BioScience Vol. 49(5), 1999 (393-404) Marine Ecosystem Sensitivity to Climate Change: Raymond C. Smith, David Ainley, Karen Baker, Eugene Domack, Steve Emslie, Bill Fraser, James Kennett, Amy Leventer, Ellen Mosley-Thompson, Sharon Stammerjohn, and Maria Vernet Raymond C. Smith, Professor Institute for Computational Earth System Science (ICESS) and Department of Geography 6812 Ellison Hall, University of California, Santa Barbara (UCSB) Santa Barbara, CA 93106-3060, USA Phone: 805-893-4709, Fax: 805-893-2578, email: ray@icess.ucsb.edu Other Authors: David Ainley, Senior Ecologist, H.T. Harvey & Associates, P.O. Box 1180, Alviso, CA 95002 (dainley@harveyecology.com) Karen Baker, Data Manager/Analyst, Scripps Institution of Oceanography (SIO), La Jolla, CA 92093-0218 (karen@icess.ucsb.edu) Eugene Domack, Professor, Department of Geology, Hamilton College, 198 College Hill Road, Clinton, NY 13323 (edo- mack@hamilton.edu) Steve Emslie, Assistant Professor, Department of Biological Sciences, University of North Carolina, Wilmington, NC 28403 (emslies@uncwil.edu) Bill Fraser, Montana State University, (wrf@gemini.oscs.montana.edu) James Kennett, Professor of Oceanography, Department of Geological Sciences, UCSB, Santa Barbara, CA 93106-9630 (kennett@magic.geol.ucsb.edu) Amy Leventer, Assistant Professor, Department of Geology, Colgate University, Hamilton, NY 13346 (aleventer@mail.colgate.edu) Ellen Mosley-Thompson, Professor, Byrd Polar Research Center, Department of Geography, Ohio State University, Columbus, OH 43210-1002 (thompson.4@osu.edu) Sharon Stammerjohn, Research Analyst, ICESS, UCSB, Santa Barbara, CA 93106-3060 (sharon@icess.ucsb.edu) Maria Vernet, Associate Research Oceanographer, SIO, La Jolla, CA 92093-0218 (mvernet@ucsd.edu) Subhead: Historical observations and paleoecological records reveal ecological transitions in the Antarctic Peninsula region Quote: This century's rapid climate warming is occurring concurrently with a shift in the population size and distribution of penguin species. Mounting evidence suggests that the earth is experiencing a period of rapid climate change. Never before has it been so important to understand how environmental change influences the earth's biota and to distinguish anthropogenic change from natural variability. Long-term studies in the western Antarctic Peninsula (WAP) region provide the opportunity to observe how changes in the physical environment are related to changes in the marine ecosystem. Analyses of paleoclimate records show that the WAP region has moved from a relatively cold regime, between about 2700 and 100 years before present (B.P.), to a relatively warm regime during the present century. Air temperature records from the last half century show a dramatic warming trend, confirming the rapidity of change in this area (Sansom, 1989; Stark, 1994; Rott et al., 1996; Smith et al., 1996). Significantly, polar ecosystem research over the last few decades (Fraser et al., 1992; Trivelpiece and Fraser, 1996) and paleoecological records for the past 500 years (Emslie, 1995; Emslie et al., 1998) reveal the occurrence of ecological transitions in response to this climate change. In this article, we summarize the available data on climate variability and trends in the WAP region and discuss these data in the context of long-term climate variability during the last 8000 years of the Holocene. We then compare the available data on ecosystem change in the WAP region to the data on climate variability. Both historical and paleoenvironmental records indicate a climate gradient along the WAP which includes a dry, cold continental regime to the south and a wet, warm maritime regime to the north. The position of this climate gradient shifts over time, depending on the dominant climate regime, and makes the WAP region a highly sensitive location for assessing ecological response to climate variability. Our findings show that this century's rapid climate warming is concurrent with a shift in the population size and distribution of penguin species. Climate change in the western Antarctic Peninsula region The marine environment of the WAP (Figure 1) includes an open ocean area of shelf-slope waters, frontal regions, a highly variable seasonal sea ice zone, and ice-free and glacier covered islands and coastal areas. High variability and long-term change constitute the setting in which this polar marine ecosystem has evolved. Understanding the history of ecosystem development in response to environmental changes presents a considerable challenge, particularly because there are few time-series data of physical and biological change in this region of the world. Modern instrumental records that do exist span at most the last half of this century. Fortunately, paleoclimate records derived from Antarctic ice sheet and marine sediment cores, when carefully analyzed, supplement and extend the limited historical observations and modern instrumental records. They provide a critical context for understanding the most recent warming trend and the present-day biotic-environmental relations. We discuss climate indicators from four sources: historical observations and modern instrumental records, ice core records, marine sediment records, and seabird distributions. All of these climate indicators are consistent in showing a rapid warming trend in the WAP beginning this century. The WAP region is the area studied by the Palmer Long-Term Ecological Research (Pal LTER) program (Smith et al., 1995), and a summary volume on ecological research for the area can be found in Ross et al. (1996). Historical observations and modern instrumental records. Historic observations of physical features can indicate climate change from the recent Antarctic past. For example, fast ice forms seasonally and is attached to a shoreline, making it easy to be observed from a coastal research station. Its duration typically reflects the annual advance and retreat of pack ice at that location. The British Antarctic Survey research station in the South Orkney Islands (Figure 1) has a 90-year time series of fast-ice duration in that area. Despite the high interannual variability, the data show an overall decreasing trend in fast ice duration (Murphy et al., 1995) that is presumably in response to regional warming. Another physical feature historically observed is the extent of ice shelves, floating ice sheets attached to the coast that are of considerable thickness and typically rising 2-50 m or more above sea level. Ice shelves are nourished by annual snowfall and often also by land glaciers, and portions may be aground. Along the Antarctic Peninsula, ice shelves exist up to a climatic limit, corresponding to the mean annual temperature isotherm of -8 degC (Vaughan and Doake, 1996), and are therefore sensitive to atmospheric and oceanic warming. The more northern ice shelves in the WAP region have a history of recession over the past fifty years, which has been related to increased air temperatures observed in the Antarctic Peninsula region (Doake, 1982; Doake and Vaughan, 1991; Vaughan and Doake, 1996). Comprehensive studies of southern hemisphere land-based surface air temperature variations over the past century indicate a long-term warming trend of about 0.5 degC (1881 to 1984) (Jones et al., 1986) to 0.6 degC (1880-1985) (Hansen and Lebedeff, 1987). These data are strongly correlated with the southern hemisphere marine temperature series of Folland et al. (1984). Subsequently, Jones (1990) combined climatological station data (Jones and Limbert, 1987) and air temperature records for 26 overwinter expeditions in the Antarctic Peninsula and Ross Sea sectors of Antarctica to extend the time-series of mean annual air temperatures from 1909 to 1987. Jones estimated annual air temperature values for each expedition site as anomalies from the reference period 1957 to 1975. Four of the five Antarctic regions studied exhibited linear trends of at least 2 degC since the beginning of this century. Jones concluded that Antarctica is now at least 1 degC warmer than it was at the beginning of the 20th Century. Most instrument records for the Antarctic date from 1957, the International Geophysical Year. These records provide