Marine Biology (1999) 134: 387±395
Ó Springer-Verlag 1999
D. J. McCaerty á I. L. Boyd T. R. Walker á R. I. Taylor
Can marine mammals be used to monitor oceanographic conditions?
Received: 16 June 1998 / Accepted: 13 February 1999
Abstract The breeding performance of higher predators has often been used to monitor ¯uctuations in the abundance of important prey stocks in marine ecosystems. The development of electronic data-loggers in recent years has also provided the opportunity of using wide-ranging marine animals to measure physical oceanographic conditions. In this study, time±depth recorders (TDRs) programmed to record temperature were deployed on female Antarctic fur seals (Arctocephalus gazella) at Bird Island, South Georgia (54°00¢S; 38°02¢W) during the breeding seasons 1994 to 1998. Temperature sensors had relatively slow response times, and thermal radiation errors occurred during the day when seals spent a large proportion of their time at the surface. Nevertheless, measurements provided temperature±depth pro®les which were typical of the vertical strati®cation of the ocean. During the early stages of a foraging trip temperature increased, suggesting that fur seals travelled northwards from South Georgia towards the warmer waters of the Polar Front. In addition, higher temperatures were recorded by females that remained at sea for longer, implying that these individuals also travelled further. Mean sea-surface temperature (SST) increased from 1 to 4 °C from December to March and agreed with SSTs from ship, buoy and satellite. Future studies on marine mammals which combine satellite tracking with oceanographic measurements are likely to provide valuable information on biophysical aspects of the ocean. Communicated by J.P. Thorpe, Port Erin D.J. McCaerty (&)1 á I.L. Boyd á T.R. Walker á R.I. Taylor British Antarctic Survey, NERC, High Cross, Madingley Road, Cambridge CB3 OET, England 1
Present address: 33 Edderston Road, Peebles EH45 9DT, Scotland e-mail: d.mcca
[email protected]
Introduction The breeding performance of higher predators has been used to monitor important prey stocks in marine ecosystems (Montevecchi 1993; Monaghan 1996). Data from seabirds and seals can provide an indication of local prey availability and signal major ¯uctuations in the abundance and distribution of prey resulting from large-scale oceanographic phenomena (Croxall et al. 1988; Trillmich et al. 1991; York 1995; Montevecchi and Myers 1996). Breeding performance integrates measurements of predator and prey behaviour over relatively long periods of time, and both predator and prey may respond to a number of environmental factors, many of which are unknown or dicult to measure. The development of electronic data-loggers has provided the opportunity of using wide-ranging animals to measure oceanographic conditions that are linked to or responsible for changes in prey availability. With marine mammals such as southern elephant seals (Mirounga leonina), time±depth±temperature recorders have provided data on the temperature strati®cation in the Southern Ocean, and sea-surface temperatures (SSTs) that are characteristic of dierent locations within the ocean (Boyd and Arnbom 1991; Hindell et al. 1991). Similarly, temperature measurements have identi®ed water masses of the North Paci®c used by northern elephant seals (M. angustirostris) and foraging areas of northern fur seals (Callorhinus ursinus) (Hakoyama et al. 1994; Gentry et al. 1998). Studies on seabirds have also provided temperature data over different spatial scales. Data from recorders on the wideranging wandering albatross Diomedea exulans were used to map SST patterns over a large area of the Southern Ocean (Weimerskirch et al. 1995). In contrast, time±depth±temperature recorders carried by penguins have been used to relate temperature characteristics of local foraging areas to prey abundance (Wilson et al. 1994).
