By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Due to essential maintenance the subscribe/renew pages will be unavailable on Wednesday 26 October between 02:00- 08:00 BST/ 09:00 – 15:00 SGT/ 21:00- 03:00 EDT. Apologies for the inconvenience.
‡Correspondence and present address: A. Kitaysky, Department of Zoology, University of Washington, Box 351800, Seattle WA 98195, USA. E-mail: email@example.com
1. Eight years of observations on seabird reproductive success and oceanographic change in Tauyskaya Bay (Okhotsk Sea, north-western Pacific) were used to evaluate the hypothesis that interannual climate change causes opposite trends in reproductive performances of planktivorous auklets (Aethia cristatella and Cyclorhinchus psittacula) vs. piscivorous puffins (Lunda cirrhata and Fratercula corniculata).
2. The climate change was assessed by examining changes in sea-surface temperature (SST), time of permanent ice disappearance (ID), wind (WV) and current vectors (CV). Changes in the distribution of zooplankton biomass in the study region were used to assess changes in prey communities. Bird reproductive success was determined as the number of chicks fledged per nest occupied.
3. There were two distinct sets of oceanographic conditions in the study region, as reflected in the SST, ID, WV and CV. Strong northerly winds in the spring produced a late ice disappearance in the study region, whereas easterly winds determined an early ice disappearance. The patterns in ice disappearance were significantly correlated with SST anomalies during the summer. A negative SST anomaly (– 1·2 °C) defined a ‘cold’ regime, whereas a positive SST anomaly ( + 1·2 °C) defined a ‘warm’ regime.
4. Reproductive success of planktivorous auklets was negatively correlated with the SST in the western part of Tauyskaya Bay, whereas reproductive success of piscivorous puffins was positively correlated with the SST. The ‘cold’ season in 1988 was characterized by a strong in-flow of water masses into the bay area. The ‘warm’ season in 1989 was characterized by well-mixed warm water inside the bay that were separated from colder water masses outside the bay. Macro-zooplankton, which were the main prey of planktivorous auklets, were more abundant during the ‘cold’ regime of the ecosystem. Meso-zooplankton, a potential prey of juvenile pelagic fish, were more abundant during the ‘warm’ regime of the ecosystem.
5. Interannual oceanographic change probably impacts alcid reproductive performances by affecting food accessibility to planktivorous auklets and piscivorous puffins in opposite ways.
Changes in marine climates may favour the reproductive success of one group of seabird species over another through fluctuations in the availability of prey. In the central and south-eastern Bering Sea, the population dynamics of birds foraging on zooplankton and fish have shown opposite trends during the last several decades ( Springer 1992). It is likely that since the 1960s, fish-eating (piscivorous) birds have decreased in numbers, whereas zooplankton-eating (planktivorous) birds have increased ( Piatt, Roberts & Hatch 1990; Konyoukhov 1991). These changes in abundance coincided with a shift in the sea-surface temperature that decreased from the mid-1950s to the late 1970s ( Royer 1993). It is not well understood why planktivorous and piscivorous seabirds have had opposite responses to the same oceanographic change.
Several hypotheses have been proposed to explain the breeding failures of piscivorous seabirds in the Bering Sea (e.g. Straty & Haight 1979; Springer 1992; Decker, Hunt & Byrd 1995; Hunt et al. 1996a; Hunt, Decker & Kitaysky 1996b). ‘Bottom-up’ hypotheses predict that the observed variation in seabird reproduction may be controlled by variation in the distribution and abundance of small pelagic fish following changes in the productivity of zooplankton communities ( Springer 1992). There are some data that provide evidence that meso-zooplankton biomass is higher during ‘warm’ compared to ‘cold’ regimes in the middle domain of the Bering Sea ( Walsh & McRoy 1986). This may increase the availability of food to juvenile fish and facilitate their better survival and growth during ‘warm’ regimes ( Walsh & McRoy 1986), thereby increasing the availability of small pelagic fish to piscivorous foraging seabirds. Decker et al. (1995) and Hunt et al. (1996a) failed to find a correlation between piscivorous bird breeding performances in the south-eastern Bering Sea and the abundance of a particular prey species in their diets. Their results suggest that seabird reproductive performances reflect a major shift of the Bering Sea ecosystem in response to a climate change rather than the abundance of particular prey. However, underlying mechanisms of the presumed ecological shift have yet to be identified.