a relatively short time-series compared to the more than hundred year instrumental records from the more temperate regions in the southern hemisphere and elsewhere. Both warming and cooling trends have been reported for some Antarctic continental regions (Rogers, 1983; Taylor et al., 1990; Weatherly et al., 1991), but their statistical significance is unclear due to the brevity of records and high interannual variability. Instrument records from several coastal stations in the WAP region show statistically stronger trends as well as an along-peninsula gradient in climate variables (e.g. surface air temperature), (Schwerdtfeger, 1970; Sansom, 1989; King, 1994; Harangozo et al., 1997). This gradient of contrasting influences of marine versus continental climate regimes, with the potential for these contrasting regimes to shift in dominance, makes the WAP region unusually sensitive to climate change. For example, several studies (King, 1994; Stark, 1994; Smith et al., 1996) have shown a warming of 4-5 degC over the past 50 years in midwinter surface air temperature in the WAP region, indicating that the maritime regime has been more dominant during this interval. The time-series of annual average surface air temperature (British Antarctic Survey, 1998), recorded at Faraday Station (Figure 1) from 1945 to 1997, shows this warming trend (Figure 2). The least-squares regression of the warming trend, which is significant at the 94% confidence level, is due in large part to the high frequency of warm years during the last three decades. Air temperatures at Faraday Station have been above average for 20 of the 2 years from 1970 to 1996. We also analyzed each month of the Faraday air temperature time-series separately. Breaking down the seasonal time-series in this way helps to eliminate autocorrelation in the error terms and shows whether the warming trend is more prevalent in any one season. Table 1 shows the trend detected for each month's time-series (as well as seasonal and annual trends for comparison), the standard error and the significance of the linear regression analysis as estimated by the F-test. The F-test shows that 9 of the 12 monthly regressions are significant at confidence levels greater than 95%. The warming trend in Faraday air temperatures is strongest in mid-winter months and peaks in June at 0.097 degC/yr, representing a 5.0 degC increase in June temperatures over the 54-year period. It is important to note that the mid-winter amplification of the warming trend can affect the extent, thickness, and concentration of the seasonal sea ice cover as well as the marine ecology associated with sea ice. Ice core records. Carefully extracted ice cores provide annual- to decadal- scale paleoenvironmental records that can supplement and extend the historical observations and modern instrumental records. Proxy histories of atmospheric conditions, including temperature, chemistry, dust content, and volcanic emissions, are contained in ice core records. These paleoenvironmental histories are inferred from the particulates, chemical species, and gases preserved in ice sheets and ice caps (Mosley-Thompson, 1992; Peel, 1992). In addition, the history of net snow accumulation may be reconstructed upon identification of individual annual layers or known time-stratigraphic horizons. The delta18O isotopic record extracted from an ice core on the Dyer Plateau (Figure 1) provides a 480-year paleoclimate history of the Antarctic Peninsula region (Figure 3) (Thompson et al., 1994; Dai et al., 1995). The delta18O composition of ice reflects a combination of four factors: the air temperature at which condensation occurs, atmospheric processes occurring along the trajectory of the air mass from the source of evaporation to the site of condensation, local conditions during the transformation of the snow into ice, and the surface elevation and latitude of the deposition site (for review, see Dansgaard et al. 1973 and Bradley 1985). The Dyer Plateau record indicates an increase in net snow accumulation since the beginning of the 20th century and a distinct warming trend beginning in the 1940's. Ideally the delta18O signal should be calibrated against in situ observations, but doing so was difficult due to the short occupation of this drilling site. However, the 18O history from the Dyer Plateau cores has been shown to be consistent with available temperature observations recorded at various Antarctic Peninsula stations over the last 50 years (Thompson et al., 1994). The 480-year Dyer Plateau delta18O history (Figure 3) shows almost no consistent pattern except for a marked cooling around 150 years B.P. (i.e., 1845), followed by nearly a century of below normal temperatures, which was followed by an abrupt warming in the last half century. Warming, as indicated by less negative delta18O, began in the 1930's and has continued to the present, with the last two decades being among the three warmest decades in the 480-year history. Marine sediment records. Marine sediment cores have been used to reconstruct past climate and associated productivity changes in the WAP region for the last 8000 years (Domack et al., 1993). The long-term history of the WAP has been marked by pulses of warm and cool episodes accompanied by coincident changes in the extent and duration of regional ice cover (Domack and McClennen, 1996; Leventer et al., 1996). These marine sediment records not only are crucial for characterizing the long-term natural variability in the WAP region, they also provide context for this century's warming trend as documented in the historical observations, modern instrumental records and ice cores. The marine sediment cores from fjords and bays along 450 km of the WAP (Figure 1) contain a paleoclimate record that varies in temporal resolution, composition, and nature of paleoenvironmental proxy. In the southern region of the WAP (Lallemand Fjord), low temperatures, perennial sea ice and low glacial meltwater give rise to slow (1-2 mm/yr) rates of ice rafted terrigenous sedimentation (i.e., sedimentation derived from rock debris that is carried by sea ice, icebergs or ice shelves and is released to the water column as ice melts, overturns or fractures). In the mid-peninsula region (Palmer Deep), the pronounced seasonality of sea ice gives rise to higher (3-4 mm/yr) rates of biogenous (i.e., produced by living organisms) and ice rafted terrigenous sedimentation. In the northern region of the WAP (Brialmont Cove), the long melt season gives rise to very high (5-10 cm/yr) rates of ice rafted terrigenous sedimentation. The variability in sedimentation content and rate along the WAP reflects the regional climate gradient and is used to infer the environmental context of deposition. Sedimentation rates determine the amount of time represented in a given length of core, the slower the sedimentation rate, the longer the record. The slow sedimentation rate found in the Lallemand Fjord core yields a deposition record extending back to at least 8000 radiocarbon years B.P. in only 5 meters of core. For the Lallemand Fjord core, the most effective paleoenvironmental proxy so far discovered is total organic carbon (TOC) which reflects the record of primary production. TOC levels have been found to be elevated during minimal sea ice conditions and depleted during persistent sea ice conditions (Shevenell et al., 1996). Thus, changes in TOC levels are inferred to reflect oscillations between warm and cold episodes. Within the time frame captured by the Lallemand Fjord core (Figure 3), we can most easily resolve millennial-scale variations in climate, such as the middle Holocene warm episodes (5000-2500 year B.P.), the Neoglacial cooling beginning at about 2500 years B. P., and the Little Ice Age from about 500 years B. P. to the present century. At the location of the Palmer Deep core (Figure 1), biogenous sedimentation is enhanced primarily by favorable sea ice conditions and a more open-ocean location. The name Palmer Deep arises from the fact that it is a 1440 meter deep basin surrounded by shallower ridges (200 to 500 meters deep) of the continental shelf. The pelagic and hemipelagic sediments of this basin were deposited in this partially