388
The aim of this study was to determine the feasibility of using Antarctic fur seals (Arctocephalus gazella) to measure the temperature of the ocean in areas surrounding their breeding sites at South Georgia. Longterm population studies of predators at South Georgia have shown that major reductions in breeding performance occur in years of low krill (Euphausia superba) abundance (Croxall et al. 1988; Lunn et al. 1993; Boyd et al. 1994; McCaerty et al. 1998). Although the factors controlling krill stocks at South Georgia are not fully understood, it is likely that large-scale atmospheric and oceanographic processes are involved (Priddle et al. 1988; Fedoulov et al. 1996). South Georgia lies in a hydrographically complex region, and considerable oceanographic research has been undertaken by shipbased surveys and remote-sensing (Stein 1979; Whitehouse et al. 1996a; Trathan et al. 1997). However, surveys may be constrained by the expense of operating ships in this remote location, and satellite measurement of SSTs are also hampered by extensive cloud cover (Trathan et al. 1997). Deployment of temperature sensors on fur seals may therefore be a method of making detailed oceanographic measurements in areas of the Southern Ocean where such data are lacking.
Materials and methods Deployments on fur seals Time-depth recorders (TDR Mk5, Wildlife Computers Inc., Redmond, USA) were deployed on female Antarctic fur seals (Arctocephalus gazella) at Bird Island, South Georgia (54°00¢S; 38°02¢W) during the austral summers 1993/1994 to 1997/1998 (hereafter referred to as 1994 to 1998). The TDRs consisted of a microprocessor with depth (1 m), temperature (0.1 C°) and light sensors (uncalibrated units) encased in an epoxy block (1.25 ´ 3.75 ´ 6.25 cm, 50 g). Fur seals were selected opportunistically from the breeding beach using standard capture methods (Gentry and Holt 1982). The TDRs either were attached with plastic cable ties to a nylon webbing strap (10 ´ 2.5 cm) that had been glued to the mid-dorsal fur using quick-set epoxy (RS Ltd, Corby, Northants, UK) or, in February and March 1994, were glued directly onto the fur. Females were also ®tted with a 40 g 165 MHz radio-transmitter (Sirtrack Ltd, Havelock North, New Zealand) to relocate individuals when they returned ashore. Several sets of deployments were carried out during the study period. TDRs were deployed on 37 individuals from 1994 to 1996 and were programmed to record depth at 10 s and temperature at 300 s intervals. The TDRs only recorded when the salt-switch was immersed in seawater. To examine the response time of TDR temperature-sensors, three individuals in 1994 and a further four individuals in 1997 were ®tted with TDRs that were programmed to record depth and temperature at 5 s intervals. Finally, the eect of attachment method was determined on four individuals from 20 to 31 March 1998. Two TDRs (10 s depth and 10 s temperature) were attached to the same individual, by glueing one TDR directly onto the fur and mounting the other on a strap (see foregoing paragraph). Prior to deployment, the TDRs were placed in a container of seawater to check calibration (see following subsection). Calibration The TDR thermistors were calibrated in seawater at 0.5 to 9.0 °C against a mercury thermometer with a precision of
0.1 C°. The eect of solar radiation on the temperature recorded by the TDRs was examined from 22 to 23 November 1994. Two Mk5 TDRs were programmed to record temperature and light at 60 s intervals, and were ®xed to a wooden platform that was buoyed 200 m from the shore. The platform held the devices '5 to 10 cm below the sea surface. One TDR was wrapped in aluminium foil to re¯ect solar radiation. The TDR light sensor was calibrated against a monochromatic ®bre-optic lamp (Model KL1500, Schott Fibre Optics Ltd, Doncaster, UK). Light level was varied from 14 to 10250 lx using neutral-density optical ®lters (HV Scan Ltd, Solihil, UK). Illumination (0.1 lux, spectral sensitivity 400 to 700 nm) was recorded on a digital luxmeter (Model TES132, TES Electrical Electronic Corporation, Taipei, Taiwan).
