We report on a wintering area off the Pacific coast of Central America for humpback whales (Megaptera novaeangliae) migrating from feeding areas off Antarctica. We document seven individuals, including a mother/calf pair, that made this migration (approx. 8300 km), the longest movement undertaken by any mammal. Whales were observed as far north as 11° N off Costa Rica, in an area also used by a boreal population during the opposite winter season, resulting in unique spatial overlap between Northern and Southern Hemisphere populations. The occurrence of such a northerly wintering area is coincident with the development of an equatorial tongue of cold water in the eastern South Pacific, a pattern that is repeated in the eastern South Atlantic. A survey of location and water temperature at the wintering areas worldwide indicates that they are found in warm waters (21.1–28.3°C), irrespective of latitude. We contend that while availability of suitable reproductive habitat in the wintering areas is important at the fine scale, water temperature influences whale distribution at the basin scale. Calf development in warm water may lead to larger adult size and increased reproductive success, a strategy that supports the energy conservation hypothesis as a reason for migration.
The long annual migrations of baleen whales from productive high-latitude areas used for feeding (feeding areas) to low-latitude oligotrophic areas used for breeding (wintering areas) are well known, but the selective forces driving them remain unclear. Several hypotheses have been proposed that include: a vestigial behaviour from when smaller ocean basins meant closer feeding and wintering areas (Evans 1987); the optimization of energy budgets by wintering in warm waters (Brodie 1975); increased calf (offspring) growth and survivorship in warm protected waters (Norris 1967); and avoidance of killer whale (Orcinus orca) predation at low latitudes (Corkeron & Connor 1999).
For humpback whales (Megaptera novaeangliae), which occur in all major ocean basins, wintering areas are at approximately 20° latitude in both hemispheres (Clapham & Mead 1999), while the feeding areas are found in temperate to polar waters within the same hemisphere. Both Mathews (1937) and Mackintosh (1942), however, reported humpback whale catches near the equator during the austral winter (July–October) off the western coasts of South America and Africa, and suggested that some Southern Hemisphere whales winter in areas north of the equator. Modern research has confirmed this off Ecuador and Colombia (approx. 0–7° N; Flórez-González et al. 1998; Félix & Haase 2001). We report on a wintering area in coastal Central America for humpback whales feeding off the Antarctic Peninsula, resulting in the longest documented mammalian migration. We also examine the water temperature at this wintering area and compare it with all other wintering areas worldwide to investigate the influence of temperature on humpback whale distribution at low latitudes.
2. Material and methods
Data collection consisted of coastal marine mammal surveys conducted yearly off Central America between latitudes 7°46′ N and 11°06′ N during the austral winter from 2001 to 2004 (figure 1; Rasmussen 2006). Additional opportunistic observations made since 1993 were also included. Whenever encountered, humpback whales were identified using photographs of the unique markings on the ventral surface of the tail flukes (Katona & Whitehead 1981). Photographs were also collected off the Antarctic Peninsula between latitudes 61°14′ S and 66°43′ S during the austral summer from 1981 to 2004, as part of the Antarctic humpback whale catalogue (AHWC, n=965; Allen et al. 2006). These two photographic collections were compared to determine individual whales common to both areas.
We reviewed the geographical distribution of current humpback whale wintering areas worldwide based on the published literature (electronic supplementary material). For inclusion as a wintering area, we considered areas where small calves had been observed and where there were other behaviours by adults indicative of breeding, such as competitive groups and singing (Winn & Winn 1978; Tyack & Whitehead 1983; Baker & Herman 1984). Long-term mean sea-surface temperature (SST) for the month of peak whale occurrence in each wintering area was extracted from satellite-derived global monthly climatologies at 4 km grid resolution for the base period 1985–2001. These climatologies are distributed by NOAA's National Oceanographic Data Center (http://www.nodc.noaa.gov/sog/pathfinder4km/).
A total of 207 whales were sighted off Central America during the 2001–2004 austral winter surveys (figure 1). Mother/calf pairs, groups of competing males and singing males were all documented (Rasmussen 2006). Forty-one whales were individually identified photographically; of these, seven were also photographed off the Antarctic Peninsula. Three of these whales were seen within the same year; one off Costa Rica 262 days after it was seen off Antarctica and two (a mother/calf pair) off Antarctica 161 days after being seen off Costa Rica (table 1). The minimum distances between these locations, as determined by great-circle distances between straight segments around land masses, ranged from 8299 to 8461 km.