isolated natural sediment trap which filters out stratigraphic noise that might otherwise result from the presence of sediment gravity flows (slumps and debris flows). For the Palmer Deep core, magnetic susceptibility is the paleoenvironmental proxy for productivity, whereby low and high magnetic susceptibility is inferred to reflect intervals of high and low productivity, respectively (Leventer et al., 1996). The nine meter long Palmer Deep core shows bi-century oscillations of primary productivity and ice rafting over the last 3700 years (Figure 3). Comparison of the Lallemand Fjord and Palmer Deep records clearly demonstrates a similar record of events for the last 3700 years, although at different temporal resolutions. The relatively high temporal resolution of the Palmer Deep record also displays both short-term cycles (200-300 years) and longer term events (~2500 year cycles). Both the long- and short-term cycles are thought to be related to global climatic fluctuations (Leventer et al., 1996). Oxygen (delta18O) and carbon (delta13C) isotopic records have proven to be highly effective proxies of paleotemperature change during the Holocene, especially in the interval after the retreat of glaciers following 6000 yrs B.P. Oxygen and carbon isotopic records have been extracted from benthic foraminifera (Bulimina aculeata) found in Palmer Deep cores. Benthic foraminiferal delta18O oscillations range between about 0.25 and 0.5 parts per thousand which suggests oscillations in bottom water temperature of about 1 to 2 degC during the late Holocene (i.e., the last 3700 years). The inferred paleotemperatures of waters in these deep basins indicate that there have been distinct climate shifts, some occurring over periods as brief as a few decades. Bottom waters in the Palmer Deep are largely derived from upwelled Circumpolar Deep Water (CDW), amounts of which may have fluctuated through time (Ishman, 1990). CDW constitutes the water mass with the greatest volume in the Southern Ocean (Sievers and Nowlin, 1984; Carmack, 1990), is marked by a temperature maximum at intermediate depths, and periodically floods the bottom waters of the WAP region where the continental shelf is relatively deep. Thus, the temperature oscillations of Palmer Deep bottom water during the last several thousand years may reflect the changing influence of CDW in the shelf-slope waters of the WAP. Because CDW is a relatively warm water mass, it may provide a heat source to surface waters, especially during periods of greater CDW upwelling. Consequently, the increased sea surface temperatures would result in decreased sea ice extent. This hypothesis is supported by the changes in the carbon isotopic (delta13C) record, which indicates that warmer intervals of the climate record are generally associated with lower delta13C values. Lower delta13C values are indicative of higher nutrient concentrations in bottom water, which is expected with stronger influence of nutrient-rich CDW in the Palmer Deep. Such a correlation is consistent with less sea ice extent and higher primary productivity (Leventer et al., 1996). In the northern region of the WAP where air temperatures are often above freezing, glacial meltwater, which is typically laden with high terrigenous material, contributes significantly to marine sedimentation (Domack and Williams, 1990; Domack and Ishman, 1993). Consequently, it is difficult to extract good chronologies from cores collected in such places as Brialmont Cove, because the biogenous sedimentary components are diluted as a result of the very high terrigenous sedimentation rates caused by glacial meltwater (Domack and McClennen, 1996). However, in two cores which had intact sediment-water interfaces, meltwater events were recognized based upon the abundance of sand and other terrigenous deposits. These features were restricted to the uppermost meter (Domack and McClennen, 1996). Domack (1990) suggest that these sand intervals resulted from changes in sedimentary processes as a result of recent warming in the region. The latest chronology for these cores (Figure 3) shows that the inferred melt-water events began about 60 years ago, with a noticeable abundance in sand about 25 to 35 years ago. This observed depositional change in the sedimentary process and inferred climate change is consistent with historical records of increasing mean monthly surface air temperatures. Seabird distributions. Seabirds are relatively long-lived (15-70 years) upper-trophic level predators that appear to integrate environmental variability over large spatial (thousands of square kilometers) and temporal (years to decades) scales (Ainley and Boekelheide, 1990; Furness and Greenwood, 1993). Seabirds also are relatively wide ranging, because they are dependent upon the patchy distribution of their prey as well as responsive to changes in their physical environment (e.g., sea ice) (Fraser et al., 1992). Consequently, changes in their abundance and distribution provide an indication of ecosystem and environmental change. For example, in the WAP region and elsewhere in Antarctica, information from both paleoecological (Baroni and Orombelli, 1991; Denton et al., 1991; Baroni and Orombelli, 1994; Emslie, 1995) and modern census studies (Taylor et al., 1990; Fraser et al., 1992) suggest that penguin distributions are undergoing a fundamental reorganization due to climatic factors that influence their long-term recruitment. Ade'lie penguins (Pygoscelis adeliae) and their sympatric, most closely related congener, the Chinstrap penguin (P. antarctica) are estimated to comprise more that 80% of the avian biomass of the Southern Ocean (Woehler, 1997). Accordingly, international efforts are underway to monitor changes in the abundance of these species through various national and international programs (e.g., Scientific Committee on Antarctic Research) (SCAR, 1987). Interannual changes in breeding population size is a standard protocol to assess population status (CCAMLR, 1992). The most recent twenty-year trends in Ade'lie and Chinstrap breeding populations on islands near Palmer Station (Figure 4) are believed to be representative of broad trends in the WAP region as a whole (Fraser and Patterson, 1997). In the WAP region, Ade'lie breeding populations have been stable or declining slowly, while Chinstrap breeding populations have increased several hundred percent (Fraser et al., 1992). In contrast, the number of Ade'lie breeding penguins in the Ross Sea area increased rapidly during the 1980's and has since remained relatively stable. The Ade'lie penguin is an obligate associate of winter pack ice (Ribic and Ainley, 1988; Ainley et al., 1994), whereas the Chinstrap penguin occurs almost exclusively in open water (Ainley et al., 1994). It has been suggested that the change in the distribution of breeding populations in recent years has resulted from changes in the environment, especially the reduction in sea-ice related habitats resulting from warming trends in WAP air temperatures (Fraser et al., 1992). Abandoned penguin rookeries in Antarctica, marked by the presence of nest stones, preserved guano and chick remains, can provide considerable information on past population changes and migration of Ade'lie and Chinstrap penguins (Baroni and Orombelli, 1991; Baroni and Orombelli, 1994; Zale, 1994; Emslie, 1995; Emslie et al., 1998). Analysis of bones and other organic remains preserved in the sediments provide information on which penguin species inhabited a site, time of occupation, and their diet. These data can be compared to other paleoclimate records to determine if the changes observed in the penguin populations are indicative of climate change. Six abandoned rookeries, all found to have been previously inhabited by Ade'lie penguins, were excavated in the near vicinity of Palmer Station (Figure 1) in March 1997 (Emslie et al., 1998). Radiocarbon dates of penguin bones and squid beaks recovered from the sediments indicate that these rookeries were occupied from 644 B.P. to the present (mean corrected date, see Emslie et al., 1998). The absence of Chinstrap or Gentoo (Pygoscelis papua) bones in the surveys of the abandoned rookeries suggests either that