Straty & Haight (1979) and Springer (1992) suggested an alternative hypothesis of ‘top-down’ control of seabird populations. These authors assumed that bird reproductive failure may have been caused by the depletion of the piscivorous birds’ prey due to increased predation by adult walleye pollock, Theragra chalcogramma Pallas. According to Springer (1992), a reduction in the abundance of small pelagic fish could result in an increase in foraging effort by fish-eating birds and consequently a decrease in their fitness. The consumption of pelagic fish by adult pollock could also result in an increase in zooplankton abundance. An increased abundance of zooplankton could enhance reproductive performances of planktivorous seabirds. Although the hypothesis that seabird population trends reflect pollock control of forage fish abundance is logical and attractive, there is insufficient evidence to test it ( Decker et al. 1995).
In this paper we present results of a meso-scale comparative study of sympatric and closely related planktivorous and piscivorous seabirds to investigate the link between ecosystem change and the reproductive ecology of seabirds. Planktivorous auklets forage mostly on macro-zooplankton ( Kitaysky 1996; Hunt et al. 1998). Piscivorous puffins prey on small forage fish that in turn prey mostly on meso-zooplankton ( Huse & Toresen 1996; Brodeur 1998). We hypothesize that oceanographic changes result in two ‘regimes’, one of which enhances stocks of macro-zooplankton and the other pelagic fish, and that these changes are reflected in opposite reproductive trends of planktivorous and piscivorous alcids.
The goals of our study are: (1) to test whether planktivorous and piscivorous alcids differ in their reproductive response to the same suit of oceanographic changes, characterized by warm and cold temperatures, and then (2) to examine mechanisms whereby changes between ‘cold’ and ‘warm’ regimes might influence the distribution and abundance of macro-zooplankton or meso-zooplankton and their fish predators. The objectives of the comparison between ‘cold’ and ‘warm’ regimes are: (1) to compare distribution of the sea-surface temperature (SST) across the entire bay area; (2) to assess the factors that might affect the SST distribution; and (3) to relate changes in physical oceanography to changes in the structure of zooplankton communities. If a relationship between bird reproductive performances and oceanographic change is found, it would support the hypothesis that seabird breeding performances indicate a shift in the marine ecosystem from one biological state to another ( Hunt et al. 1996a). Furthermore, variation in zooplankton communities in relation to that shift could provide a clue to the linkage between temperature changes and variation in seabird reproduction.
Materials and methods
Data were collected from 1987 to 1994 in the northern part of Okhotsk Sea (north-western Pacific [Fig. 1]). Within the study area, which is the shelf and adjacent Tauyskaya Bay waters, the following oceanographic processes are relevant to this study.
The Kuril warm current (originating in the Pacific Ocean) meets the cold Jamskoe current (originating in the northern Okhotsk Sea) at approximately 57°30′ N latitude. Under the influence of Jamskoe’s cold waters and the bottom topography, the Kuril current splits into two main branches ( TINRO 1988). After splitting, the first branch continues flowing north-easterly, whereas the other turns north-west. This second branch interacts with the Jamskoe current to form an oceanographic front (Kuril–Jamskoe oceanographic front, KJF) from the mouth of Shelikhov Bay to the western side of Tauyskaya Bay. The position of the western border of this front fluctuates yearly, periodically reaching the vicinity of Talan Island ( Cherniavski 1985).
The KJF is characterized by steep horizontal sea-surface gradients of temperature and salinity ( Cherniavski 1985). The Jamskoe non-stratified waters north of the convergence can be easily identified by low temperature (5–8 °C) and high salinity. The Kuril water south of the convergence is strongly stratified with the upper layer (25–30 m) warmed by solar radiation to 10–14 °C, and there is a strong thermocline.
The KJF enters the Tauyskaya Bay area near Zavialova Island ( Fig. 1). The Tauyskaya Bay’s warm (10–14 °C) coastal waters (CW) flow from the mouth of the Tauy River. Large tidal velocities prevent strong stratification of the CW. When the KJF and CW meet west of Zavialova Island, they form a double-convergence zone and a large cold-core eddy ( Cherniavski 1985; TINRO 1990). One of the most important features of this eddy is a constant downwelling of the dense, cold Jamskoe surface water which may create significant concentrations of zooplankton ( Cherniavski 1984, 1985). The convergence and eddy have quasi-permanent positions. They migrate offshore during the summer under the influence of solar heating, winds and tidal currents. Usually, the convergence and eddy occur from early July until late August in the vicinity of Tauyskaya Bay. As they migrate from the mouth of the bay, the double convergence and the eddy gradually lose their structure.