Data analysis On recovery of TDRs, data were transferred to a computer and decoded, using custom-written software, to produce an ASCII list of depth and temperature readings and a summary of individual dives (Boyd et al. 1997). The response time of temperature sensors was examined using TDR depth and temperature data recorded at 5 s sampling intervals. To ensure that the temperature during each dive was not in¯uenced by earlier dives and the sensor was given sucient time to equilibrate after each dive, dives were selected according to the criteria that no diving occurred 5 min or more before the start and following the end of each dive. Surface temperature was de®ned as the temperature recorded immediately prior to the start of the dive. The response time was de®ned as the time taken following the end of the dive for the temperature to return to within 0.2 C° of the surface reading (i.e. maximum calibrated resolution ´ 2). Because of the response time of temperature sensors and the diel temperature pattern recorded by the fur seals, the mean SST during each foraging trip was determined from night-time records when the depth did not exceed 20 m in the previous 5 min. Night time was determined as the time between dawn and dusk, whereby these were de®ned as nautical twilight (i.e. when the sun was 9° below the horizon). Independent measurements of SST from ship, buoy and satellite data were obtained from the United States National Centre for Atmospheric Research (NCAR). Monthly 1° latitude ´ 1° longitude optimum-interpolation grids were loaded into a marine ``Geographical Information System'' (Trathan et al. 1993). Data from December 1993 to March 1996 were obtained for three grids to the NW of South Georgia (centred at 53.5°S, 38.5°W; 53.5°S, 39.5°W; and 53.5°S, 40.5°W). All data manipulation and statistical analysis were made with SAS Version 6.11 (SAS Institute Inc., Cary, USA).
Results Calibration There was a straight-line relationship between the TDR temperature output and the thermometer reading over the range in temperatures (R2 = 0.92±0.99, n = 12, p < 0.0001 in all cases), and individual calibration coecients and intercepts were not dierent from one and zero respectively (p > 0.09 in all cases). The manufacturer's temperature calibration was therefore used for all measurements. There was an exponential relationship between illumination (lux) and sensor output [ln(lux) = 0.13 output ) 12.07, R2 = 0.99, F1,17 = 9663, p < 0.0001].
389
Sensor response TDR records showed that the temperature reading lagged behind the depth reading (Fig. 1). The response time of TDRs, de®ned as the time taken for the TDR to record the surface temperature following a dive, averaged 3.9 s (SE = 0.87 s, range 0 to 150 s) during 431 dives with a mean dive depth of 14.9 m (SE = 1.18 m, range 2 to 131 m) and mean dive duration of 38 s (SE = 2.3 s, range 10 to 205 s). Although the average response time was 30 m) varied with hour of day (p < 0.0001 in all cases: Fig. 6). Fur seals did not spend an equal amount of time at the surface during each hour of the day (chi-square test: v2 = 268394, p < 0.0001, n = 7929267: surface readings), but they did spend a greater proportion of their time at the surface during daylight hours (Fig. 6). However, fur seals dived to greater depths during daylight hours (Kruskal±Wallis test: v2 = 32001, df = 23, p < 0.0001; number of dives = 576 534; Fig. 6). Temperature changes during foraging trips Changes in temperature were recorded during foraging trips (Fig. 7). A repeated-measures ANOVA was used to examine changes in temperature throughout the foraging trip (stage) by determining the eect of individual (tag) and year, as well as the interaction between stage and individual (stage ´ tag) and year (stage ´ year), (Table 2). In December and January, there were signi®cant dierences in temperature across years and between dierent stages of the foraging trip. In December
Fig. 7 Relationship between TDR temperature, mean dive depth, percentage of time submerged and proportion of foraging trip made by Antarctic fur seals, Arctocephalus gazella. SE bars are shown
392 Table 2 Results from repeated-measures ANOVAs on TDR temperatures recorded in each month and on foraging behaviour of fur seals, Arctocephalus gazella (% time spent submerged and dive depth), during foraging trips in 1994 to 1996. Table shows probability values calculated for between-subject eects (female tag and year) and within-subject eects (stage of foraging trip and interactions) Between-subject eects Within-subject eects tag Temperature Dec. Jan. Feb./Mar.
