Influence of the Antarctic Ice Sheet on the Southern Ocean

1. Resumé

We wish to understand the influence of the Antarctic Ice Sheet ice sheet on the Southern Ocean during the past, present, and possible future climatic regimes. To achieve this aim we need to know the effects of melting and freezing at the base of ice shelves, and the oceanographic impact of ice bergs that might melt some distance from the coast.

The controls on the production of High Salinity Shelf Water (HSSW), the source water for much of the present-day sub-ice shelf circulation, will be studied by instrumenting the Ronne shorelead. Local meteorological conditions recorded by AWS's will be coupled with sea-surface temperature and ice extent data to provide estimates of HSSW fluxes. The degree to which the HSSW is modified prior to its arrival at the base of the ice shelf will be studied using sub-ice shelf instrumentation, as will the properties of the resulting Ice Shelf Water (ISW) plumes.

The sources of energy for vertical mixing will be investigated by analyzing the results of a tidal model for the ice shelf, and by direct sampling of the internal wave regime existing beneath the ice shelf. Parameterization of the distribution of mixing energy is necessary for realistic models of the sub-ice shelf circulation. The test of our understanding of the impact of the ice shelves on the Southern Ocean will be our ability to develop and apply numerical models that reproduce the observations. These modelling efforts will include a tidal model; modelling of the evolution of ISW plumes; modelling HSSW production, and, in collaboration with external groups, applying a large scale ocean model to the sub-ice shelf domain. Pacific-sector ice shelves will also be studied to help test models predicting the impact if warmer water were to reach the cavity beneath the larger Antarctic ice shelves.

Additional work will include an investigation into the way in which the saline ice deposited beneath some Antarctic ice shelves evolves with time to see whether these deposits contain paleo-oceanographic information.

Synthetic Aperture Radar (SAR) data from ERS-1 will be used to generate quasi-synoptic ice-berg counts for the Southern Ocean. This, coupled with an analysis of the local oceanographic impact of an ice-berg, will help us quantify the impact of the total Southern Ocean ice-berg load.

2. Scientific background

The Southern Ocean is modified by the Antarctic Ice Sheet in two ways. The direct transfer of latent heat from the ocean to melt the ice results in a cooling and freshening of the ocean, and the indirect cooling that is mediated by the atmosphere. This project is concerned with the direct ice-ocean interaction, that is, the melting of ice-bergs, and the melting and freezing at the base of ice shelves. Present estimates suggest that ice-berg melting accounts for nearly four times as much ice sheet loss as ice shelf melting (Jacobs et al, 1992). There are two principal reasons why the impact of the two sources of meltwater will be different. First, melting of ice-bergs takes place at a much reduced depth compared with ice shelves. The ice bergs with the greatest draft calve from ice shelves, and so will be no thicker than the thinnest part of that ice shelf, namely the ice front. According to Jacobs et al, nearly 70% of all ice shelf melting takes place more than 100 km from the ice front, and so will be from the deepest parts of the ice shelves. 60% of this melting comes from the interior region of the thickest Antarctic ice shelf: Ronne-Filchner Ice Shelf. The second important difference concerns the geographical site of the melting. Ice shelves melt over the continental shelf, while most ice-berg melting occurs north of the shelf edge, often in the circumpolar current.

The importance of the depth at which melting takes place arises out of the pressure dependence of the freezing point of seawater (Jenkins (1991)). At a temperature of -1.9C and a salinity of greater than 34.7, HSSW is dense but also about one degree warmer than the freezing point of seawater at the typical Ronne Ice Shelf grounding line depths of 1300 m. If HSSW reaches the grounding line unmodified by the ice shelf it is clearly capable of sustaining high melt rates. This leads to the generation of a buoyant plume of ISW. As the ISW plume ascends the ice shelf base the pressure-dependent freezing point rises and can ultimately reach the local temperature of the plume. Ice precipitated from the plume then accumulates at the ice shelf base. Under the central Ronne Ice Shelf, for example, the saline ice deposits can account for more than half the total ice depth (Thyssen et al 1993). Ice cores penetrating the saline ice have been retrieved from the central Ronne Ice Shelf (Oerter et al 1992a; 1992b) and from Amery Ice Shelf (Morgan, 1972) and its structure might contain information about the oceanographic regime that existed when it was deposited. With the present-day hydrography in the Weddell Sea the ISW plume finally emerges from beneath the ice shelf, potentially supercooled, at mid-depth (Foldvik et al, 1985). ISW flowing down the continental slope from the Filchner Depression has been observed to reach great depth before strong mixing occurs (Foldvik and Gammelsrød, 1988). Since at greater depth the Warm Deep Water (WDW) with which it is mixing is cooler, smaller volumes of ISW can yield the same flux of Weddell Sea Bottom Water (WSBW), and ultimately Antarctic Bottom Water (AABW). It is not yet clear how important AABW production via the ISW mechanism is to total AABW productivity, compared with the shelf-break mixing process discussed by Foster et al (1987).