The data on bird reproductive performance were obtained for colonies on Talan Island ( Fig. 1). Talan Island is a small rocky island in the western corner of Tauyskaya Bay that provides nesting habitats for a minimum of 2 million seabirds of 12 species ( Kondratyev et al. 1992). The island is 2·5 km long, with a tundra plateau above talus slopes and steep granite cliffs up to 200 m high.
Reproductive performance data for crested (Aethia cristatella Pallas) and parakeet (Cyclorhinchus psittacula Pallas) auklets, and tufted (Lunda cirrhata Pallas) and horned (Fratercula corniculata Naumann) puffins were obtained on Talan Island between 1987 and 1994. To record bird reproductive success, 20–100 nests of each species were monitored from egg laying until chick fledging ( Table 1). Breeding sites were checked at 2–7-day intervals. Auklets and puffins lay one-egg clutches and raise a single chick per successful reproductive bout. Reproductive success was estimated as the mean number of chicks fledged per occupied nest. In 1994, we were able to visit the colony only during the late chick-rearing stage so bird reproductive success was estimated as the mean number of chicks found alive in the nest per occupied nest. Excluding the data on birds’ reproductive success in 1994 did not significantly change the main results of the study and, thus, we included them into the analyses.
Table 1. . Sampling effort for analyses of alcid reproductive success on Talan Island during 1987–94
Year of observations 1987
No. of occupied nests observed
Chick diet composition was determined from food samples obtained from parents that were delivering food to their chicks. Auklet food samples were collected by netting adult birds returning to the colonies with food for chicks. Puffin food samples were obtained by netting adults delivering food, and collecting individual food loads dropped by parents in the nest burrows or elsewhere at the colonies. Crested auklet food samples were preserved in 5% formaldehyde prior to determining prey species composition and prey biomass, which was determined with an accuracy of ± 0·01 g. For tufted puffins, prey species composition and prey biomass were determined while food samples were fresh. Individual fish biomass was determined with an accuracy of ± 0·1 g. Because a single food load of auklets and puffins consisted of several prey items, we calculated percentage-biomass of prey species as the proportion of wet biomass of a particular prey species in relation to the total wet biomass of a single food load.
Climate data were obtained from the Meteorological Company at Magadan City (Ocean Department of the Kolymskoe Agency). Daily SST (reported as daily mean), wind speed and wind direction were measured (at a standardized time of day) on Spafarieva Island, which is 7 km south-east of Talan Island ( Fig. 1). Ice maps were based on the weekly aerovisual observations of ice distribution during April to June of each year. Information on the distribution of SST, bottom water temperature and zooplankton biomass in the Tauyskaya Bay area were obtained from the Magadan Branch of the Pacific Institute of Fisheries and Oceanography. These data were collected in 1988 and 1989 at 61 set stations evenly covering the entire bay area. Water temperatures were measured by using a deep-water thermometer with an accuracy of ± 0·1 °C. Meso-zooplankton biomass through the water column (bottom to surface) was sampled by using a standard Jetty conical net (diameter 37 cm, mesh size = 0·168 mm). Macro-zooplankton biomass through the water column (bottom to surface) was obtained by using an Isaacs–Kidd mid-water trawl (mesh size = 0·588 mm). Wet (after preservation in 5% formaldehyde) zooplankton biomass was estimated volumetrically (using methods described in Lubny-Gercig 1953) with an accuracy of ± 0·1 g.
We used Pearson correlation to test for relationships between the date of ice disappearance and the SST anomaly in the July–August period of each year. In particular, SST anomalies were calculated by subtracting a mean SST during the July August period of each year from the long-term mean SST during the same period between 1986 and 1994. The date of ice disappearance was taken as the date when fields of permanent ice moved below 59°00′ N latitude.
To test for interannual differences in SST, we compared variation in daily SST from 1 July to 30 August with one-way anova with year as the factor. For this comparison we chose the chick-rearing period of alcids on Talan Island ( Kondratyev et al. 1992), and this period is the most critical for overall reproductive performance ( Golubova 1992). To test for differences between the reproductive performance of planktivorous and piscivorous alcids in response to changes in the July–August SST, we conducted a heterogeneity test (test for a parallelism of slopes) with an ancova, where the dietary preferences (planktivory vs. piscivory) were used as a factor and the SST (yearly averages for the period from 1 July to 30 August) as a covariate. The interaction term between the dietary preferences and the covariate represent differences in the reproductive responses of piscivorous and planktivorous alcids to the same sets of oceanographic conditions. Comparisons of the species-specific reproductive responses to variation in SST were carried out with ancova, where species were used as a factor and SST as a covariate. Planktivorous species and piscivorous species were analysed separately.
