Coastal polynyas can be regions of intense ocean-atmosphere heat transfer. In these polynyas, relatively warm ocean is exposed to cold, dry continental air, resulting in high sea-ice production rates. We have calculated ocean-atmosphere heat fluxes over an area of the Mertz Glacier polynya, East Antarctica, from atmospheric data collected in August 1999. Air-temperature and humidity data were measured using a probe extending from a helicopter undercarriage. Flights were made over the most vigorously ice-producing part of the polynya, within 20 km of the coast in southeast Buchanan Bay, both in very strong katabatic winds (20 m s–1) and during calmer conditions (5 m s–1). The total turbulent heat loss from the surface was nearly 575 W m–2 for the windy case and 250 W m–2 for the calmer one, with a ratio of sensible to latent heat in both cases of slightly more than four. These fluxes are in general agreement with other estimates of heat loss from Antarctic polynyas. Strong katabatic winds, sometimes exceeding 40 m s–1, were common in Buchanan Bay and the heat losses during the strong-wind cafe are probably typical of the region. We suggest that this inner part of the Mertz Glacier polynya has a very high ice-production rate.

The Greenland Sea is one of the few places in the World Ocean where deep convection takes place. The convection process is initiated by a density increase originating from rapid cooling and/or a salt flux to the upper layer of the ocean due to brine rejection from ice formation (Rudels 1990; Visbeck and others, 1995). The predominant ice types in the Greenland Sea are frazil/grease ice and pancake ice. A numerical model has been developed relating ice formation and decay of these ice types as observed by the SMMR and SSM/I microwave radiometers and evaluating their contribution to salt redistribution in the Greenland Sea. The model has been used to calculate spatial distribution of the annual integrated net salt flux to the Greenland Sea from ice production and advection for the period 1979−97.

An observational account of research carried out in July-August 1999 shows that grounded iceberg and related fast-ice distributions, and periodic "break-outs" of fast ice (in winter as well as at other times), have an important impact on the size and behaviour of the Mertz Glacier polynya, East Antarctica, and a smaller polynya to the east. Analysis of satellite and in situ data shows that a semi-constant "stream" of thick broken-out fast ice and other large floes from the east extends westwards from north of the glacier terminus to form a compact barrier to the net west-northwesterly export of ice formed in the polynya. An annual fast-ice promontory to the west further narrows the outlet path. As a result of this and high ice-production rates, the polynya periodically "back-fills", significantly reducing the open-water area present. Intervening "flush-outs" by synoptic storm events clear the polynya region to some extent before it back-fills again. This cycle continued from mid-March until early October in 1999, when a significant change in the regional ice drift occurred. A preliminary comparison with data from 1998 indicates that the timing and magnitude of the processes may vary interannually. Similar morphological features were also observed in 1963 (on a declassified photoreconnaissance satellite image).

In July-September 1999, an extensive oceanographic survey (87 conductivity-, temperature-and depth-measuring stations) was conducted in the Mertz Glacier polynya over the Adélie Depression off the Antarctic coast between 145° and 150° E. We identify and describe four key water masses in this polynya: highly modified circumpolar deep water (HMCDW), winter water (WW), ice-shelf water (ISW) and high-salinity shelf water (HSSW). Combining surface velocity data (from an acoustic Doppler current-profiler) with three hydrographic sections, we found the HMCDW to be flowing westward along the shelf break (0.7 Sv), the WW and HSSW flowing eastwards underneath Mertz Glacier (2.0 Sv) and that there was a westward return flow of ISW against the continent (1.2 Sv). Using a simple box model for the exchanges of heat and fresh water between the principal water masses, we find that the polynya was primarily a latent-heat polynya with 95% of the total heat flux caused by sea-ice formation. This heat flux results from a fresh-water-equivalent sea-ice growth rate of 4.9−7.7 cm d−1 and a mass exchange between HMCDW and WW of 1.45 Sv The inferred ocean heat flux is 8−14 W m−2 and compares well with other indirect estimates.

