Skip navigation


You are in:  Home » Publications » Reports » Report 13

SCAR Report No. 13, November 1996 

SCAR Group of Specialists on Global Change and the Antarctic (GLOCHANT)

Report of the 1995 meeting of the GLOCHANT Task Group 3 on
Ice Sheet Mass Balance and Sea-Level (ISMASS)Chamonix, France, 17 September 1995

Members of the ISMASS Group present: Professor C R Bentley (Chairman, USA), Dr C S M Doake (UK), Dr I D Goodwin (GLOCHANT coordinator, Australia), Dr P Holmlund (Sweden), Dr T H (J) Jacka (Australia), Dr B K Lucchitta (USA), Dr H Oerter (Germany), Professor G Orombelli (Italy).

Apologies: Dr F Nishio (Co-chairman, Japan), Dr Q Dahe (China).

Other participants: Professor W F Budd (Australia), Ms C Richardson (Sweden).

1. Introduction

The overall theme of the meeting was, primarily, the coordination of ice radar sounding surveys of the grounding zone of the entire Antarctic Ice Sheet, and secondarily, the coordination of surface mass balance measurements and surface ice velocities. No formal agenda was tabled.

The Chairman opened the meeting with a review and history of the development of SCAR GLOCHANT. He outlined the overall aim of Ice Sheet Mass Balance Programme (ISMASS), to evaluate the present Antarctic mass balance and its affect on global sea-level. I D Goodwin expressed a need for the group to establish a 3–5-year plan with

milestones to enable a cooperative international approach to the execution of this aim.

The national representatives were then invited to make presentations of their country's present and future plans for radio-echo sounding (RES) and overall mass balance measurements. Each nation was previously assigned a sector of Antarctica to concentrate their RES surveys. The sectors were identified by the Chairman from the paper by Bentley and Giovinetto (1991) and are referred to alphabetically.

2. National Reports

2.1 Germany (Presented by Dr. Hans Oerter)

The Chairman's proposed that Germany could contribute by carrying out ice thickness measurements in two sectors, namely JIK and KKII, ie the southern and eastern boundary of Ronne–Filchner ice shelf approximately from Skytrain Ice Rise (80°W) to Vahsel Bay (35°W), and from there to the northeast as far as Cap Norvegia (12°W), respectively.

German activities have taken place within these two sectors, but especially between K and KII. The UK (eg from Halley Station), Norway, and Sweden are also active in this sector making ground-based and airborne RES measurements. The grounding zone of sector K–KII is not an area with special scientific interest to Germany.

Available measurements

The present knowledge of subglacial and seabed topography beneath Ronne–Filchner ice shelf was compiled in a map published by IfAG (1994). The corresponding ice surface topography was compiled in an earlier map by IfAG (1993).

During the 1994–95 field season new ice-thickness measurements by airborne RES were carried out along the grounding zone of the Ronne–Filchner Ice Shelf between Institute Ice Stream (approx 74°W) and Dufek Massif (approx 50°W) (see Hempel and Oerter, in press).

Ice-velocity measurements, available for the Foundation Ice Stream, were made by the Technical University of Braunschweig (see Riedel et al, in press).

Ice-thickness measurements were also made by Swedish glaciologists along Bailey Ice Stream during the Nor-wegian expedition of Dr M Kristensen in 1992, in the vicinity of the grounding zone of the Filcher Ice Shelf.

Also, in the 1994–95 season the grounding zone of the Brunt Ice Shelf was overflown several times during a combined airborne aeromagnetic and RES programme from Halley Station. Ice-thickness measurements in the grounding zone of Ekström Ice Shelf are also available from earlier measurements of the University of Munster (Thyssen and Grosfeld, 1988) and from a survey in 1994 by Alfred-Wegener-Institut für Polar- und Meeresforschung (AWI) (Mayer and Huybrechts, 1994). Ice-velocity values were also published in the Hoppe and Thyssen (1988) report on ice-thickness measurements in Western Neuschwabenland by airborne RES (see also IfAG, 1989). The grounding line area of Riiser-Larsenisen (approx 15°–11°W) was included in these measurements.