The precursor to ISW, HSSW, is formed in the shorelead north of the ice front. A prolific source of HSSW appears to be at the western end of Ronne Ice Front, above the Ronne Depression (Foldvik et al 1985; Grumbine 1991). Here, the shorelead is maintained as a latent heat polynya by the southerly barrier winds that persist in that region (Schwerdtfeger, 1984). It is not clear how much of the HSSW flows into the Ronne Depression, beneath the ice shelf, and how much flows north under the influence of the Weddell gyre and the local wind stress.

The oceanographic conditions on the continental shelf of the Pacific sector are quite different to those of the colder Atlantic sector and provide an important comparison. In fact, the Antarctic Peninsula acts as a barrier between two extremes of the Antarctic coastal marine environment. In deep enough regions of the Bellingshausen and Amundsen seas, the warm Circumpolar Deep Water (CDW) is able to penetrate on to the continental shelf, and allow water some 3C warmer than the in situ freezing point to come into contact with the base of the ice shelves that exist there. This results in the relatively high average basal meltrate of 2.1 m yr-1 for George VI Ice Shelf (Potter and Paren 1985), and more than 10 m yr-1 for Pine Island Glacier. By the time CDW reaches the continental shelf north of Ronne-Filchner Ice Shelf, it has been transformed to Warm Deep Water (WDW), which is slightly fresher, and cooler. WDW, modified by admixture of Winter Water (WW) has been observed to intrude occasionally onto the continental shelf as Modified Warm Deep Water (MWDW) with temperatures as high as -1.3C (Foldvik et al 1985). At present, this water is thought never to penetrate far beneath the ice shelf, and certainly not as far as the deep grounding line.

Until recently, the sparsity of ice-berg observations have prevented a reliable evaluation of the effect of ice-bergs on the Southern Ocean as a whole (Orheim, in press). There are few observations of the local oceanographic impact of ice-berg melt, and the theoretical analyses have yet to be confirmed. However, the indications so far are that the impact is small, except possibly on Antarctic Surface Water (Jacobs et al 1979).

3. Research and methodology

The question that this project will address is "How do ice shelves and ice-bergs influence the Southern Ocean at the present time, during past climatic periods, and during possible future climates?". We propose to make progress with this question by study in the following areas: formation of the principal water mass (HSSW) interacting with Ronne-Filchner Ice Shelf; mechanisms by which water masses are transported to the base of ice shelves; interaction between ice shelves and the ambient water; properties of the products of that interaction; the distribution and size of ice-bergs; and the local impact of ice-berg melt.

The principal source water mass for interactions between Ronne Ice Shelf and the southern Weddell Sea is HSSW. We propose a study of HSSW-production in the shorelead north of Ronne Ice Front. This will involve deploying an AWS network on the ice front and the Peninsula to provide local meteorological forcing data. ARIES data will give the areal extent of the lead, and satellite-derived SST's will help indicate when ice is being formed during the late summer and autumn periods. The principal water mass impinging on the ice shelves in the eastern Pacific Sector is CDW. We propose a cruise programme to study the bathymetry and hydrography between Ronne Entrance and the continental shelf break, probably using HMS Endurance as the platform. When coupled with data from cruises of the Nathaniel B Palmer, this will help build a picture of the interannual variability of the hydrography, and assist in our understanding of the reasons for CDW intruding so far on to the continental shelf in the present day.

To understand the way Ronne Ice Shelf modifies the HSSW that flows beneath it we must determine the history of a parcel of HSSW: what route does it take and how is it modified prior to its main interaction with the ice shelf? What are the principal controls on the flow and evolution of the resulting ISW plumes as they ascend the ice shelf base? We propose to deploy instrument strings through hot-water-drilled access holes at a site west of Korff Ice Rise, and, in a collaboration with AWI, at locations south of Berkner Island, and between Berkner Island and Henry Ice Rise. The resulting datasets, together with those collected during past drilling seasons, will be used to help control a large scale iso-pycnic model that will be applied to the sub-ice shelf cavity in a collaboration with David Holland at the Hadley Centre.

The physics of the interaction between the ice-shelf base and the underlying water must be understood before the sub-ice shelf environment can be realistically modelled. Tides probably play an important role in providing energy for mixing the dense, relatively warm HSSW up though the water column to enable it to interact with the ice-shelf base (eg MacAyeal (1984), Sheduikat and Olbers (1990)). In a collaboration with the Proudman Oceanographic Laboratory the tides of the Weddell Sea, and the Ronne-Filchner Ice Shelf in particular, will be modelled and their contribution to vertical mixing studied and quantified. Another source of mixing energy is internal waves, and these will be studied using long-term datasets from instrument moorings. The evolution of an individual plume of ISW will be studied in detail, both for the Ronne and other ice shelves, particular attention being paid to the controls on the deposition of ice precipitating from the ascending potentially supercooled water. Ice cores retrieved from the thick deposits of saline ice found at the base of Ronne Ice Shelf might contain paleo-oceanographic information. In a collaboration with the Department of Applied Mathematics and Theoretical Physics at the University of Cambridge we propose to use numerical models and laboratory experiments to investigate the post-depositional evolution of the saline ice, and to determine whether anything can be learnt about the oceanographic conditions that prevailed at the time of deposition.