Ship and iceberg drifts, along with data from modern satellite-tracked drifting buoys, portray generally northward flow in the southwest Ross Sea. Less is known, however, about the ocean variability in this region. Here we use Defense Meteorological Satellite Program (DMSP) Operational Linescan System (OLS) imagery to describe a cyclonically rotating gyre in Terra Nova Bay, Antarctica (75° S, 165° E). The image sequence is taken from an exceptionally clear, calm period during winter 1998. Surface currents, derived from tracking drifting ice with a maximum cross-correlation algorithm, were found to be ∼ 20 to 30 cm s−1 Based on scaling arguments, the observed currents appear to be best explained as an oceanic response to a rapidly changing, offshore, katabatic windjet.

Monitoring ice motion provides insight into the relative contributions of atmospheric and oceanic forcing to polynya dynamics In this study ice kinematics in the North Water (NOW) region, northern Baffin Bay Canadian Arctic are determined using time-sequential RADARSAT-1 ScanSARWide imagery (from the period January-December 1998), processed by the automated ice-tracking algorithm (Tracker) currently used at the Canadian Ice Service (CIS). A case-study is then carried out on the Smith Sound region from January till the end of june 1998 to monitor polynya formation, maintenance and disintegration. This has two objectives: (1) to evaluate satellite ice-motion products as a means of better understanding the relative contributions of latent- and sensible-heat mechanisms responsible for the formation and behaviour of the NOW polynya, and (2) to study the influence of winds on regional sea-ice motion. Derived ice-motion maps were validated using in situ ice beacons deployed onto floes in the NOW region during the 1998 ship-based science experiment. Tracker-derived displacement and directional regressions were 0.93 and 0.79, respectively, with a total standard error of 3.6 km in magnitude and 38.8° in direction. Analysis of monthly mean ice-motion maps shows that a significant export of ice occurs from the polynya. A comparison of weekly-averaged ice motion with the mean wind-field data suggests that ice export in Smith Sound is influenced by synoptic-scale atmospheric pressure systems.

A coupled sea-ice-ocean model of the Weddell Sea, Antarctica, has been developed as part of the Bremerhaven Regional Ice-Ocean Simulations (BRIOS) project. It is based on the s-Coordinate Primitive Equation ocean Model (SPEM) and a dynamic-thermodynamic sea-ice model with viscous-plastic rheology which also provides the thermohaline forcing at the base of the Antarctic ice shelves. Model runs are forced with wind, cloudiness, temperature and precipitation fields of the European Centre for Medium-range Weather Forecasts and U.S. National Centers for Environmental Prediction re-analyses. Model results show good agreement with observations of ice extent, thickness and drift. Water-mass properties and the large-scale circulation are in good agreement with observations. Fresh-water fluxes from sea-ice formation as well as from ice-shelf basal melting and from precipitation are computed and compiled to the fresh-water budget of the Weddell Sea. Supporting estimates based on hydrographic observations, model results indicate that fresh-water loss due to sea-ice formation and export (34mSv) is roughly balanced by ice-shelf basal melting (9 mSv) and net precipitation (19 mSv). Furthermore, sea-ice formation appears to be a necessary condition for bottom-water production in the Weddell Sea.

The Ronne polynya formed in the Weddell Sea, Antarctica, during the period November 1997−February 1998 to an extent not seen previously in the 25 years of all-weather satellite observations. The vessel HMS Endurance traversed the polynya region and took sea-ice, physical oceanographic and meteorological measurements during January and early February 1998. These observations, together with satellite imagery and weather records, were analyzed to determine the causes of the anomalous condition observed and to provide comparisons for numerical modeling experiments. The polynya area, analyzed from satellite imagery, showed a linear, nearly constant, increase with time from mid-November 1997 through February 1998. It had a maximum open-water area of 3 × 105 km2 and extended 500 km north of the Ronne Ice Shelf (at 76° S) to 70° S. The ice and snow structure of floes at the northern edge of the polynya showed the ice there had formed in the previous mid- to late winter (October 1997 or earlier) and had been advected there either from the eastern Weddell Sea or from the front of the Ronne Ice Shelf. Analyses of the wind fields showed anomalous spring-summer wind fields in the polynya year, with a strong southerly to southwesterly component compared to the mean easterly winds typical of summer conditions. These southerly wind conditions, in both magnitude and direction, therefore account for the drift of ice northward. The predominant summer easterly winds usually fill the southern Weddell Sea with ice from the east, and the high-albedo surfaces reflect the solar radiation, preventing warming of the surface ocean waters and consequent sea-ice melt. Instead, high incident solar radiation from November 1997 to February 1998 was absorbed by the open water, rather than being reflected, thereby both melting ice and preventing ice formation, and thereby sustaining the polynya. We conclude that open-water-albedo feedback is necessary to allow the observed polynya formation, since similar drift conditions prevail in winter (arising from southerly winds also) and usually result in extensive new ice formation in front of the Ronne Ice Shelf. The strong southerly winds therefore have quite opposing seasonal effects, leading to high ice production in winter as usually found, and extensive open water if they occur in spring and summer, as seen in this atypical event in 1997/98. In this case, the atypical southerly winds may be associated with an El Niño-Southern Oscillation (ENSO)-induced atmospheric circulation pattern.

