus, sooty shearwaters appear to seek out biologically productive areas to over-winter in the Northern Hemisphere summer and similar regions in the Southern Hemisphere summer when breeding. Indeed, 95% of all recorded dives (n = 4,370) occurred within these highly productive regions in both hemispheres (Fig. 2). The pursuit of an “endless summer” by transiting between hemispheres allows sooty shearwaters to integrate peak oceanic resources on a global scale throughout the year (Fig. 3).
Sooty shearwater latitudinal movements (white open circles, n = 4,375 filtered locations) and primary productivity in the South (yellow) and North (red) Pacific throughout the year. Primary productivity is represented by the mean (±SD) of 8-day productivity from 1997 to 2005 encompassing the geographic regions where sooty shearwaters most frequently occurred (see Materials and Methods). The shaded regions represent the time periods for breeding (BR), migration, and prenuptial (PN) phases. Note that chick-hatching (A) occurs during the austral summer when productivity becomes higher in the South Pacific than in the North Pacific. Upon completion of breeding, the onset of migration coincides with the period when primary productivity becomes higher in the North Pacific (B). Primary productivity in the North Pacific peaks several months before shearwaters return to the South Pacific. Shearwaters conduct the reverse migration in October when productivity is still higher in the North Pacific. However, the timely return is required so adults can court during the prenuptial phase and lay an egg that will hatch (A) when productivity is the highest available at that time (i.e., between A and B).
The ability of sooty shearwaters to travel throughout the Pacific Ocean is a key element of their migration strategy that may have evolved as a mechanism to buffer the birds from interannual variability in Pacific-wide biological productivity, in particular during El Niño Southern Oscillation events (21). Importantly, climate change (22) may impact sooty shearwater populations if warming trends influence major ecosystem dynamics of the North Pacific (e.g., ref. 23). Indeed, sooty shearwater populations have declined in recent years both at breeding colonies in New Zealand (24) and at wintering grounds in the eastern North Pacific (12, 13). These trends were associated with concomitant increases in oceanic temperatures, which may have limited regional biological productivity (12, 13, 24). Significant food limitation over the course of a long-distance migration ultimately could lead to increased shearwater mortality or to an abstention of breeding upon return to the Southern Hemisphere. Distributional shifts between wintering regions of the North Pacific also could account for the observed declines (12, 13). However, this hypothesis seems less likely because our study confirms that shearwaters range throughout the North Pacific Rim, without movement between regions, during a given migration cycle. Fisheries interactions also impose a significant risk for sooty shearwaters as birds are caught as bycatch in several driftnet fisheries in the western North Pacific (14). Hence, the capacity of sooty shearwaters to fly and integrate resources over trans-Pacific scales, combined with recent declines in their abundance, implies that they may serve as a sensitive species to monitor the impacts of global climate change and health of oceanic ecosystems.
Materials and Methods
In January–March 2005, we deployed 33 archival tags (Lotek 2400; Lotek Wireless, St. Johns, NF, Canada) on breeding sooty shearwaters at two colonies in New Zealand (Codfish Island/Whenua Hou, 46.8°S, 167.7°E, n = 25 birds and Mana Island, 41.1°S, 174.8°E, n = 8 birds). Breeding begins in late November to early December and ends in April to early May (6). Archival tags recorded the daily activities of each bird during the latter part of breeding, the complete trans-Pacific migration cycle, and, for some birds, the prenuptial period before the start of the next breeding cycle. All birds were captured in their underground burrows by using methods described in ref. 25. For most study burrows, only a single partner was equipped with an archival tag; however, we equipped both members of three breeding pairs from Codfish Island/Whenua Hou to determine the dispersion of the pair (Fig. 1 b–d). Gender was determined by molecular methods (26). All protocols were approved by the University of California, Santa Cruz, Institutional Animal Care and Use Committee.