To compare wind and current patterns between ‘cold’ and ‘warm’ regimes, we used data for wind speed and direction collected between 1 July and 31 August in 1988 and 1989. To calculate a wind vector, we multiplied mean wind speed by the monthly frequency of wind from a particular direction (all winds were assigned to 16 major directions). To calculate a monthly mean of wind vectors and to compare wind vectors between the seasons, we followed methods described in Zar (1984) for data on circular scales. Accordingly, means of wind vector lengths were compared with a t-test, whereas means of wind vector angles were compared by using a two-sample Hotelling test. Wind-driven water currents were assumed to be from 45° (at the surface) to 90° (at depth) to the right of the wind vector, to adjust current vectors for the Coriolis force ( Pickard & Emery 1990).
To compare distributions of meso- and macro-zooplankton, and the SST in Tauyskaya Bay, we used data collected during July and August of 1988 and 1989. These are the only available complete data sets for ‘warm’ and ‘cold’ regimes. We treated the means of parameters for each of the oceanographic stations as independent samples. We compared means of the parameters at each station between 1988 and 1989 years using a paired-sample analysis (paired by station location). Contour plots used in this study were generated with the surfer package ( Keckler 1994).
To test for interannual differences in diet composition of crested auklets and tufted puffins, we used data collected during July and August of 1988 and 1989. A single food load was used as a sample unit. We compared means of the percentage-biomass of each prey species between 1988 and 1989 years using non-parametric Kruskall–Wallis anova (statistical values are reported as chi-square approximation, c2, degrees of freedom, d.f. and P-value). We compared means of wet biomass of a single food load of tufted puffins using independent sample t-tests.
All calculations and statistical tests were performed using systat ( Wilkinson 1992). For each analysis, the data were tested for assumptions required by a test (according to Sokal & Rohlf 1981). If data violated assumptions for parametric tests, we used non-parametric equivalents. We used the criterion of P < 0·05 to reject a null hypothesis.
Ice conditions in the spring and SST anomaly in the summer
Two distinct states of ice conditions during the spring–early summer period were observed in Tauyskaya Bay and adjacent waters. During 1989, 1990, 1991 and 1994 solid ice disappeared from the area above 59°00′ N latitude between 7 May and 14 May. During 1989, 1990, 1992 and 1994 solid ice dissapeared from the area between 21 May and 7 June.
July–August daily SST varied significantly interannually (F8,552= 156·3, P < 0·001). July–August SST were below 9·3 °C (the long-term mean from 1986 to 1994) in 1986, 1987, 1988, 1992 and 1993, which indicated a ‘cold’ temperature regime ( Fig. 2). July August SST above 9·3 °C indicated a ‘warm’ temperature regime in 1989, 1990, 1991 and 1994 ( Fig. 2).
July–August SST anomaly during 1987–94 was negatively correlated with yearly dates of the disappearance of solid ice from the area above 59°00′ N latitude (r2= − 0·804, P= 0·016, n= 8; Fig. 3).
Interannual variation in July–August SST anomaly and bird reproductive performance
Alcid reproductive success on Talan Island varied interannually during the period of observations from 1987 to 1994 ( Fig. 2). However, piscivorous and planktivorous alcids differed significantly in their reproductive responses to the change of SST (as indicated by significant interaction between the dietary preferences and variation in SST: F1,28= 18·678, P < 0·001, Fig. 4). Horned and tufted puffins showed similar patterns of interannual variation in reproductive success (F1,13= 0·555, P= 0·481), which was significantly associated with variation in the SST (F1,13= 12·275, P= 0·004). Reproductive success of piscivorous puffins was positively correlated with SST (r2= 0·475, P= 0·003, n= 16). Crested and parakeet auklets showed similar patterns of reproductive success (F1,13= 0·447, P= 0·447), and these patterns were significantly associated with variation in SST (F1,13= 6·774, P= 0·022). However, in contrast to the piscivorous puffins, reproductive success of planktivorous auklets was negatively correlated with SST (r2= 0·332, P= 0·019, n= 16).
Comparison of ‘cold’ and ‘warm’ regimes of the tauyskaya bay ecosystem
A negative SST anomaly in the vicinity of Spafarieva Island represented a ‘cold’ temperature regime of the Tauyskaya Bay ecosystem in July August of 1988 ( Fig. 2). Similarly, a positive SST anomaly indicated a ‘warm’ temperature regime of the bay in July August of 1989 ( Fig. 2).