Future plans of the German programme

During the next two field seasons (1995–96 and 1996–97) German activities will concentrate on the EPICA pre-site survey in Dronning Maud Land, but glaciological activities on the Ronne–Filchner ice shelf will continue in the 1996–97 field season. A flight line east of Dufek Massif and a flight line south of Neumayer Station or Cap Norvegia, respectively, should also be possible during one of the forthcoming field seasons. A more detailed study of the grounding zone of the Filchner Ice Shelf, at the inflow of Recovery and Slessor glaciers, would need a fuel depot in the area. This is beyond the range of German logistics but is in near-future plans.

References:

Hempel, L and Oerter, H, 1995: Airborne radio-echo sounding during the Filchner V field season. In: H Oerter (ed) FRISP Report No 9, Alfred-Wegener-Institut für Polar- und Meeresforschung, Bremerhaven, in press.

Hoppe, H and Thyssen, F, 1988: Ice thickness and bedrock elevation in Western Neuschwabenland and Berkner Island, Antarctica. Annals of Glaciology 11, 42-45.

IfAG, 1989: Maps of Ice Shelf Kinematics. Topographic Map and Satellite Image Map 1 : 500 000, Ekströmisen, SR29-30/SW, Antarktis. Institut für Angewandte Geodasie (IfAG), Frankfurt am Main.

IfAG, 1993: Topographic Map (Satellite Image Map) 1: 2 000 000, Ronne/Filchner Schelfeis, Antarktis. Institut für Angewandte Geodasie (IfAG), Frankfurt am Main.

IfAG, 1994: Map of subglacial and seabed topography 1: 2 000 000 Ronne/Filchner Schelfeis / Weddell Sea, Antarktis. Institut für Angewandte Geodasie (IfAG), Frankfurt am Main.

Mayer, C and Huybrechts, P, 1994: The Ekström grounding line experiments - first results. In: H Oerter (ed) FRISP Report No 8, Alfred-Wegener-Institut für Polar- und Meeresforschung, Bremerhaven, 39-44.

Riedel, B, Karsten, A, Ritter, B and Niemeier, W, 1995: Geodetic field work along Foundation Ice Stream. In: H Oerter (ed.) FRISP Report No 9, Alfred-Wegener-Institut für Polar-und Meeresforschung, Bremerhaven, in press.

Thyssen, F and Grosfeld, K, 1988: Ekström Ice Shelf, Antarctica. Annals of Glaciology 11, 180-83.

2.2 United Kingdom (Presented by C S M Doake)

  1. Ice thickness
    The portion of the coast identified by the Chairman as being assigned to British Antarctic Survey (BAS) will be incorporated into the BAS field plans as high priority areas for airborne RES. During the 1995–96 field season, BAS will be concentrating on the area around the western coast of Ronne Ice Shelf. There might also be a chance to do some flying from the Argentine Marambio Station at the northern end of the Antarctic Peninsula, thereby covering the grounding line area of Larsen Ice Shelf. For 1996–97, there will be less opportunity to survey the grounding line as the radar will be flown in support of aerogravity and aeromagnetic surveys.
    The latest BAS airborne radar operates at 150 MHz and uses an arbitrary waveform generator to provide a number of operational modes such as pulse, chirp or pcm.
    Parameters such as pulse length, attenuation, integration time etc can be set in software. Data are recorded digitally. The radar is fitted to a Twin Otter aircraft with a four-dipole antenna beneath each wing.
  2. Velocities
    BAS experience with Synthetic Aperture Radar (SAR) interferometry suggests that obtaining accurate surface velocities near the grounding line requires good calibration data, probably best obtained from radar reflector arrays surveyed with GPS. The vertical velocity component can cause errors of 10% or more in the inferred horizontal velocity if left uncorrected because of the amplification due to the increased sensitivity to vertical motion. A minimum of five coherent images, in at least two different look directions, is necessary to separate topography from movement, but so far no one has demonstrated that this is sufficient. With additional data on surface slopes and/or velocities, fewer interferograms could be used. Caution should be exercised about the accuracies that can be obtained.
    Feature tracking in satellite SAR and visible images will be used to give velocities over Ronne Ice Shelf and elsewhere. Surface velocities can be measured accurately with GPS techniques, but even with continuous profiling it would result in a patchy coverage along the grounding line and be time consuming. A few individual sites on the Zumberge Coast will be set up and some old ones reoccupied by BAS during 1995–96. Surface velocities need to be converted to depth averaged ones for flux calculations.
  3. Accumulation
    Measurements on shallow firn cores drilled in Dronning Maud Land in 1995–96 will be used to determine recent accumulation rates there.
  4. Database
    BAS is interested in helping to define the requirements for creating and running a database of relevant information, primarily ice thickness and surface elevation, but also encompassing accumulation, velocity and temperature. Thought should be given to how data from disparate sources can be combined in a proper way, so that a high-quality product can be provided.