The Storfjorden (Svalbard) ice cover has been observed via European remote-sensing satellite (ERS-2) synthetic aperture radar (SAR) over two winters from 1997 to 1999 that showed a high interannual variability in atmospheric forcing. This has resulted in different scenarios in the formation of the ice cover. We reconstruct the ice-cover evolution with a simple ice-drift and polynya model consistent in parameterization for both winters and with ERS-2 SAR and ground-truth observations. The observations showed that, during the rather cold winter of 1997/98, a fast-ice cover developed over the whole northern part of Storfjorden that closed the ice advection through two sounds. A polynya of up to 6000 km2 developed persistently under northerly winds. The mild winter of 1998/99 showed a different pattern in the Storfjorden ice cover: the fast-ice extent was reduced to the coastal region, so ice motion was more important than in the previous winter. A larger thin-ice and open-water area over the whole fjord resulted in lower total ice production, but relocated the strong ice production sites to shallower waters. Thus, freezing in a large region of shallow waters in the north of Storfjorden in 1998/99 produced denser waters than freezing in a smaller polynya over the deeper centre of Storfjorden during the colder winter 1997/98.

The transformation of ocean water masses at high latitudes is closely related to the freezing and melting processes during the year. Downward salt fluxes from brine rejection during freezing increase the salinity and density of the water column underneath. Fresh-water input from river run-off and melting of sea ice reduces the density, mainly of the surface layer. Hydrographic profiles collected in Storfjorden, Svalbard, in spring and summer, show the strong seasonal and interannual variability of the water masses. Using, in addition, data from moorings, a ship-borne acoustic Doppler current profiler and drifting buoys, and results from models of ice drift, polynya evolution, ice formation and convection processes during freezing, we document the seasonal water-mass transformation and try to explain its interannual variability. The advection of ice and water through the two northern sounds and over the sill in the south of the fjord is examined. The release of brine-enriched bottom water over the sill must be balanced by advection from the Barents Sea in the upper layers. The interannual variability of the brine-enriched bottom layer is very high, and higher salinities are observed in a milder winter. The density anomaly resulting from freezing might depend more on the ice cover and geographical position of the polynya than on the total atmospheric forcing during winter.

In order to extend diagnoses of recent sea-ice variations beyond the past few decades, a century-scale digital dataset of Arctic sea-ice coverage has been compiled. For recent decades, the compilation utilizes satellite-derived hemispheric datasets. Regional datasets based primarily on ship reports and aerial reconnaissance are the primary inputs for the earlier part of the 20th century. While the various datasets contain some discrepancies, they capture the same general variations during their period of overlap. The outstanding feature of the time series of total hemispheric ice extent is a decrease that has accelerated during the past several decades. The decrease is greatest in summer and weakest in winter, contrary to the seasonality of the greenhouse changes projected by most global climate models. The primary spatial modes of sea-ice variability diagnosed in terms of empirical orthogonal functions, also show a strong seasonality. The first winter mode is dominated by an opposition of anomalies in the western and eastern North Atlantic, corresponding to the well-documented North Atlantic Oscillation. The primary summer mode depicts an anomaly of the same sign over nearly the entire Arctic and captures the recent trend of sea-ice coverage.