The distribution of SST in the Tauyskaya Bay area was characterized by a strong temperature stratification of cold (temperature range from 5° C to 9° C) and warm (temperature above 13° C) water masses in 1988 ( Fig. 5). The area of the mixed cold and warm water masses (temperature range from 9° C to 13° C) was relatively small ( Fig. 5). The Kuril–Jamskoe oceanographic front (KJF) was well defined only in the western part of the bay and in the vicinity of Spafarieva Island ( Fig. 5). The large cold-core eddy was situated in the central part of Tauyskaya Bay ( Fig. 5). In contrast, the bay area was well separated by KJF from the oceanic waters in 1989 ( Fig. 5). The distribution of SST was homogeneous and there were no signs of a temperature stratification of surface waters in the central part of the bay in 1989 ( Fig. 5). The cold-core eddy appeared farther east in the Tauyskaya Bay area in 1989 compared to 1988 ( Fig. 5).
The distribution of bottom temperatures in 1988 and 1989 generally resembled the distribution of SST in Tauyskaya Bay ( Fig. 6). Cold water masses (temperature range from – 1° C to – 0·5° C) were present in the south-western part of the bay during the ‘cold’ regime in 1988 ( Fig. 6). Warm bottom waters (temperature range above 0° C) were present in the north-western and eastern parts of the bay in 1988 ( Fig. 6). In contrast, only small patches of cold bottom waters were present in the bay area during the ‘warm’ regime in 1989.
Differences in the distribution of surface and bottom water temperatures in the bay were probably related to the wind-driven water circulation. Wind strengths, as indicated in lengths of wind vectors, were indistinguishable between the ‘cold’ regime in 1988 and the ‘warm’ regime in 1989 (t= – 0·439, P= 0·644, n= 32). However, the orientations of wind vectors were significantly different between years (F2,29= 238·8, P < 0·001, Fig. 7). In 1988, the wind vector was southerly orientated, and wind-driven currents flowed toward south-west and west. In 1989, the wind vector was east-north-easterly orientated, and wind-driven currents were orientated toward east-south-eastern and south-south-eastern directions.
Distribution of meso-zooplankton
The distribution of zooplankton biomass in Tauyskaya Bay area differed considerably between the 1988 and 1989 seasons ( Fig. 8). Dense concentrations of meso-zooplankton [mostly copepods Pseudocalanus, Calanus, Acartia, Metridia and Neocalanus genera ( TINRO 1988, 1990)] were present in the coastal areas, in the cold-core eddy area and in the eastern corner of the bay in 1988. In 1989, high densities (2000–2500 mg 1000 m–3) of meso-zooplankton appeared north of Talan Island ( Fig. 8). Meso-zooplankton biomass, on average for the bay area, was significantly lower during the ‘cold’ regime in 1988 compared to the ‘warm’ regime in 1989 ( Table 2).
Table 2. . Biological characteristics of the Tauyskaya Bay ecosystem during ‘cold’ and ‘warm’ states
Distributions of macro-zooplankton in the bay area were remarkably different between the ‘cold’ regime in 1988 and the ‘warm’ regime in 1989 ( Fig. 9). Adult euphausiids, Thysanoessa rashii Sars, were the most abundant and composed about 75% of the total macro-zooplankton wet biomass ( TINRO 1988, 1990). In 1988, large concentrations of macro-zooplankton in the water column were recorded in the western and central parts of the Bay, and in the cold-core eddy area ( Fig. 9). In 1989, a very low biomass of macro-zooplankters was recorded in the western and central parts of the bay ( Fig. 9). High concentrations of macro-zooplankton were recorded only along the KJF area north of the bay and in the eastern part of the bay. Biomass of macro-zooplankton, on average for the bay area, was significantly higher during the ‘cold’ year of 1988 compared to the ‘warm’ year of 1989 ( Table 2).
Crested auklet chick diets
Diet compositions of crested auklet chicks consisted of mostly euphausiids, Thysanoessa rashii, in both the ‘cold’ regime in 1988 and the ‘warm’ regime in 1989 ( Table 3). Relative proportions of meso- and macro-zooplankton were not significantly different between the crested auklet diets in 1988 and 1989. However, the proportion of juvenile fish increased in the diet in the ‘cold’ regime of 1989 ( Table 3). Similarly, proportions of calanoid copepods, Neocalanus plumchrus and Calanus glacialis varied significantly between the ‘cold’ and the ‘warm’ regimes ( Table 3). The proportion of N. plumchrus in the diet was significantly higher during the ‘cold’ regime in 1988. Conversely, the proportion of C. glacialis was significantly higher during the ‘warm’ regime in 1989.