A total of 24 wintering areas were determined worldwide, all within 30° of the equator (figure 2; electronic supplementary material). A humpback whale population inhabiting the Arabian Sea was excluded from the analysis because it does not appear to undertake seasonal migrations (Mikhalev 1997). Climatological SSTs in all wintering areas were 21.1–28.3°C (electronic supplementary material), with no significant difference between Northern and Southern Hemispheres (t-test on the log-transformed data to homogenize the variance, p-value=0.23). The interquartile SST range for all areas was 23.4–25.7°C, and the average was 24.6°C (±1.9°C s.d.).
The migrations reported here (up to 8461 km minimum travelling distance) are the longest movements documented for any mammal. The longest previously known migration was reported by Stone et al. (1990) for a humpback whale that moved between the Antarctic Peninsula and Colombia (7878 km according to our method; their reported 8334 km is probably an overestimate).
These trans-equatorial migrations between Antarctica and Central America are common for at least part of the humpback whale population in the eastern South Pacific. During this study, whales were regularly sighted off Panama and Costa Rica in the austral winter as far north as 11° N. Group composition (mother/calf pairs) and behaviours (competing males and singing males) were indicative of an area used for calving and mating. This area is probably an extension of the wintering area off Ecuador and Colombia (figure 1). The combined extent of these wintering areas (figure 2) may be a result of whales migrating further north to Central America due to space limitation or other density-dependent process.
Another unique aspect of this area is the spatial overlap with whales from the Northern Hemisphere. The work we have conducted off Central America during the boreal winter season (December–April) since 1996 indicates that this is also a wintering area for eastern North Pacific humpback whales migrating from feeding areas off California (Calambokidis et al. 2000). In fact, anecdotal whale sightings have been reported monthly off Central America (Rasmussen 2006), although the population identity of animals seen during the non-peak occurrence months remains undetermined. Eastern North and South Pacific populations share genetic traits indicating a trans-equatorial exchange, probably off Central America (Medrano-González et al. 2001).
The climatological August SST in the Central American wintering area is 28°C (electronic supplementary material). Temperatures between 24 and 28°C were reported at other humpback whale wintering areas (Dawbin 1966; Herman & Antinoja 1977; Whitehead & Moore 1982), consistent with our global SST analysis (figure 2; electronic supplementary material). Coastal upwelling and cold tongue development during the austral winter result in cool surface waters extending from the South American coast into the eastern equatorial Pacific, such that SSTs greater than 24°C only occur north of the equator (figures 1 and 2). This implies that humpback whales in the eastern South Pacific need to migrate farther north to wintering areas off Ecuador, Colombia and Central America. This pattern is observed again off the western coast of Africa, where an anomalously northerly wintering area for eastern South Atlantic humpback whales (Findlay et al. 1994; Walsh et al. 2000; Van Waerebeek et al. 2001) coincides with the occurrence of coastal upwelling and an equatorial cold tongue (figure 2). Thus, while the availability of suitable reproductive habitat at the wintering areas is important at the fine scale (e.g. Ersts & Rosenbaum 2003; Félix & Haase 2005), water temperature influences their distribution at the basin scale.
Our analysis shows that worldwide humpback whale wintering areas are found in warm coastal waters irrespective of latitude. In the ongoing debate on the reasons for migration, this result supports previous ideas linking temperature at the wintering areas to energetic strategies without the need to invoke killer whale avoidance. Clapham (2001) suggests that, as in some terrestrial mammals, energy conserved during offspring development can be devoted to growth, leading to larger size and increased reproductive success in adulthood. Thus, the temperature regime at the wintering areas, regardless of distance to the feeding areas, probably constitutes a major selective force driving humpback whale migration.
Funding was provided by the Homeland Foundation and the National Fish and Wildlife Foundation. Photographs were contributed by K. Robertson Chater and I. McWhirter. DMP was supported by award no. N00014-05-1-0045 from the US Office of Naval Research, National Oceanographic Partnership Program. Helpful comments were provided by P. J. Clapham, J. T. Harvey, R. L. Brownell, S. J. Bograd and an anonymous reviewer.