year
0.02 0.002 0.12 0.04 0.0004 0.77
stage
stage ´ stage ´ tag year
0.0001 0.0001 0.0001
0.27 0.0009 0.0001
0.004 0.01 0.0001
0.57 0.03
0.07 0.02
Foraging behaviour % time spent submerged 0.0001 0.0001 0.0001 Dive depth (m) 0.14 0.0007 0.0001
Table 3 Monthly mean sea-surface temperature (SST) at South Georgia, taken from ship, buoy and satellite data (see ``Materials and methods ± Data analysis'') and SSTs measured by TDRs on fur seals, Arctocephalus gazella, during 1994 to 1996 breeding seasons (Dec refers to December in previous calendar year) Month
SST (oC) 1994
1995
Measured by ship, buoy and satellite Dec. 2.2 2.1 (1.8±2.6, 3) (1.8±2.4, Jan. 3.4 3.2 (3.2±3.6, 3) (2.7±3.8, Feb. 4.1 4.3 (3.9±4.3, 3) (3.5±5.1, Mar. 3.4 4.5 (3.2±3.7, 3) (3.8±5.1, Measured by TDRs Dec. 2.4 (2.3±2.5, 3) Jan. 2.5 (2.0±3.5, 2) Feb. 3.1 (2.6±3.6, 10)
1996
3) 3) 3) 3)
1.9 (1.7±2.0, 36) 2.6 (2.5±2.7, 40) 3.8 (3.6±3.8, 36)
1.7 (1.3±2.1, 3.5 (3.1±3.9, 4.4 (4.1±4.7, 4.8 (4.5±5.0,
3) 3)
Fig. 8 Relationship between sea-surface temperature (SST) and date (Days since 1 December) recorded by TDRs on Antarctic fur seals, Arctocephalus gazella (s), and mean monthly SST at South Georgia (h) recorded by ship, buoy and satellite. Quadratic regression line was ®tted to TDR data by least-squares regression
SST was correlated with foraging-trip duration (Pearson correlation: r = 0.40, p < 0.0001), however, foraging-trip duration was also highly correlated with date (days since 1 December r = 0.42, p < 0.0001). The relationship between mean SST and foraging-trip duration was therefore determined by examining the residuals from the quadratic regression of SST with date (F2,272 = 644.2, p < 0.0001: Fig. 8). Higher SSTs were associated with longer foraging trips, as the residuals were correlated with foraging-trip duration (r = 0.18, p < 0.0001).
3) 3)
1.8 (1.7±1.9, 43) 2.9 (2.8±2.9, 57) 3.7 (3.6±3.8, 47)
there was a signi®cant dierence across years (F2,266 = 6.8, p = 0.045), the interaction between month and season was also signi®cant (F4,265 = 3.0, p = 0.02). Examination of 95% con®dence intervals indicated that only in December of the 1994 season were SSTs signi®cantly greater than in the subsequent two years (Table 3). The mean SST measured by ship, buoy and satellite data also showed a seasonal increase (Fig. 8). ANOVA revealed signi®cant dierences in SST between months (F3,24 = 344.5, p < 0.0001) and across years (F2,24 = 10.8, p < 0.0005), but the interaction between month and season was also signi®cant (F6,24 = 16.3, p < 0.0001). The only dierence in SST between years occurred in March, when SST in 1994 was cooler than in 1995 and 1996 (Table 3).