Sea-ice fluctuations in the Sea of Okhotsk and the Bering Sea during the winters of 1992−99 were investigated by using the Special Sensor Microwave/ Imager dataset and a new ice-property retrieval algorithm This algorithm can distinguish between ice types such as fast ice floes, young ice and new ice, in an area covered by concentrations of >80% ice, and also has improved display resolution because it uses one of the 85 GHz channels. The ice thicknesses derived from the ice-thickness parameter of the new algorithm were compared with ship-based ice-thickness measurements, and were assumed to be 1−10, 11−34, 35−85 and 86−120 cm for new ice, young ice, floes (first-year ice) and fast ice, respectively. The results showed that ice volume can be small even if the ice area is large, due to thinness of the ice (e.g. in 1999 in the Sea of Okhotsk). A significant out-of-phase response, i.e. ice volume is larger in the Sea of Okhotsk when ice volume is smaller in the Bering Sea, was observed. The period of this see-saw showed two different time-scales, which were short (1 week) and long (2−4 weeks).

Recent observations of a decreasing ice extent and a possible thinning of the ice cover in the Arctic make it imperative that detailed studies of the current Arctic environment are made, especially since the region is known to be highly sensitive to a potential change in climate. A continuous dataset of microwave, thermal infrared and visible satellite data has been analyzed for the first time to concurrently study in spatial detail the variability of the sea-ice cover, surface temperature, albedo and cloud statistics in the region from 1987 to 1998. Large warming anomalies during the last four years (i.e. 1995−98) are indeed apparent and spatially more extensive than previous years. The largest surface temperature anomaly occurred in 1998, but this was confined mainly to the western Arctic and the North American continent, while cooling occurred in other areas. The albedo anomalies show good coherence with the sea-ice concentration anomalies except in the central region, where periodic changes in albedo are observed, indicative of interannual changes in duration and areal extent of melt ponding and snow-free ice cover. The cloud-cover anomalies are more difficult to interpret, but are shown to be well correlated with the expected warming effects of clouds on the sea-ice surface. The results from trend analyses of the data are consistent with a general warming trend and an ice-cover retreat that appear to be even larger during the last dozen years than those previously reported.

The interaction between sea ice and glaciers has been studied for the floating tongue of Nioghalvfjerdsfjorden glacier, northeast Greenland (79°30’N, 22° W). Information from glacial geological studies, expedition reports, aerial photographs and satellite imagery is used to document the glacier front position and fast-ice conditions on millennial to decadal time-scales. The studies indicate that the stability of the floating glacier margin is dependent on the presence of a protecting fast-ice cover in front of the glacier. In periods with a permanent fast-ice cover, no calving occurs, but after fast-ice break-up the glacier responds with a large calving activity, whereby several years of accumulated glacier-ice flux suddenly breaks away. Climate-induced changes of sea-ice conditions in the Arctic Ocean with seasonal break-up of the near-shore fast ice could lead to disintegration of the floating glaciers. The present dominant mass loss by bottom melting would then to a large extent be taken over by grounding-line calving of icebergs. The local influx of fresh water from the north Greenland glaciers to the sea would be reduced and the local iceberg production would increase.

Variability of the sea-ice cover (extent) in the Northern Hemisphere (Arctic and subpolar regions) associated with the Arctic Oscillation (AO) is investigated using historical data from 1901 to 1997. A principal-component analysis (empirical orthogonal functions (EOFs)) was applied to sea-ice area (SIA) anomalies for the period 1953−95. The leading EOF mode for the SI A anomaly shows an in-phase fluctuation in response to the AO and is called the Arctic sea-ice oscillation (ASIO). Arctic sea ice experiences seasonal variations that differ in timing and magnitude. Four types of seasonal variation are identified in the Arctic sea ice, and are superimposed on long-term interannual to decadal variability. Consistent with the total Arctic SIA anomaly eight regional SIA anomalies have shown significant in-phase decrease (downward trend) since 1970, possibly part of a very long-term (century) cycle. Thus, it is recommended that SIA anomalies in the sensitive seasons be used to better capture interannual, interdecadal and longer (century) variability. Major decadal and interdecadal time-scales of SIA anomalies are found at 12−14 and 17−20 years. In the Sea of Okhotsk, a century time-scale is evident. The reduction rate (negative trend) of the total Arctic sea-ice cover in the last three decades is −4.5% per decade, with the summer rate being the highest (-10.2% per decade). The contribution to this total reduction varies from region to region, with sea-ice cover in the Greenland and Norwegian Seas experiencing the highest reduction rate of −20.2 % per decade.