To help validate predictions of what might happen if climatic shifts bring warmer water under Ronne-Filchner Ice Shelf, our models of ice-ocean interaction will be applied to the ice shelves of the warmer Pacific sector. An ability to reproduce present-day conditions under the southern George VI Ice Shelf, for example, will lend confidence to our predictions. A glacial maximum is the other climatic extreme. During these periods the Antarctic Ice Sheet is thought to have been grounded out to the continental shelf break. We propose to develop a coupled ice shelf/ocean model to study the way in which the influence of the ice sheet on the ocean changed as the ice sheet floated off the continental shelf to form the ice shelves.

The influence of ice-bergs will be studied by analyzing the near-field oceanographic impact of the presence of an ice-berg in the various Antarctic oceanographic regimes, and then integrating that effect over the Southern Ocean as a whole. The distribution of ice-bergs, a necessary input to the study, will be determined by routine analysis of SAR imagery.

4. Wider implications

The Weddell Sea is Antarctica's most productive factory for AABW, a water mass that helps drive the abyssal circulation of the World Ocean and one that helps cool and oxygenate everywhere it upwells. There is little doubt that ISW formed beneath the Ronne-Filchner system of ice shelves plays a role in AABW production, but the importance of that role is uncertain. One of the primary aims for ICD5 is to determine the present day importance of ISW to AABW production, to predict how that contribution changes during a glacial cycle, and how it might change in the future.

5. References

Foldvik, A. and T. Gammelsrød. 1988. Notes on Southern Ocean hydrography, sea-ice and bottom water formation. Palaeogeogr. Palaeoclimatol. Palaeoecol. 67, 3-17.

Foldvik, A., T. Gammelsrød, N. Slotsvik and T. Tørresen. 1985. Oceanographic conditions on the Weddell Sea Shelf during the German Antarctic Expedition 1979/80. Polar Res. 3, 209-226.

Foster, T.D., A. Foldvik and J.H. Middleton. 1987. Mixing and bottom water formation in the shelf break region of the southern Weddell Sea. Deep-Sea Res. 34, 1771-1794.

Grumbine, R.W. 1991. A model of the formation of high-salinity shelf water on polar continental shelves. J. Geophys. Res. 96, 22,049-22,062.

Jacobs, S.S., A.L. Gordon and A.F. Amos. 1979. Effect of glacial ice melting on the Antarctic surface water. Nature 277, 469-471.

Jacobs, S.S, H.H. Helmer, C.S.M. Doake, A. Jenkins and R.M. Frolich. 1992. Melting of ice shelves and the mass balance of Antarctica. J. Glaciol. 38, 375-387.

Jenkins, A. 1991. A one-dimensional model of ice shelf-ocean interaction. J. Geophys. Res. 96, 20671-20677.

MacAyeal, D.R. 1984. Thermohaline circulation below the Ross Ice Shelf: a consequence of tidally induced vertical mixing and basal melting. J. Geophys. Res. 89, 597-606.

Morgan, V.I. 1972. Oxygen isotope evidence for bottom freezing on the Amery Ice Shelf, Nature 238, 393-394.

Oerter, H., C. Drücker, J. Kipfstuhl, U. Nixdorf and W. Graf. 1992. The Filchner IV campaign and the 320 m deep ice core B15. Filchner Ronne Ice Shelf Programme Report 6, 47-53.

Oerter, H, J. Kipfstuhl, J. Determan, H. Miller, A. Minikin and W. Graf. 1992. Ice-core evidence for basal marine shelf ice in the Filchner-Ronne Ice Shelf. Nature 358, 399-401.

Orheim, O.O. (In Press) Ice-berg calving rates and the mass balance of Antarctica. Ann. Glaciol.

Potter, J.R. and J.G. Paren. 1985. Interaction between ice shelf and ocean in George VI Sound, Antarctica. In Oceanology of the Antarctic Continental Shelf. Ed. Jacobs, S.S. American Geophysical Union, Washington DC. Antarctic Research Series, 43, 35-58.

Scheduikat, M. and D.J. Olbers. 1990. A one-dimensional mixed layer model beneath the Ross Ice Shelf with tidally induced vertical mixing. Ant. Sci. 2, 29-42.

Schwerdtfeger, W. 1984. Weather and climate of the Antarctic. In Developments in Atmospheric Acience, 15. Elsevier, Amsterdam, pp 261.

Thyssen, F, A. Bombosch and H. Sandhäger. 1993. Elevation, ice thickness and structure mark maps of the central part of the Filchner-Ronne Ice Shelf. Polarforschung 62, 17-26.