Discussion Temperature measurement Over many years, TDRs have provided detailed information on the diving behaviour of a wide range of pinnipeds, including the Antarctic fur seal, Arctocephalus gazella (reviews by Boyd and Croxall 1996; Le Boeuf et al. 1996). However, the development of smaller data-loggers with increased ``memory'' capacity has extended the capability of these devices to record temperature, swim-speed and light (e.g. Ponganis et al. 1990; DeLong et al. 1992; Wilson et al. 1992; Boyd et al. 1995). If these devices are to be used to record a range of environmental variables, then the sensors must be accurate and, in the case of temperature measurements, where temperature changes with depth, the rate of response of the sensor must be faster than the vertical swim speed of the animal. This study has shown that Mk5 TDRs on fur seals were unable to respond fast enough to changes in temperature during deep dives. Although the thermistor on the Mk5 TDR is surrounded by a small amount of heat-sink compound, the entire device is encased in epoxy which has a relatively low thermal conductivity. Previous studies on Antarctic fur
393
seals and Southern elephant seals (Boyd and Arnbom 1991; Boyd and Croxall 1992) used MK3 TDRs which were housed in metal, and therefore sensors responded faster. However, close examination of TDR traces shows that even these devices showed temperature lags following deep dives. TDRs recorded a diurnal increase in temperature which was related to the diel pattern of fur seal diving. This increase was partially explained by an increase in sea temperature as a result of solar heating at the surface [which is likely to be less exaggerated or absent where greater mixing occurs further oshore (Mann and Lazier 1991)], but was also due to radiative heating of the instrument surface, a well-known phenomenon with temperature sensors (McIlven 1986). Errors due to the response time and radiative heating of sensors may be relatively minor for species such as elephant seals, which remain underwater for long periods of time and where the TDR may rarely break the surface. However, for species such as fur seals and penguins which make dives of relatively short duration, errors are likely to be greater. TDRs glued directly onto the fur recorded greater temperatures than TDRs mounted on a strap. Temperatures recorded by TDRs which were glued were strongly in¯uenced by the pattern of diving, indicating that these instruments were also recording the heat ¯ux from the animal. Although this was accounted for, there was a high potential for error in the 1994 deployments using TDRs mounted on the fur. Recently developed TDRs with small, externally-mounted thermistors will not be in¯uenced by the seal's own body heat and are likely to improve sensor response-time and reduce radiation errors associated with these devices. Despite the problems associated with TDR design, deployments on fur seals provided temperature pro®les which were characteristic of this region of the Southern Ocean (Stein 1979; Whitehouse et al. 1996b; Pakhomov et al. 1997); i.e. a mixed layer extending to 50 m and a thermocline of 1 to 2 C° down to 100 m. Although fur seals were recorded as having dived to >180 m, only 0.4% of dives were >100 m, and therefore provided limited data on the overall vertical temperature distribution of the ocean. In addition, the overall accuracy of the TDRs in recording temperature as a function of depth and dierences in diving behaviour between years means that fur seals are unlikely to provide detailed measurements of vertical temperature strati®cation. Nevertheless, biological productivity is greatest within the mixed layer (Mann and Lazier 1991), and therefore fur seals may provide measurements on seasonal and annual variation in this region of the water column. Temperature changes during foraging trips During January to March 1995 and 1996, there was a systematic increase in temperature recorded during foraging trips. This suggested that fur seals travelled northward towards the warmer waters of the Antarctic