Table 3. . Composition of crested auklet chick diets during ‘cold’ and ‘warm’ temperature regimes of the Tauyskaya Bay ecosystem
‘Cold’ 1988 wet biomass (%) n= 243 Mean (SE)
‘Warm’ 1989 wet biomass (%) n= 97 Mean (SE)
†Prey species, for which proportions in the diets were significantly (Kruskal–Wallis test χ2 approximation ≥23·2, 1 d.f., P < 0·001) different between the seasons, are shown in bold font. The data were collected on the Talan Island colonies during the same period from 7 July to 2 August in both years; sample size (n) represents the number of individual food loads.
Diet compositions of tufted puffin chicks were more diverse ( Table 4) during the ‘cold’ regime in 1988 (13 species of prey) than during the ‘warm’ regime in 1989 (seven species of prey). Conversely, mass of a single food load was significantly higher in 1989 compared to 1988 ( Table 4). Relative proportions of 1 + age class herring (Clupea harengus) were not significantly different between the tufted puffin diets in 1988 and 1989. However, the proportion of 1 + age class sandlance (Ammodytes hexapterus) increased significantly in the diet in the ‘warm’ regime of 1989 ( Table 4). Conversely, proportions of very small (0 + age class) Osmeridae, flat fish and the poucher, Aspidophoroides bartonii, in the diet were significantly higher during the ‘cold’ regime in 1988. Similarly, the proportion of squid was significantly higher during the ‘cold’ regime in 1988.
Table 4. . Composition of tufted puffin chick diets during ‘cold’ and ‘warm’ temperature regimes of the Tauyskaya Bay ecosystem
‘Cold’ 1988 wet biomass (%) n= 72 Mean (SE)
‘Warm’ 1989 wet biomass (%) n= 80 Mean (SE)
† Prey species, which proportions in the diets were significantly (Kruskal–Wallis tests χ2 approximation ≥4·53, 1 d.f., at *P≤ 0·001, **P < 0·05) different between the seasons, are shown in bold font. †Annual means of mass of a single food load were analysed with independent sample t-test, t= 2·332, 1 d.f., P= 0·02. The data were collected on the Talan Island colonies during the same period from 27 July to 22 August in both years; sample size (n) represents the number of individual food loads.
In this study we have evaluated the hypothesis that interannual climate change causes contrasting trends in reproductive performances of planktivorous and piscivorous alcids. Delayed ice cover in Tauyskaya Bay and the adjacent areas in the spring was associated negative SST anomalies in the eastern part of the bay in the summer. We found that during ‘cold’ years the reproductive performance of planktivorous auklets was higher than those of piscivorous puffins, whereas during ‘warm’ years puffins were more successful than auklets.
For the comparison of the physical and biological changes of the Tauyskaya Bay ecosystem during ‘cold’ and ‘warm’ years, data for only a single ‘warm’ and single ‘cold’ season were available. This sample size is insufficient to demonstrate conclusively that the results for 1988 and 1989 characterize physical and biological phenomena related to the changes in climate during 1987–94. However, it is possible that the physical and biological characteristics of the Tauyskaya Bay ecosystem during the ‘cold’ and the ‘warm’ regimes in 1988 and 1989 are indicative of ‘cold’ and ‘warm’ regimes in general.
Cold and warm water masses were well stratified during the ‘cold’ regime of the ecosystem in 1988, as reflected in the strong stratification of sea-surface and bottom water temperatures ( Figs 5 and 6). A stratification of the water masses probably was the result of the westerly orientated current vectors that were orientated in parallel to the Kuril–Jamskoe oceanographic front. In contrast, a homogeneous distribution of the sea-surface and bottom temperatures characterized a ‘warm’ regime of the Tauyskaya Bay ecosystem in 1989. It implies a strong mixing of the cold and warm water masses in the bay area, which was probable due to the easterly orientated current vectors that were perpendicular to the KJF. A significant difference in the orientations of wind vectors between 1988 and 1989 allow us to suggest that the orientation of wind-driven currents is important in determining a temperature regime of the Tauyskaya Bay ecosystem. The distribution of SST, the variable that we used to characterize the ecological conditions of the system, was probably related to the wind-driven water circulation in Tauyskaya Bay in 1988 and 1989.