these two species are relatively recent arrivals to the Palmer area or that their remains have not yet been located. The first hypothesis seems most likely, because the Chinstrap and Gentoo penguins currently inhabiting the Palmer region are near the southern boundary of their breeding range, and the present colonies are believed to have been established in the last 20-50 years (Parmelee, 1992). Thus, records from excavated abandoned rookeries in the Palmer region (Figure 3) show that Ade'lie penguins have been permanent occupants at least over the past 600 years, whereas the recent population expansions by Chinstrap and Gentoo penguins southward along the WAP appear to be correlated with regional warming during the past 50 years (Fraser et al., 1992). Ecological response to climate change Investigations of the impact of possible global warming and climate change on entire ecosystems are becoming increasingly important for understanding interactions among the atmosphere, ocean, and biosphere and for testing models depicting these interactions. Long-term studies in the WAP provide the rare opportunity to integrate time-series data related to the physical environment, biology and paleoecology of the Antarctic marine ecosystem. These data reveal that ecological responses have occurred over the past 500 years in association with well-documented climate changes in the WAP region. The complex trophic relationships in Antarctica are often short and involve relatively few species (see, for example, Figure 3 in Smith et al., 1995 or Figure 11.3 in Ainley and Demaster, 1990). Krill, a major herbivore for the transfer of energy within the Antarctic marine ecosystem, is closely coupled to sea ice during various periods of its annual life cycle (Ross and Quetin, 1986; Quetin and Ross, 1991; Ross and Quetin, 1991; Loeb et al., 1997) and is expected to be an important, but not yet fully understood, link between phytoplankton and penguins. How krill populations change with climate is therefore an important question when considering the consequent influence on penguin populations. At the present time we can not address this and many other questions concerning linkages in the Antarctic marine ecosystem due to a dearth of long-term data. We have focused instead on the ecosystem components for which we do have long term data and have developed two conceptual models depicting Antarctic biotic-environmental relations. In one of the models we use seabirds as a surrogate for change in the upper trophic levels, noting that we, at this time, are unable to disentangle the effects of changes in bird populations that are directly tied to environmental changes versus those that are mediated by trophic interactions (i.e., predator-prey dynamics). Over time we hope that more long-term data on the Antarctic marine ecosystem will become available to test and modify our conceptual models and to create new ones so to more fully describe the processes involved in ecosystem response to climate change. Conceptual model of biogenic flux optimum. A simplified diagram (Figure 5a) illustrates idealized relationships between the sedimentary record and overlying biological production and terrigenous influx. This model represents the modern continuum from polar to subpolar conditions along a south to north gradient roughly parallel to the WAP. Over long time scales, as recorded in paleoclimate records, this system will migrate north or south in response to climate change. At the cold end of this climate gradient, perennial sea ice is present in surface waters. Under these conditions, the contribution of sedimentation from melting ice, which is often rafted with rock debris (i.e., terrigenous material), is low. Thick, multi-year sea ice also reduces light penetration, resulting in little to no biological production in either sea ice or surface waters. The net result is low accumulation rates of biogenous and terrigenous sediments (about 0.01 to 0.02 mm/year) (Domack et al., 1995; Shevenell, 1996; Shevenell et al., 1996). With climate change, the advance and retreat of annual sea ice controls the composition and volume of material settling to the sea floor. Although terrigenous input increases with increased meltwater, the impact of warming on the biological system is even more significant. High primary production is often associated with the marginal ice zone, because melting sea ice (which is fresh relative to sea water) stabilizes the upper water column (Hart, 1934; Smith and Nelson, 1985; Mitchell and Holm-Hansen, 1991). Under these conditions, material produced by living organisms (i.e., biogenous material) dominates the particle flux, and overall sediment accumulation rates range from 1 to 4 mm/year (Domack et al., 1993; Domack and McClennen, 1996; Leventer et al., 1996). This situation currently corresponds to the area just south of Anvers Island, which may be an optimum location for the detection of subtle changes in sea ice extent and its consequent influence on phytoplankton production and flux of carbon to the sediment (Domack and McClennen, 1996). Anvers Island is also hypothesized to be the northern boundary for the distribution of Ade'lie penguins associated with sea ice from the Bellingshausen Sea (Fraser and Trivelpiece, 1996) and roughly the southern boundary for Chinstrap penguins (Fraser et al., 1992). At the warm end of this climate gradient, the region experiences longer ice-free periods. Without sea ice melt to stratify the upper water column, the open water system undergoes increased wind-mixing, resulting in lower annual average primary production and organic flux to the sea floor. However, due to the relative warm conditions, glacial meltwater input of terrigenous material is high, leading to sediment accumulation rates of up to 10 cm/year (Domack, 1990; Domack and McClennen, 1996). Leventer and co-workers (1996) provide a schematic model (see their Figure 12) to illustrate how changes in water column stratification can lead to contrasting phytoplankton bloom conditions leading to different sediment records. Their model, which provides testable hypotheses linking sea ice, upper water column stratification, primary productivity and vertical flux to the sea floor, indicates that of all the factors that may influence primary productivity in Antarctic waters, the variability of sea ice coverage is among the most important. Because the ecological impact of sea ice is a complex space-time matrix of biology and physical forcing, sea ice indexes have been developed to give a common context with which to link variability in sea ice coverage to variability in the marine ecosystem (Smith et al., 1998). Several workers (Sansom, 1989; Weatherly et al., 1991; King, 1994; Smith et al., 1996) have found that the anticorrelation between surface air temperature and sea ice extent is much stronger in the WAP region than elsewhere in the Antarctic. In the WAP region, the statistically significant increase in surface air temperature is associated with an observed decrease in sea ice, particularly summer sea ice, as recorded by passive microwave satellite data over the last two decades (Stammerjohn and Smith, 1997). The decrease in sea ice coverage is in contrast to an observed increase in sea ice coverage for the Antarctic as a whole (Cavalieri et al., 1997; Stammerjohn and Smith, 1997). For the WAP region, if the observed anticorrelation between air temperature and sea ice is assumed to have held for the full 53-year period for which we have air temperature data (1945-1997), there would have been a southward shift of 100 to 200 km in seasonal sea ice coverage along the WAP. Consequently, modern instrumental and satellite records are consistent with the sediment records in suggesting a warming trend with a southward shift in seasonal sea ice extent. Since primary productivity is strongly influenced by the presence or absence of sea ice, the southward shift in sea ice extent is accompanied by a southward shift in the dominance of biotic versus abiotic sedimentation (Figure 5a). Conceptual model of Ade'lie penguin population growth and sea ice. Ade'lie penguins are obligate inhabitants of pack ice whereas their congeners, the Chinstrap penguins, occur