Geolocations (n = 4,736) based on ambient light levels from each archival tag were recorded, processed, and error-checked according to previous validation studies (27, 28), which yielded mean errors of 184 ± 114 km and 202 ± 171 km, respectively (see Fig. 4, which is published as supporting information on the PNAS web site). All filtered locations (93% of original locations) were interpolated at 10-min intervals (29) for subsequent analyses of movement patterns and for estimating dive locations of each bird. In addition, the tags recorded pressure and temperature at 432-s intervals throughout the tracking period, which provided a subsample of diving depths and environmental temperatures from activity at sea. The environmental temperatures were used in combination with light-based longitudes to improve or obtain latitude estimates (28, 30) during the equinox periods when light-based geolocation is unreliable (31). All data are presented as mean ± SD.
Twenty shearwaters (15 at Codfish/Whenua Hou and 5 at Mana) were recaptured at their same breeding burrows between October 2005 and February 2006. Each shearwater returned with a tag, and with the exception of minor callusing on the legs of some birds, all adults were in good physical condition with body masses similar to those of nonequipped shearwaters (n = 37; 14 chosen at random for comparison with tracked birds; t test, t 26 = −1.56, P = 0.131). Nineteen of 20 recovered tags provided complete records of tracks during breeding and migration (Fig. 1), and 11 tags provided records of diving depths and environmental temperatures (Fig. 2 a and b). Nonrecovered tags could have fallen off at sea, although no equipped shearwaters were recaptured without a tag. Birds may have avoided detection and changed breeding burrows, or potentially they did not survive the postbreeding migration.
Vertically integrated primary productivity rates for the ocean were calculated for successive 8-day periods from October 1997 to December 2005 (time series analysis only, see below) from measurements of Chlorophyll a, temperature, and photosynthetically available radiation (PAR) by using the method of Behrenfeld and Falkowski (32). Estimates of Chlorophyll a (33) and PAR (34) were derived from the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) carried aboard the Orbview-2 spacecraft. Sea surface temperatures were adapted from Reynolds et al. (35), and estimates of day length were derived from basic celestial mechanics (36). Data were extracted from the resulting global time series for boxes with a spatial extent of 1° longitude and 2° latitude centered at each calculated position and time (2005 only) occupied by the birds (Fig. 2 c). The dimensions of this box were chosen based on the error of the geolocation method (27, 28). To compare the seasonal signal in primary production with the shearwaters’ migration patterns, time series (1997–2005) were extracted for two large boxes that bounded 85% of all filtered geolocations and 90% of all dive locations (box 1: 35°S–70°S by 120°E–140°W in the south; box 2: 30°N–60°N by 130°E–120°W in the north). These time series then were used to develop climatologies for each region frequented by the shearwaters (Fig. 3).
Acknowledgments
We thank the members of the field crew for assistance with tag deployments and recovery; the Whenua Hou Management Committee, Rakiura T̄ıt̄ı Islands Administering Body, and Southland Department of Conservation in New Zealand for permission and support of this project; S. Ruault and C. L. Baduini for conducting the gender analysis; and D. A. Croll for intellectual stimulation of this project. Sea-Viewing Wide Field-of-View Sensor satellite data were provided courtesy of the National Aeronautics and Space Administration’s Goddard Space Flight Center and Orbimage, Inc., and SST data were provided by the National Oceanic and Atmospheric Administration’s (NOAA) National Climate Data Center. This research was conducted as part of the Tagging of Pacific Pelagics program. Funding for production and distribution of primary productivity was provided by the NOAA Satellites and Information, Research and Operation Program; the Gordon and Betty Moore Foundation, the David and Lucile Packard Foundation, the Sloan Foundation, and the National Science Foundation provided financial assistance.
Footnotes
- †To whom correspondence should be addressed. E-mail: shaffer{at}biology.ucsc.edu
-
Author contributions: S.A.S., H.W., D.S., D.R.T., P.M.S., H.M., and D.P.C. designed research; S.A.S., D.S., D.R.T., P.M.S., and G.A.T. performed research; Y.T. contributed new reagents/analytic tools; S.A.S., Y.T., and D.G.F. analyzed data; and S.A.S., Y.T., H.W., D.S., D.R.T., P.M.S., H.M., D.G.F., B.A.B., and D.P.C. wrote the paper.
-
Conflict of interest statement: No conflicts declared.
-
This paper was submitted directly (Track II) to the PNAS office.
- © 2006 by The National Academy of Sciences of the USA