polar front before returning to Bird Island. Satellitetracking of a dierent group of females in 1995 and 1996 did indeed show that fur seals travelled as far as 350 km to the NW of Bird Island and that diving behaviour was associated with the edge of the continental shelf of South Georgia (Boyd et al. 1998). Controlling for seasonal changes in foraging-trip duration and temperature, longer foraging trips were also correlated with higher SSTs. Females that stayed at sea for longer periods were either foraging in areas of warmer water or were travelling further. Satellite tracking has con®rmed that fur seals on longer foraging trips also travel further from Bird Island (Boyd unpublished data). In addition, during January 1994, TDRs ®tted to two females that abandoned their pups for 20 to 32 d recorded temperatures of 7 to 8 °C after 14 d at sea (at which point the data-logger ``memory'' was full) (McCaerty et al. unpublished data). These females either moved into a warm-water protrusion which was recorded on satellite images SW of Shag Rocks, or they moved north of the Antarctic polar front (Whitehouse et al. 1996b; Trathan et al. 1997). The only previous record of an Antarctic fur seal crossing the polar front was of a pup tagged on Bird Island that was subsequently found dead in Tierra del Fuego (Payne 1979). Intra and inter-annual variation SSTs recorded by TDRs increased from 1 °C at the start of December to 4 °C by the beginning of March, and were similar to seasonal temperature changes in the vicinity of South Georgia recorded by ship and satellite (Deacon 1977; Whitehouse et al. 1996a). In comparison to 1995 and 1996, there were few measurements in 1994 and it was therefore dicult to determine if the higher SSTs recorded by the TDRs in December were simply due to the fact that females travelled further from Bird Island (see McCaerty et al. 1998). Although there was no dierence in the temperature across years when averaged across all foraging trips (Table 1), the repeatedmeasures ANOVA which accounted for temperature recorded by individual seals suggested that even at the start of foraging trips, temperatures in December and January 1994 were greater than in subsequent years (Fig. 7). A previous study (McCaerty et al. 1998) found that SSTs recorded by fur seals in 1994 were greater in all months. However, the 1998 study included a greater number of measurements from dierent types of instruments and failed to account for dierences in attachment methods. Ship, buoy and satellite data showed that there were no large-scale dierences in SST between 1994 and subsequent years (Table 3). However, incursions of warm water were recorded in 1994 (Whitehouse et al. 1996b; Trathan et al. 1997), and therefore it is possible that TDRs recorded localised areas of warm water not detected using large-scale interpolated measurements of SST. This observation of elevated SSTs in 1994 is
394
interesting, as ship-based hydro-acoustic surveys also recorded extremely low densities of krill in the vicinity of South Georgia (Brierley and Watkins 1996; Brierley et al. 1997) and a lower proportion of fur seal scats contained krill (McCaerty et al. 1998). Antarctic krill is the predominant prey item in the diet of a range of predators, including fur seals (Reid and Arnould 1996). The consequence of this was that pup survival, pup growth and weaning mass were amongst the lowest on Bird Island in 16 yr (McCaerty et al. 1998). Conclusions If wide-ranging marine animals are to be used as platforms for oceanographic measurement (Ancel et al. 1992; Wilson et al. 1994; Weimerskirch et al. 1995), it is essential that we determine the accuracy of instruments and understand how measurements are aected by an animal's behaviour. In this study, temperature sensors on TDRs had relatively slow response times and were prone to radiation errors when animals spent long periods at the surface. Nevertheless, deployments on Antarctic fur seals were able to obtain data on the vertical and horizontal temperature distribution in areas surrounding their breeding sites. Fur seals have relatively shallow dive depths and limited foraging ranges when breeding. However, Antarctic fur seals are krill specialists and may therefore make physical measurements at the temporal and spatial scales necessary to understand key aspects of krill distribution and abundance. Provided that improvements are made in sensor design and instruments are rigorously tested, future satellite telemetry-studies on wide-ranging marine animals are likely to provide valuable physical oceanographic data. However, measurements from marine mammals are biased towards favoured areas of the ocean due to the non-random foraging patterns of these animals. Future studies must aim to examine how measurements from animals apply to wider spatial scales, if they are to be used to help understand oceanographic variability and its eects on important prey stocks. Acknowledgements Thanks to all British Antarctic Survey sta at Bird Island for assistance throughout the study, in particular to I. Staniland and K. Barlow for TDR deployments in 1998. We are grateful to P.N. Trathan for providing SST data and for reviewing an earlier draft of this manuscript.