almost exclusively in ice-free waters from the Antarctic to the sub-antarctic. These two species have a similar breeding cycle of courtship, egg laying, incubation, brooding and fledging, but the timing of the Ade'lie breeding cycle is roughly one month earlier than the Chinstrap's. Ade'lie penguins begin egg laying from early to mid-November, and after 50-55 days the young birds fledge and depart for the sea in late January to mid-February. Chinstrap penguins lay eggs in late to early December, with fledging and departure of young birds in late February to mid-March. The timing associated with this relatively fixed breeding chronology, in association with interannual variability in sea ice cover and the life histories of primary and secondary producers, provides the ecological context that determines breeding success. Figure 5b illustrates a conceptual model for seabird population growth in Antarctica using Ade'lie penguins as an example. This Ade'lie penguin model shows highest population growth occurring during conditions of moderate sea ice coverage (between extremes of excessive and insufficient sea ice coverage) (Trivelpiece and Fraser, 1996). Ade'lie penguins are true Antarctic penguins - that is, they live within or near the pack ice zone year-round. They show strong fidelity to both a winter sea ice habitat and to their natal rookery to which they return in spring for breeding. Breeding colonies are established in coastal areas whose topography permits them to build pebble nests in places where neither snow nor meltwater accumulate (Wilson et al., 1990). Consequently, the location and long term survival of colonies are associated with several sea ice-mediated factors (Baroni and Orombelli, 1994), including relatively ice-free coastal areas in late spring, the availability of suitable nesting sites (free of snow or meltwater), and sufficient abundance of prey within the foraging range of these sites. The abundance of prey is in turn linked to environmental conditions controlling primary production, in particular the presence or absence of sea ice, as discussed above. Our conceptual model integrates these environmental conditions and is roughly analogous to the intermediate disturbance model (Connell, 1978) which hypothesizes an optimum condition intermediate between extremes of disturbance. Not all mechanisms controlling these environmental conditions are fully understood at present. However, we do know that regions of heavy and persistent sea ice cover or, conversely, no sea ice cover provide unsatisfactory conditions for breeding Ade'lie penguins. Over the past 20 years, the frequency of heavy sea ice years in the WAP region has been decreasing (Figure 2) in conjunction with decreasing Ade'lie populations (Figure 4). In contrast, Ade'lie populations in the Ross Sea (Figure 4) have been increasing (Taylor et al. 1990). At Cape Royds in the Ross Sea region (Figure 1), where the southernmost colony of Ade'lie penguins is located, the recent warming trend and consequent decreasing sea ice is postulated to have improved the nesting success and increased food availability for these colonies (Taylor et al., 1990). Our Ade'lie penguin population growth model provides a hypothesis explaining the different high trophic-level responses to climate change in these regions. Optimum sea ice conditions for Ade'lie penguins no longer exist in the WAP region and populations continue to decline there, whereas in the Ross Sea region, the optimum sea ice and habitat conditions have not yet occurred and populations are increasing. The paleoecology records also support this hypothesis. In the Ross Sea region, there is evidence that Ade'lie populations were larger during 2800 to 4200 B.P. (Baroni and Orombelli, 1994), which, according to the ice core from Dome C (Figure 1), was a warm phase (Lorius et al., 1979). In the WAP region, paleoecology evidence suggests that Ade'lie populations have occupied sites in the area of Palmer Station since at least the coldest period of the Little Ice Age, whereas Chinstrap populations expanded during warm phases of the Little Ice Age (Figure 3) (Emslie, 1995; Emslie et al., 1998). Our conceptual model predicts that, with present day warming, Ade'lie populations in the Ross Sea will reach a peak, then begin to decline. In the WAP region, Ade'lie populations will continue to decline and the locus of their distribution forced further south along the peninsula, while Chinstrap populations will continue to increase. Mechanisms of climate change The mechanisms responsible for the cycles and trends observed in the modern instrumental and paleoclimate records are poorly known. Large-scale atmospheric processes, perhaps driven by global mechanisms (Stuiver and Brazuinas, 1993), acting on the unique geographic features of the WAP, could provide the overall forcing. The Antarctic Peninsula is the only area in Antarctica where the mean position of the circumpolar atmospheric low-pressure trough (i.e., the Antarctic Convergence Line - ACL) crosses land. The seasonal cycle displayed in temperature, pressure, wind, and precipitation (van Loon, 1967; Schwerdtfeger, 1984) is linked to both increased cyclonic activity and a southward shift between 60 to 70 degS of the mean position of the ACL during spring and autumn. The ice edge is near its extreme equatorward (spring) or poleward (autumn) positions when the mean position of the ACL is nearest the Antarctic continent, and conversely, the ACL is on average furthest equator-ward when the sea ice edge is at an intermediate position (van Loon, 1967). Therefore, the relative position of the ACL will influence not only the semiannual cycle of temperature, pressure, wind and precipitation, but also sea ice distribution. The effect of increasing or decreasing sea ice extent could lead to additional feedbacks. For example, greater storminess has been associated with regions of higher temperatures and lower sea ice extent (Ackley and Keliher, 1976). Consequently, large-scale atmospheric processes that influence the intensity of the westerlies and the mean position of the ACL (van Loon, 1967; Schwerdtfeger, 1984; Enomoto and Ohmura, 1990; Harangozo, 1994; Smith et al., 1996) may be the atmospheric mechanisms influencing sea ice extent. Alternatively, or in concert, the oceans may play a critical role in the distribution of sea ice (Broecker, 1990; Stuiver and Brazuinas, 1993; Broecker, 1997) by changes in thermohaline circulation which could influence both the production of oceanic deep water and the strength and distribution of CDW (circumpolar deep water). As noted above, CDW is the water mass with the largest volume in the Southern Ocean (Sievers and Nowlin, 1984; Carmack, 1990) and is found along the margin and over the shelf of the western Antarctic Peninsula at depths greater than a few hundred meters (Domack et al., 1992; Hofmann et al., 1996; Jacobs et al., 1996). Jacobs and Comiso (1993, 1997) suggest that greater upwelling of CDW could reduce sea ice thickness and coverage. Consequently, seasonal heating of additionally exposed open water would lead to later sea ice formation in autumn (Jacobs and Comiso, 1989). Because this relatively warm water mass is sufficient to melt sea ice (Hofmann et al., 1996), a change in the intensity or distribution of CDW, driven either locally or globally, can potentially effect changes in the timing and spatial extent of sea ice in the WAP region. Future Challenges All of the modern and paleoclimate records are consistent in showing a rapid warming trend in the WAP region during this century. This trend is evident in spite of large interannual variability associated with shorter term fluctuations in climate (e.g., El Nino/Southern Oscillation). Before this century's warming trend, the marine sediment record indicates a cooler climate, roughly coincident with the Little Ice Age (beginning about 500 yrs B.P.), which was in turn preceded by a warmer period beginning about 2700 yrs B.P. (Bjo"rck et al., 1991; Domack and McClennen, 1996). In addition, cyclical fluctuation in organic matter preservation on 200 to 300 year time scales in the mid-WAP region (around Anvers Island) resulted from cycles in primary productivity (Domack et al., 1993; Leventer et al., 1996). Evidence for such cycles have not been found in cores at more northern or southern locations along the WAP, and these workers have suggested that in the mid-WAP region, slight changes in sea ice extent might amplify the influence of climate variability on primary production and the subsequent influence on higher trophic levels. Consequently, the climatic gradient found along the WAP is a valuable scale for assessing ecological response to climate variability. Contrasts in habitat preference (sea ice-obligate Ade'lie penguins and sea ice-intolerant Chinstrap penguins) provide a gauge for assessing ecological change along the WAP. Before the warming in the 20th century, evidence suggests that the previous several centuries had been cooler, with greater and more persistent sea ice coverage in the mid-WAP region, thus providing Ade'lie penguins with a more suitable habitat than at present. Conversely, the increased sea ice in the 18th and 19th centuries along the WAP provided a generally less suitable habitat for Chinstrap penguins. Warming within this century has reversed these habitat conditions, and Ade'lie populations are declining while Chinstrap populations are increasing. These results are consistent with both the modern penguin censuses and the paleoecology records excavated from penguin rookeries. These results, while consistent with one another, leave several important issues unresolved. A key question is whether the recent warming is part of a natural cycle or not. Leventer and co-workers (1996) hypothesized that the increased productivity during the 200-300 year productivity cycles in the WAP was due to warmer atmospheric and sea surface temperatures and a corresponding reduction in sea ice. However, these mechanisms remain unproven. A shift in the Antarctic Convergence Line (ACL) to the south will bring increased storminess, warmer air from the northwest and less sea ice (Ackley and Keliher, 1976), but the global and regional mechanisms influencing the position of the ACL are not fully understood. Similarly, an increase of CDW onto the WAP shelf will lead to enhanced heat flux to the surface, less sea ice, and correspondingly warmer temperatures, but the mechanisms controlling CDW upwelling also are not fully understood. Primary productivity, as suggested by the conceptual models shown in Figures 5a and 5b, is linked to the marginal ice zone, but there are other controls on primary production we have yet to consider. Also, the trophic linkages and their association with sea ice are not fully understood. Sea ice clearly influences primary production and habitat suitability for the Ade'lie and Chinstrap penguins, but the direct and indirect mechanisms underlying our conceptual models need to be refined with future research. One ideal location for that future research appears to be the WAP region whose climate gradient provides the opportunity to investigate ecosystem response to climate change. In particular, the 200-300 year cycles seen in the paleoclimate records of the WAP challenge us to understand these fundamental processes, if we hope to distinguish, now and in the future, natural from anthropogenic causes of climate change. Acknowledgements. The Palmer LTER work is supported by Office of Polar Programs, National Science Foundation (NSF grant no. OPP-96-32763). A supplement to this grant provided focus for this manuscript effort through a workshop held 20-23 August 1997. D.A.'s participation in this work was made possible through NSF grant OPP-9526865. References Cited Ackley, S. F., and Keliher, T. E., 1976. Antarctic sea ice dynamics and its possible climatic effects. AIDJEX Bulletin, 3, 53-76. Ainley, D. G., and Boekelheide, R. J. (Eds.)., 1990. Seabirds of the Farallon Islands: ecology, dynamics and structure of an upwelling-system community. 450 pp., Stanford University Press, Stanford, California. Ainley, D. G., Ribic, C. A., and Fraser, W. R., 1994. Ecologi- cal structure among migrant and resident seabirds of the Sco- tia-Weddell confluence region. Journal of Animal Ecology, 63, 347-364. Baroni, C., and Orombelli, G., 1991. Holocene raised beaches at Terra Nova Bay, Victoria Land, Antarctica. Quaternary Research, 36, 157-177. Baroni, C., and Orombelli, G., 1994. Abandoned penguin rook- eries as Holocene paleoclimatic indicators in Antarctica. Geology, 22, 23-26. Bjo"rck, S., Hakansson, H., Zale, R., Karlen, W., and Jonsson, B. L., 1991. A late Holocene lake sediment sequence from Liv- ingston Island, South Shetland Islands, with palaeoclimatic implication. Antarctic Science, 3, 61-72. Blackburn, N., Taylor, R. H., and Wilson, P. R., 1991. An interpretation of the growth of the Ade'lie penguin rookery at Cape Royds, 1955-1990. New Zealand Journal of Ecology, 15, 117-121. Bradley, R. S., 1985. Quaternary paleoclimatology: methods of paleoclimatic reconstruction. Allen and Unwin, Boston, MA. 472 pp.. Broecker, W. S., 1990. Salinity history of the Northern Atlantic during the last deglaciation. Paleoceanography, 5, 459-467. Broecker, W. S., 1997. Thermohaline circulation, the achilles heel of our climate system: will man-made CO2 upset the cur- rent balance? Science, 278, 1582-1588. CCAMLR, 1992. CEMP standard methods. CCAMLR ecosystem monitor- ing program, Hobart, Australia. Carmack, E. C., 1990. Large-scale physical oceanography of polar oceans. in Polar Oceanography. edited by Smith, W. O., pp. 171-222, Academic press, San Diego. Cavalieri, D. J., Gloersen, P., Parkinson, C. L., Comiso, J. C., and Zwally, H. J., 1997. Observed hemispheric asymmetry in global sea ice changes. Science, 278, 1104-1106. Connell, J. H., 1978. Diversity in tropical rainforests and coral reefs. Science, 199, 1302-1310. Dai, J. C., Thompson, L. G., and Mosley-Thompson, E., 1995. A 485 year record of atmospheric chloride, nitrate and sulfate: results of chemical analysis of ice cores from Dyer Plateau, Antarctic Peninsula. Annals of Glaciology, 21, 182-188. Dansgaard, W., Johnsen, S. J., Clausen, H. B., and Gundestrup, N., 1973. Stable isotope glaciology. Meddelelser om Gronland, 197, 1-53. Denton, G. H., Bockheim, J. G., Wilson, S. C., and Stuiver, M., 1991. Late Wisconsin and early Holocene glacial history, inner Ross embayment, Antarctica. in West Antarctic Ice Sheet Ini- tiative Volume 2: Discipline Reviews, edited by Bindschadler, R. A., p. 55-86, NASA Conference Publication, Washington (DC). Doake, C. S. M., 1982. State of balance of the ice sheet in the Antarctic peninsula. Annals of Glaciology, 3, 77-82. Doake, C. S. M., and Vaughan, D. G., 1991. Rapid disintegration of the Wordie ice shelf in response to atmospheric warming. Nature, 350, 328-330. Domack, E. W., 1990. Laminated terrigenous sediments from the Antarctic Peninsula: the role of subglacial and marine pro- cesses. in Glacimarine Environments: Processes and Sediments. edited by Dowdeswell, J. A., and Scourse, J. D., (Geol. Soc. Spec.Publ, vol. 53), pp. 91-103. Domack, E. W., and Ishman, S. E., 1993. Oceanographic and phys- iographic controls on modern sedimentation within Antarctic fjords. Geological Society of America Bulletin, 105, 1175-1189. Domack, E. W., and McClennen, C. E., 1996. Accumulation of glacial marine sediments in fjords of the Antarctic Peninsula and their use as late Holocene paleoenvironmental indicators. in Foundations for Ecological Research West of the Antarctic Peninsula. edited by Ross, R. M., Hofmann, E. E., and Quetin, L. B., (AGU Antarctic Research Series, vol. 70), pp. 135-154. Domack, E. W., and Williams, C. R., 1990. Fine structure and suspended sediment transport in three Antarctic fjords. Con- tributions to Antarctic Research, 50, 71-89. Domack, E. W., Schere, E., McClennen, C., and Anderson, J., 1992. Intrusion of circumpolar deep water along the Belling- shausen Sea continental shelf. Antarctic Journal of US, 27, 71. Domack, E. W., Mashiotta, T. A., Burkley, L. A., and Ishman, S. E., 1993. 