The climate change in the bay ecosystem in 1988 and 1989 was associated with changes in zooplankton communities. A significant increase in meso-zooplankton biomass during the ‘warm’ regime compare to the ‘cold’ regime might have been due to the distribution of meso-zooplankton in the bay area. There is also a possibility that the productivity of meso-zooplankton was affected by differences in oceanographic conditions between the ‘warm’ and the ‘cold’ year. For instance, Paul & Coyle (1993) and Paul, Coyle & Haldorson (1991) investigated the food resources of juvenile fish in Auke Bay (Alaska), and found that production and biomass of meso-zooplankton occurred earlier and were higher during ‘warm’ years compared to those in ‘cold’ years. Similarly in Tauyskaya Bay, the biomass of meso-zooplankton was significantly higher during the ‘warm’ regime in 1989 compared to the ‘cold’ regime in 1988 ( Table 2).
An increase in meso-zooplankton biomass can influence the prey accessibility to foraging piscivorous puffins which could cause the observed increase of their reproductive performances during ‘warm’ seasons. Proportions of sandlance significantly increased in the diet of tufted puffins during the ‘warm’ regime of the ecosystem in 1989 compared to the ‘cold’ regime in 1988 ( Table 4). Conversely, proportions of small fish and juvenile squid decreased in the diet of tufted puffins in 1989 ( Table 4). Piscivorous alcids, when they have a choice, are known to select large prey items to feed their young ( Bradstreet & Brown 1985; Piatt 1987; Vermeer, Sealy & Sanger 1987). An increase in proportions of 1 + age sandlance and a decrease in proportions of small 0 + age fish in the diet of tufted puffin chicks resulted in an increase of mass of a single food load delivered by tufted puffins in 1989 ( Table 4). A change in the diet might indicate that puffin parents were able to choose larger prey to feed their chicks in 1989 but not in 1988. At the same time, proportions of juvenile fish increased in the diet of usually planktivorous crested auklets during the ‘warm’ regime of the ecosystem in 1989 compared to the ‘cold’ regime in 1988 ( Table 3). These indirect evidences imply that the abundance of juvenile fish increases in the bay area during a ‘warm’ regime. Other studies have suggested that ‘warm’ oceanographic regimes may be close to an optimum for pelagic fish breeding and foraging in the coastal areas ( Pitt 1958; Reynolds, Thomson & Casterlin 1977; Rose & Leggett 1989). A similar conclusion was reached by Piatt, Pinchuk & Kitaysky (1992) when they analysed distributions of pelagic forage fish and piscivorous seabirds in the northern Bering Sea. They found that the major part of the acoustically determined pelagic fish biomass was in the warmer coastal waters with evenly distributed meso-zooplankton. Their study supported the earlier suggestions by Springer & Roseneau (1985), Springer et al. (1987) and Springer, McRoy & Turco (1989) that piscivorous seabird foraging distribution coincides with distribution of meso-zooplankton communities. We suggest that ‘warm’ events (which are characterized by warmer water temperature and higher meso-zooplankton biomass) can attract pelagic fish to forage in Tauyskaya Bay. An increase in the prey accessibility to puffins can cause the observed increase of the reproductive performances of these piscivorous seabirds with an increase in the SST.
Despite a dramatic difference in the composition of zooplankton communities of Tauyskaya Bay between the ‘cold’ and ‘warm’ regimes, crested auklet diet composition did not change significantly during 1988–89 ( Table 3). The constancy of diet composition implies that crested auklets probably exploited zooplankton patches with similar prey species compositions in the two years. Although there was an increase in meso-zooplankton biomass during the ‘warm’ regime in 1989, small zooplankton such as Acartia spp. and Pseudocalanus spp. dominate meso-zooplankton communities in Tauyskaya Bay (A.I. Pinchuk, personal communication). These zooplankters are probably too small to be captured by crested auklets in substantial amounts ( Table 3). It is also possible that crested auklets require a certain suite of oceanographic conditions that concentrate prey of similar body sizes or swimming abilities in patches that would be optimal for foraging. Observations on foraging activities of planktivorous auklets in relation to micro-oceanographic features suggest that this is possible ( Piatt et al. 1992; Hunt et al. 1998).