The western Antarctic Peninsula (WAP) region has experienced a statistically significant warming trend during the past half-century. In addition, a statistically significant anticorrelation between air temperatures and sea-ice extent, as determined from satellite passive-microwave data during the past two decades, has been observed for this region. Consistent with this strong coupling, sea-ice extent in the WAP area has trended down during this period of satellite observations. Further, much of the variability in both air temperature and sea ice in the WAP region has been shown to be influenced by contrasting maritime (warm, moist) and continental (cold, dry) climate regimes. As part of the Palmer Long Term Ecological Research program, the ecological influence of these trends and variability is being studied, and effects have already been demonstrated at all trophic levels. Here we extend earlier observations to include the past decade and focus on the annual cycles of air temperature and sea-ice extent for the past few years, with the aim of placing these recent observations within the context of changes seen in the longer-term records. The more recent years have seen an increasing maritime influence in the WAP region, with corresponding effects on the marine ecosystem.

A box model combining the different physical components of the coupled ocean, atmosphere, sea-ice and land-ice climate system is used to study the possible role of sea ice in the glacial-interglacial oscillations. In this paper we investigate the teleconnection between the two hemispheres, and suggest a mechanism which can explain the observed Southern Ocean sea-ice variability during glacial-interglacial cycles. The teleconnection mechanism involves changes in the temperature of the water transported from the Northern to the Southern Hemisphere by the thermohaline circulation (THC), rather than the amplitude of the THC as has often been suggested. The sequence of events proposed by this model indicates that it would be especially interesting to obtain accurate information on the relative timing of sea-ice and land-ice variations. Fuller general circulation models and new sea-ice proxy data are needed to quantify the relative role of the proposed teleconnection mechanism.

The Queen Elizabeth Islands of the Canadian Arctic Archipelago exhibit one of the most complex sea-ice regimes in the Northern Hemisphere. Time series of minimum monthly passive-microwave sea-ice area (1979−98), minimum sea-ice extent, melting degree-days (1961−98) and minimum sea ice from the new Canadian Ice Service digital database (1968−98) are examined. The extreme nature of the amount of sea-ice melt in the summers of 1998 and 1962 is evident in these time series. The 38 year record of minimum ice, to date, shows no significant trend. Details of the sea-ice behavior during summer 1998 were then examined within 13 individual sea-ice regimes. The multi-year fast-ice plugs in both Sverdrup Channel and Nansen Sound broke up and became truly mobile in 1998. Discussion focuses on the areas surrounding the multi-year plugs, relating sea-ice conditions to weather. Results emphasize the importance of the timing of synoptic events in combination with strong thermal preconditioning in determining the sea-ice conditions in this area during summer 1998.

The Q.-flux Goddard Institute of Space Studies (GISS) global climate model, in which an atmospheric general circulation model is coupled to a mixed-layer ocean with specified horizontal heat transports, is used to simulate the transient and equilibrium climate response to a gradual increase of carbon dioxide (1% per year increase of CO2 to doubled CO2). The results indicate that the current GISS model has a high sensitivity with a global annual warming of about 4°C for doubled CO2 . Enhanced warming is found at higher latitudes near sea-ice margins due to retreat of sea ice in the greenhouse experiment. Surface warming is larger in winter than in summer, in part because of the reductions in ice cover and thickness that insulate the winter atmosphere from the ocean. The annual mean reduction of sea-ice cover due to doubled CO2 is about 30% for the Northern Hemisphere. The CO2 experiment has a 70% reduction of sea-ice area and 55% thinning of ice in August in the Northern Hemisphere. Noticeable reduction of sea-ice cover has been found in both historical records and satellite observations. The largest reduction of simulated sea-ice extent occurs in summer, consistent with observations.

A global atmosphere-ocean-sea-ice general circulation model (GCM) is used in simulations of climate with greenhouse gas concentrations and sulfate aerosols prescribed from observational data (1870−1995) and future projections (1995−2100). Simulations that include the variability in solar flux from 1870 through 1995 are also performed. The variation in solar flux of ± 2 W m−2 produces a global temperature change of ± 0.2°C in the model. The more recent simulated warming trend produced by increasing greenhouse gases exceeds this solar-flux warming, although the solar flux contributes to some of the simulated present-day warm temperatures. The future increases in greenhouse gases produce an increase in global temperature of 1.2°C over 70 years, with significant decreases in Arctic ice thickness and area. The model exhibits an atmospheric pressure mode similar to the Arctic Oscillation, with different correlation indices between the North Atlantic and North Pacific pressure anomalies.