300-year cyclicity in organic matter preservation in Antarctic fjord sediments. in The Antarctic Paleoenvironment: A perspective on Global Change (Part 2). edited by Kennett, J. P., and Warnke, D. A., (Antarctic Research Series), pp. 265-272, American Geophysical Union, Washington,DC. Domack, E. W., Ishman, S. E., Stein, A. B., McClennen, C. E., and Jull, A. J., 1995. Late Holocene advance of the Muller Ice Shelf, Antarctic Peninsula: sedimentological, geochemical and palaeontological evidence. Antarctic Science, 7, 159-170. Emslie, S. D., 1995. Age and taphonomy of abandoned penguin rookeries in the Antarctic peninsula. Polar Record, 31, 409-418. Emslie, S. D., Fraser, W., Smith, R. C., and Walker, W., 1998. Abandoned penguin colonies and environmental change in the Palmer Station region, Anvers Island, Antarctic Peninsula. Antarctic Science, 10, 255-266. Palmer LTER Contribution #158. Enomoto, H., and Ohmura, A., 1990. Influences of atmospheric half-yearly cycle on the sea ice extent in the Antarctic. Journal of Geophysical Research, 95, 9497-9511. Folland, C. K., Parker, D. E., and Kates, F. E., 1984. World- wide marine temperature fluctuations 1856-1981. Nature, 310, 670-673. Fraser, W. R., and Patterson, D. L., 1997. Human disturbance and long-term changes in Ade'lie penguin populations: A natu- ral experiment at Palmer Station, Antarctic Peninsula. in Antarctic Communities, Species, Structure and Survival. edited by Battaglia, B., Valencia, J., and Walton, D. W. H., pp. 445-452, Cambridge University Press, New York. Fraser, W., and Trivelpiece, W., 1996. Factors controlling the distribution of seabirds: winter-summer heterogeneity in the distribution of Ade'lie penguin populations . in Foundations for Ecological Research West of the Antarctic Peninsula. edited by Ross, R. M., Hofmann, E. E., and Quetin, L. B., (AGU Antarctic Research Series, vol. 70), pp. 257-272. Palmer LTER Contribution #081. Fraser, W. R., Trivelpiece, W. Z., Ainley, D. G., and Trivel- piece, S. G., 1992. Increases in Antarctic penguin popula- tions: reduced competition with whales or a loss of sea ice due to environmental warming? Polar Biology, 11, 525-531. Furness, R. W., and Greenwood, J. J. D. (Eds.)., 1993. Birds as monitors of environmental change. 356 pp., Chapman & Hall, London. Hansen, J., and Lebedeff, S., 1987. Global trends of measured surface air temperature. Journal of Geophysical Research, 92, 13345-13372. Harangozo, S. A., 1994. Interannual atmospheric circulation-sea ice extent relationships in the Southern Ocean: an analysis for the west Antarctic Peninsula region, p. 364-367, AMS 6th conference on climate variations, Nashville. Harangozo, S. A., Colwell, S. R., and King, J. C., 1997. An analysis of a 34-year air temperature record from Fossil Bluff (71S, 68W), Antarctica. Antarctic Science, 9, 355-363. Hart, T. J., 1934. On the phytoplankton of the south-west Atlantic and the Bellingshausen Sea 1929-31. (Discovery Reports), University Press, Cambridge. Hofmann, E. E., Klinck, J. M., Lascara, C. M., and Smith, D., 1996. Water mass distribution and circulation West of the Antarctic Peninsula and including Bransfield Strait. in Foun- dations for Ecological Research West of the Antarctic Penin- sula. edited by Ross, R. M., Hofmann, E. E., and Quetin, L. B., (AGU Antarctic Research Series, vol. 70), pp. 61-80. Ishman, S. E., 1990. Quantitative analysis of Antarctic benthic foraminifera: application to paleoenvironmental interpreta- tions. 266 pp., Ph.D. Thesis, Ohio State University, Columbus (OH). Jacobs, S. S., and Comiso, J. C., 1989. Sea ice and oceanic processes on the Ross Sea continental shelf. Journal of Geo- physical Research, 94, 18,195-18,211. Jacobs, S. S., and Comiso, J. C., 1993. A recent sea-ice- retreat west of the Antarctic Peninsula. Geophysical Research Letters, 20, 1171-1174. Jacobs, S. S., and Comiso, J. C., 1997. Climate variability in the Amundsen and Bellingshausen Seas. Journal of Climate, 10, 697-709. Jacobs, S. S., Hellmer, H. H., and Jenkins, A., 1996. Antarctic ice sheet melting in the Southeast Pacific. Geophysical Research Letters, 23, 957-960. Jones, P. D., 1990. Antarctic temperatures over the present century - a study of the early expedition record. Journal of Climate, 3, 1193-1203. Jones, P. D., and Limbert, D. W. S., 1987. A data bank of Antarctic surface temperature and pressure data, Office of Energy Research, Office of Basic Energy Sciences, Carbon Diox- ide Research Division, Washington (DC). Jones, P. D., Raper, S. C. B., and Wigley, T. M. L., 1986. Southern Hemisphere surface air temperature variations: 1851-1984. Journal of Climate and Applied Meteorology, 25, 1213-1230. King, J. C., 1994. Recent climate variability in the vicinity of the Antarctic Peninsula. International Journal of Climatol- ogy, 14, 357-369. Leventer, A., Domack, E. W., Ishman, S. E., Brachfeld, S., McClennen, C. E., and Manley, P., 1996. Productivity cycles of 200-300 years in the Antarctic Peninsula region: understand- ing linkages among the sun, atmosphere, oceans, sea ice, and biota. Geological Society of America Bulletin, 108, 1626-1644. Loeb, V., Siegel, V., Holm-Hansen, O., Hewitt, R., Fraser, W., Trivelpiece, W., and Trivelpiece, S., 1997. Effects of sea-ice extent and krill or salp dominance on the Antarctic food web. Nature, 387, 897-900. Lorius, C., Merlivat, L., Jouzel, J., and Pourchet, M., 1979. A 30,000-yr isotope climatic record from Antarctic ice. Nature, 280, 644-648. Mitchell, B. G., and Holm-Hansen, O., 1991. Observations and modeling of the Antarctic phytoplankton crop in relation to mixing depth. Deep Sea Research, 38, 981-1007. Mosley-Thompson, E., 1992. Paleoenvironmental conditions in Antarctica since AD 1500: ice core evidence. in Climate since A.D. 1500. edited by Bradley, R. S., and Jones, P. D., pp. 572-591, Routledge, London. Murphy, E. J., Clarke, A., Symon, C., and Priddle, J. J., 1995. Temporal variation in Antarctic sea-ice - analysis of a long term fast-ice record from the South Orkney Islands. Deep Sea Research, 42, 1045-1062. Parmelee, D. F., 1992. Antarctic birds. University of Minnesota Press, Minneapolis. 203 pp.. Peel, D. A., 1992. Ice core evidence from the Antarctic Penin- sula region. in Climate since A.D. 1500. edited by Bradley, R. S., and Jones, P. D., pp. 549-571, Routledge, London. Quetin, L. B., and Ross, R. M., 1991. Behavioral and physiolog- ical characteristics of the Antarctic krill, Euphausia superba. American Zoologist, 31, 49-63. Ribic, C. A., and Ainley, D. G., 1988. Constancy of seabird species assemblages: an exploratory look. Biological Oceanog- raphy , 6, 175-202. Rogers, J. C., 1983. Spatial variability of Antarctic tempera- ture anomalies and their association with the southern hemi- sphere atmospheric circulation. Annals of the Association of American Geographers, 73, 502-518. Ross, R. M., Hofmann, E. E., and Quetin, L. B. (Eds.)., 1996. Foundations for Ecological Research West of the Antarctic Peninsula. (AGU Antarctic Research Series, vol. 70), 448 pp. Palmer LTER Contribution #136. Ross, R. M., and Quetin, L. B., 1986. How productive are Antarctic krill? BioScience, 36, 264-269. Ross, R. M., and Quetin, L. B., 1991. Ecological physiology of larval euphausiids, Euphausia superba (Euphausiacea). Memoirs of the Queensland Museum, 31, 321-333. Palmer LTER Contribu- tion #058. Rott, H., Skvarca, P., and Nagler, T., 1996. Rapid collapse of northern Larsen ice shelf, Antarctica. Science, 271, 788-792. SCAR, 1987. SCAR Manual. 61 pp., Scientific Committee on Antarctic Research, Cambridge. Sansom, J., 1989. Antarctic surface temperature time series. Journal of Climate, 2, 1164-1172. Schwerdtfeger, W., 1970. The climate of the Antarctic. in Cli- mates of the Polar Regions. edited by Orvig, S., (World Sur- vey of Climatology, Volume 14), pp. 253-355, Elsevier Publish- ing Company, New York. Schwerdtfeger, W., 1984. Weather and climate of the Antarctic. (Developments in Atmospheric Science, 15), 261 pp., Elsevier Science Pub. Co., New York. Shevenell, A. E., 1996. Record of Holocene paleoclimatic change along the Antarctic Peninsula: evidence from glacial marine sediments, Lallemand Fjord. 