References Ancel A, Kooyman GL, Ponganis PJ, Gendner J-P, Lignon J, Mestre X, Huin N, Thorson PH, Robisson P, LeMaho Y (1992) Foraging behaviour of emperor penguins as a resource detector in winter and summer. Nature, Lond 360: 336±339 Boyd IL, Arnbom T (1991) Diving behaviour in relation to water temperature in the southern elephant seal: foraging implications. Polar Biol 11: 259±266 Boyd IL, Croxall JP (1992) Diving behaviour of lactating Antarctic fur seals. Can J Zool 70: 919±928
Boyd IL, Croxall JP (1996) Dive durations in pinnipeds and seabirds. Can J Zool 74: 1696±1705 Boyd IL, Croxall JP, Lunn NJ, Reid K (1994) Foraging behaviour of Antarctic fur seals during periods of contrasting prey abundance. J Anim Ecol 63: 703±713 Boyd IL, McCaerty DJ, Reid K, Taylor R, Walker TR (1998) Dispersal of male and female Antarctic fur seals. Can J Fish aquat Sciences 55: 845±852 Boyd IL, McCaerty DJ, Walker TR (1997) Variation in foraging eort by lactating Antarctic fur seals: response to simulated increased foraging costs. Behavl Ecol Sociobiol 40: 135±144 Boyd IL, Reid K, Bevan RM (1995) Swimming speed and allocation of time during the dive cycle in Antarctic fur seals. Anim Behav 50: 769±784 Brierley AS, Watkins JL (1996) Acoustic targets at South Georgia and the South Orkney Islands during a season of krill scarcity. Mar Ecol Prog Ser 138: 51±61 Brierley AS, Watkins JL, Murray AWA (1997) Interannual variability in krill abundance at South Georgia. Mar Ecol Prog Ser 150: 87±98 Croxall JP, McCann TS, Prince PA, Rothery P (1988) Reproductive performance of seabirds and seals at South Georgia and Signy Island, South Orkney Islands, 1976±1987: implications for Southern Ocean monitoring studies. In: Sahrhage D (ed) Antarctic ocean and resource variability. Springer-Verlag, Berlin, pp 261±285 Deacon GER (1977) Seasonal and annual variations in water temperature and salinity near South Georgia 1925±1937. Report No. 49. Institute of Oceanographic Sciences, Godalming DeLong RL, Stewart BS, Hill RD (1992) Documenting migrations of Northern elephant seals using day length. Mar Mammal Sci 8: 155±159 Fedoulov PP, Murphy E, Shulgovsky KE (1996) Environment± krill relations in South Georgia marine ecosystem. CCAMLR Sci 3: 13±30 Gentry RL, Holt JR (1982) Equipment and techniques for handling northern fur seals. NOAA natn mar Fish Serv tech Rep US Dep Commerce 758: 1±15 Gentry RL, Pierson MO, Vladimirov VA (1998) Female foraging behaviour: eects of continental shelf width. In: Gentry RG (ed) Behavior and ecology of the Northern fur seal. Princeton University Press, Princeton, pp 260±276 Hakoyama H, Le Boeuf BJ, Naito Y, Sakamoto W (1994) Diving behavior in relation to ambient water temperature in northern elephant seals. Can J Zool 72: 643±651 Hindell MA, Burton HR, Slip DJ (1991) Foraging areas of southern elephant seals, Mirounga leonina, as inferred from water temperature data. Aust J mar Freshwat Res 42: 115±128 Le Boeuf BJ, Morris PA, Blackwell SB (1996) Diving behaviour of juvenile northern elephant seals. Can J Zool 74: 1632±1644 Lunn NJ, Boyd IL, Barton T, Croxall JP (1993) Factors aecting the growth rate and mass at weaning of Antarctic fur seals at Bird Island, South Georgia. J Mammal 74: 908±919 Mann KH, Lazier JRN (1991) Dynamics of marine ecosystems. Biological±physical interactions in the oceans. Blackwell Scienti®c, Oxford McCaerty DJ, Boyd IL, Walker TR, Taylor RI (1998) Foraging responses of Antarctic fur seals to changes in the marine environment. Mar Ecol