The composition of crested auklet chick diets appeared to be insensitive to the variation in abundance of meso- and macro-zooplankton in Tauyskaya Bay during the ‘cold’ regime in 1988 and the ‘warm’ regime in 1989 ( Table 3). Furthermore, meso-zooplankton occurred in larger densities and closer to the colony during the ‘warm’ regime in 1989 ( Fig. 8). However, it is likely ( Fig. 9) that macro-zooplankton were either more dispersed in the bay during ‘warm’ years or located more distant from the breeding colonies and, consequently, are less available to foraging parent auklets. For instance, relatively high densities of macro-zooplankton occurred west of the colony and were also present in the central Tauyskaya Bay areas during the ‘cold’ regime in 1988. During the ‘warm’ regime in 1989, macro-zooplankton did not occur in the central Tauyskaya Bay areas at all ( Fig. 9). During our work at the colony, we always observed crested auklets leaving the colony in north-western directions and subsequently returning with food for their chicks from the same directions. Although we do not have data on distribution of crested auklets at sea, we believe that birds could have been foraging in water masses associated with the cold-core eddy in the central Tauyskaya Bay ( Fig. 1). During the ‘cold’ regime in 1988, the cold-core eddy occurred closer to the colonies compared to its position during the ‘warm’ regime in 1989 ( Fig. 5). Other studies of at-sea distribution of planktivorous birds documented long-term associations of foraging planktivorous auklets with quasi-permanent oceanographic features ( Piatt et al. 1992; Hunt et al. 1998). We suggest that a low availability of macro-zooplankton in the central Tauskaya Bay area during ‘warm’ years is probably responsible for the low reproductive success of planktivorous auklets that breed on Talan Island.
The change in proportions of different species of meso-zooplankton may be indicative of a change in composition of the meso-zooplankton communities between the ‘cold’ and the ‘warm’ regimes. During the ‘cold’ regime in 1988, the oceanic N. plumchrus was a predominant species of copepods in the crested auklet diets ( Table 3). Inversely, during the ‘warm’ regime in 1989, the C. glacialis was the most important species of copepods that had been taken by foraging crested auklets ( Table 3). It is probable that a strong in-flow of oceanic waters inside of Tauyskaya Bay during the ‘cold’ regime in 1988 brought oceanic copepods to the foraging grounds of crested auklets. During the ‘warm’ regime in 1989, bay waters were isolated by the Kuril–Jamskoe oceanographic front from oceanic waters and in-flow of oceanic waters inside of the bay was probably weak. This would explain a smaller proportion of oceanic N. plumchrus in crested auklet diets during a ‘warm’ regime in 1989. Subsequently, relative abundance of the C. glacialis would increase during ‘warm’ years as indicated in crested auklet diets during the ‘warm’ regime in 1989. Unfortunately, it is not possible to evaluate, with the data we have, whether C. glacialis dominates the meso-zooplankton community of Tauyskaya Bay during ‘warm’ years, whereas N. glacialis is a predominant species during ‘cold’ years. However, the data on the composition of crested auklet diets during the ‘cold’ and the ‘warm’ regimes in 1988 and 1989 suggest this is true.
It is possible that the relationships between alcid reproductive success and oceanographic changes (reflected in SST fluctuations) that we observed in the Northern Okhotsk Sea occur on larger spatial and temporal scales. Fluctuations of the marine climate in the northern Sea of Okhotsk and in the central and south-eastern Bering Sea have been related to interactions between the Siberian Anticyclone and Aleutian Depression ( Afanasiev 1989; Royer 1989). In the Bering sea, a ‘cold’ oceanographic regime in the Bering Sea occurred between the mid-1950s and late 1970s. According to our study, during this period piscivorous birds would experience a decreased reproductive success, whereas planktivorous birds could have an enhanced reproductive success. These contrasting reproductive responses of piscivorous and planktivorous birds may explain the observed decline of piscivorous seabirds populations and the increase of populations of planktivorous auklets in Bering Sea since 1960s ( Piatt et al. 1990; Konyoukhov 1991; Springer 1992). The results of our study support ‘bottom-up’ hypotheses ( Springer 1992; Decker et al. 1995; Hunt et al. 1996a) that oceanographic changes influence zooplankton abundance and/or distribution and hence foraging costs of birds.
We thank Olga Kalashnikova, Irina Ryazantseva, Alexei Pinchuk, Luba Kondratieva, Konstantin Yakovlev, Valeri Zarudni, Alexander Pil’nikov, Sergei Pleshenko and Alexander Kondratiev for their great assistance in the field and in the office. George Hunt, John Piatt and Alexander Andreev provided stimulation and financial support for this study. Financial support was provided by the US National Science Foundation, Division of Polar Programs grant DPP 85–21178 to George L. Hunt, by National Biological Survey to John Piatt, and by the Northern Biological Institute of the Russian Academy of Sciences.
Received 24 October 1998;revisionreceived 7 June 1999