49 pp., BA Thesis, Hamilton Col- lege. Shevenell, A. D., Domack, E. W., and Kernan, G. M., 1996. Record of Holocene paleoclimate change along the Antarctic Peninsula: evidence from glacial marine sediments, Lallemand Fjord. in Climate Succession and Glacial Record of the South- ern Hemisphere. edited by Banks, M. R., and Brown, M. J. (Papers and Proceedings of the Royal Society of Tasmania, vol. 130), pp. 55-64. Sievers, H. A., and Nowlin, W. D., 1984. The stratification and water masses at Drake Passage. Journal of Geophysical Research, 89, 10489-10514. Smith, W. O., and Nelson, D. M., 1985. Phytoplankton bloom pro- duced by a receding ice edge in the Ross Sea: spatial coher- ence with the density field. Science, 227, 163-166. Smith, R. C., Baker, K. S., Fraser, W. R., Hofmann, E. E., Karl, D. M., Klinck, J. M., Quetin, L. B., Prezelin, B. B., Ross, R. M., Trivelpiece, W. Z., and Vernet, M., 1995. The Palmer LTER: A long-term ecological research program at Palmer Station, Antarctica. Oceanography, 8, 77-86. Smith, R. C., Stammerjohn, S., and Baker, K. S., 1996. Surface air temperature variations in the western Antarctic peninsula region. in Foundations for Ecological Research West of the Antarctic Peninsula. edited by Ross, R. M., Hofmann, E. E., and Quetin, L. B., (AGU Antarctic Research Series, vol. 70), pp. 105-121. Smith, R. C., Baker, K. S., and Stammerjohn, S. E., 1998. Exploring sea ice indexes for polar ecosystem studies. Bio- Science, 48, 83-93. Palmer LTER Contribution #127. Stammerjohn, S. E., and Smith, R. C., 1997. Opposing Southern Ocean climate patterns as revealed by trends in regional sea ice coverage. Climatic Change, 37, 617-639. Stark, P., 1994. Climatic warming in the central Antarctic Peninsula area. Weather, 49, 215-220. Stuiver, M., and Brazuinas, T. F., 1993. Sun, ocean, climate and atmospheric 14CO2: an evaluation of causal and spectral relationships. The Holocene, 3, 289-305. Taylor, R. H., and Wilson, P. R., 1990. Recent increase and southern expansion of Ade'lie penguin populations in the Ross Sea, Antarctica, related to climatic warming. New Zealand Journal of Ecology, 14, 25-29. Taylor, R. H., Wilson, P. R., and Thomas, B. W., 1990. Status and trends of Ade'lie penguin populations in the Ross sea region. Polar Record, 26, 293-304. Thompson, L. G., Peel, D. A., Mosley-Thompson, E., Mulvaney, R., Dai, J., Lin, P. N., Davis, M. E., and Raymond, C. F., 1994. Climate since AD 1510 on Dyer Plateau, Antarctic Penin- sula: evidence for recent climate change. Annals of Glaciol- ogy, 20, 420-426. Trivelpiece, W., and Fraser, W., 1996. The breeding biology and distribution of Ade'lie penguins: adaptations to environmental variability. in Foundations for Ecological Research West of the Antarctic Peninsula. edited by Ross, R. M., Hofmann, E. E., and Quetin, L. B., (AGU Antarctic Research Series, vol. 70), pp. 273-285. van Loon, H., 1967. The half-yearly oscillations in middle and high southern latitudes and the coreless winter. Journal of The Atmospheric Sciences, 24, 472-483. Vaughan, D. G., and Doake, C. S. M., 1996. Recent atmospheric warming and the retreat of ice shelves on the Antarctic Penin- sula. Nature, 379, 328-330. Weatherly, J. W., Walsh, J. E., and Zwally, H. J., 1991. Antarctic sea ice variations and seasonal air temperature relationships. Journal of Geophysical Research, 96, 15119-15130. Wilson, K. J., Taylor, R. H., and Barton, K. J., 1990. The impact of man on Ade'lie penguins at Cape Hallett, Antarctica. Ecological change and the conservation of Antarctic ecosys- tems. (Proceedings of the 5th SCAR symposium on Antarctic biology), Springer Verlag, Berlin. Woehler, E. J., 1997. Seabird abundance, biomass and prey con- sumption within Prydz Bay, Antarctica, 1980/1981-1992/1993. Polar Biology, 17, 371-383. Zale, R., 1994. Changes in size of the Hope Bay Ade'lie penguin rookery as inferred from Lake Boeckella sediment. Ecography, 17, 297-304. Figure legends Figure 1. Antarctic Peninsula with the positions of geographic locations and features mentioned in the text. Dyer Plateau ice core drill site (70 40'16" S; 64 52'30" W; mean annual temperature -21 degC, altitude 2002 meters), Palmer Station (P) (southern coast of Anvers Island), Faraday Station (F) and the marine sediment core sites (Brialmont Cove, Palmer Deep, Lallemand Fjord) are indicated. The Pal LTER regional sampling grid extends along the peninsula from Alexander to the South Shetland Islands (Smith et al., 1995). The insert of Antarctica shows the relative locations of Ross Island (RI), the Antarctic Peninsula (AP), and the South Orkney Islands (SOI). Figure 2. Faraday Station annual average air temperatures from 1945 to 1997. The solid line is the least-squares regression line with a gradient of 0.050 degC/year and the dotted lines indicate +-1 standard deviation from this line. The linear regression model using the effective number of independent observations (N=12.9) is significant at the 94.0% confidence level (see Smith et al., 1996 for statistical details). Data are from the British Antarctic Survey web site (http://www.nbs.ac.uk/public/icd/data.html). Figure 3. Comparison of paleoenvironmental records from the Antarctic Peninsula region with different temporal resolutions and duration. From top to bottom: (i) Lallemand fjord record (Shevenell et al., 1996) represents last 8000 radiocarbon years and is based upon total organic carbon (TOC %) in marine sediments. Note onset of Neoglacial cooling at around 2500 years BP and carbon minimum at about 400 years BP corresponding to the Little Ice Age. Mid-line is arbitrary and does not represent mean value. (ii) Marine sediment record from Palmer Deep (Leventer et al., 1996) illustrates magnetic susceptibility as a paleoproductivity proxy for the last 3700 radiocarbon years. Note correspondence with Lallemand Fjord record and high frequency ~200 year oscillations in productivity. Record is terminated at about 270 years BP due to incomplete core recovery. Mid-line is arbitrary and is does not represent mean value. (iii) Ice core record for the last 480 years from Dyer Plateau (Thompson et al., 1994) demonstrates reconstructed temperature based upon oxygen isotope analyses (delta18O). Record ends around 1989, the time the core was collected. Note that the last two decades (1970's and 1980's) are among the warmest in the 480-year record. Mid-line is at mean value. (iv) Penguin occupation history for the vicinity of Palmer Station: A = Ade'lie penguin, C = Chinstrap penguin, G = Gentoo penguin. Data on Ade'lie history are from abandoned rookeries and radiocarbon chronology. Data on Chinstrap and Gentoo penguins are from historical records (Parmelee, 1992). (v) Glacial meltwater record from marine sediments in Brialmont Cove for last 100 years based upon 210Pb age determinations (Domack and McClennen, 1996). Peaks in sand abundance represent "warm" melt events in a glacier proximal setting. (vi) Meteorological records of monthly mean temperature for Palmer and Faraday stations (after Domack 1996). Note recent increase in both winter and summer mean temperatures since the 1960's. Mid-line is at 0 C. Figure 4. Twenty-year trends in Chinstrap (open circles) and Ade'lie (closed circles) penguin populations at Arthur Harbor (Palmer Station). For comparison, Ade'lie (+) breeding population trends are also shown for Cape Royds which is located on Ross Island in the Ross Sea (see insert Figure 1) (Taylor and Wilson, 1990; Taylor et al., 1990; Blackburn et al., 1991) (data augmented by P. Wilson pers. comm.). In the WAP region, the two species have exhibited opposite trends, with Chinstrap penguins increasing over 500% since the mid-1970's and Ade'lie penguins decreasing by nearly 25%. Trends in Ross Sea, Cape Royds Ade'lie penguin populations show an increase through the late 1980's and roughly stable populations thereafter. Ade'lie data are given as percent of 1975 counts, whereas Chinstrap are given as percent since 1977 because there were no recorded breeding pairs in 1975 in the Palmer area. Figure 5. Ecological response to climate change. (a) Conceptual model indicating the relationship between the sedimentary record and overlying biological production. (b) Conceptual model indicating the direction of Ade'lie penguin population changes at Ross Sea and Palmer Station in relation to global warming and decline in the frequency of heavy ice years. At Ross Sea, Ade'lie populations will continue to rise until reaching the optimum, then decline. Populations at Palmer Station will continue to decline to extirpation.