Prog Ser 166: 285±299 McIlven JFR (1986) Basic meteorology: a physical outline. Van Nostrand Reinhold, Wokingham Monaghan P (1996) Relevance of the behaviour of seabirds to the conservation of marine environments. Oikos 77: 227±237 Montevecchi WA (1993) Birds as indicators of change in marine prey stocks. In: Furness RW, Greenwood JJD (eds) Birds as monitors of environmental change. Chapman & Hall, London, pp 217±266 Montevecchi WA, Myers RA (1996) Dietary changes of seabirds indicate shifts in pelagic food webs. Sarsia 80: 313±322 Pakhomov EA, Verheye HM, Atkinson A, Laubscher RK, Taunton-Clark J (1997) Structure and grazing impact of the meso-
395 zooplankton community during late summer 1994 near South Georgia, Antarctica. Polar Biol 18: 180±192 Payne MR (1979) Fur seals Arctocephalus tropicalis and A. gazella crossing the Antarctic Convergence at South Georgia. Mammalia 43: 93±98 Ponganis PJ, Ponganis EP, Ponganis KV, Kooyman GL, Gentry RL, Trillmich F (1990) Swimming velocities in otariids. Can J Zool 68: 2105±2112 Priddle J, Croxall JP, Everson I, Heywood RB, Murphy EJ, Prince PA, Sear CB (1988) Long-term ¯uctuations in mesoscale abundance of krill ± a discussion of possible causes. In: Sahrhage D (ed) Antarctic ocean and resources variability. Springer-Verlag, Berlin, pp 169±182 Reid K, Arnould JPY (1996) The diet of Antarctic fur seals Arctocephalus gazella during the breeding season at South Georgia. Polar Biol 16: 105±114 Stein M (1979) Temperature salinity and dissolved silicate in the Scotia Sea during austral summer 1975/6. Meeresforsch Rep mar Res (Sdbd) 27: 57±74 (Ber dt wiss Kommn Meeresforsch) Trathan PN, Brandon MA, Murphy EJ (1997) Characterization of the Antarctic Polar Frontal Zone to the north of South Georgia in summer 1994. J geophys Res 102: 10483±10497 Trathan PN, Murphy EJ, Symon CJ, Rodhouse PG (1993) Ecological and oceanographic relationships in the Southern Ocean. GIS Europe 2: 34±36 Trillmich F, Ono KA, Costa DP, DeLong RL, Feldkamp SD, Francis JM, Gentry RL, Heath CB, LeBoeuf BJ, Majiluf P, York AE (1991) The eects of El NinÄo on pinniped populations
in the Eastern Paci®c. In: Trillmich F, Ono KA (eds) Pinnipeds and El NinÄo. Responses to environmental stress. SpringerVerlag, Berlin, pp 247±270 (Ecol Stud 88) Weimerskirch H, Wilson RP, Guinet C, Koudil M (1995) Use of seabirds to monitor sea-surface temperatures and to validate satellite remote-sensing measurements in the Southern Ocean. Mar Ecol Prog Ser 126: 299±303 Whitehouse MJ, Priddle J, Symon C (1996a) Seasonal and annual change in seawater temperature, salinity, nutrient and chlorophyll a distributions around South Georgia, South Atlantic. Deep-Sea Res I 43: 425±443 Whitehouse MJ, Priddle J, Trathan PN, Brandon MA (1996b) Substantial open-ocean phytoplankton blooms to the north of South Georgia, South Atlantic, during summer 1994. Mar Ecol Prog Ser 140: 187±197 Wilson RP, Culik BM, Bannasch R, Driesen HH (1992) Monitoring penguins at sea using data loggers. In: Paolo M, Sando F, Christina G, Remo B (eds) Proceedings of XII International Symposium on Biotelemetry. Editrice Universitaria Litogra®ca Felici, Pisa, pp 205±214 Wilson RP, Culik BM, Bannasch R, Lage J (1994) Monitoring Antarctic environmental variables using penguins. Mar Ecol Prog Ser 106: 199±202 York AE (1995) The relationship of several environmental indices to the survival of juvenile male northern fur seals (Callorhinus ursinus) from the Pribilof Islands. Can Spec Publ Fish aquat Sciences 121: 317±327