2.3 Russia

The Chairman presented a map showing the completed coverage of Russian/Soviet RES and aeromagnetic/aerogravity flight lines. Some of the aerial surveys were completed in 1993–94. There was a discussion about the need for the Russian SCAR delegates to inform the Russian scientists of the SCAR-related activities and to encourage participation in GLOCHANT Task Groups. The Chairman reported on a discussion that he had had recently with Dr V Masolov, Head of the Antarctic Division of Polar Marine Geological Research Expeditions (PMGRE), in St Petersburg, Russia. Dr Masolov told the Chairman that he would recommend to his logistics people the Task Group's recommendations on future coast-parallel RES airflights. It was suggested that Dr Alexander Golynsky, a colleague of Dr Masolov at PMGRE, be invited to become a member of the ISMASS group.

2.4 Sweden (Presented by P Holmlund)

Present climate and present state of the ice sheet

It is proposed to conduct:

The output from these studies will include: accumulation rates and trends, velocities and ice fluxes, physical data for decision on a drilling site for EPICA, and ice fluxes at the grounding line along the coast.

Last 500 years

It is intended to drill medium-depth firn cores for chemical analyses. The expected output is a climate record, detailing climate changes since the beginning of the Little Ice Age, and changes in anthropogenic-induced aerosols and tracers.

Last 100,000 years

A deep ice core will be drilled at the EPICA site in the future. Ice velocity, ice thickness and accumulation measurements will be collected to support modelling studies. The expected outcomes will include a model of climate changes and changes in ice fluxes and shape of the ice sheet over the last ice-age cycle.

Proposed fieldwork 1996–97

Using Norwegian logistics, a Scandinavian oversnow traverse is planned from the coast at SANAE Station inland to 3,000 m elevation, via Troll Station. Snow radar soundings will be carried out using a CW-radar (1–2 Ghz), with detailed calibration using firn cores. The cores will be analysed using density, conductivity (identification of volcanic horizons) and beta radioactivity horizons, to determine annual accumulation layers. A small number of firn cores will be returned to Sweden for further analysis. If logistics are available, airborne RES surveys will also be conducted using a CW-radar, 150–180 MHz.

Proposed fieldwork 1997–98

Using Swedish logistics, a Scandinavian oversnow traverse is planned from Wasa to 3,000 m elevation (Amundsenisen) via Kibergdalen. A survey of the EPICA site, ice velocities (GPS), shallow coring (accumulation rate, chemistry, oxygen isotope (18O, 14C and conductivity), snow radar (spatial distribution of snow cover) soundings (using a CW-radar, 1–2GHz ) and RES soundings (bedrock topography and internal layering) using a CW-radar (150–180 MHz) will be conducted on the traverse. Shallow firn coring will also be conducted for snow radar calibration. A medium-length firn-coring programme is also planned to depths of 100–200 m. One site is planned at 500 m asl and one at 3,000 m asl. Duplicate cores will be drilled at each site. The cores will be analysed for radionuclides and concentrations of organic halogens in ice. Airborne RES surveys will be flown along the grounding line, either by fixed-wing aircraft or by helicopter.

The following Institutes are collaborating in the above projects:

Department of Physical Geography, Stockholm University, Sweden.

The Swedish Polar Research Secretariat, Stockholm.

Department of Meteorology, Stockholm University, Sweden.

Institute for Marine and Atmospheric Research Utrecht, The Netherlands.

Environmental Surveillance Technology Programme, Lillestrøm, Norway.

The Swedberg Laboratory, University of Uppsala, Sweden.

Water and Environmental Studies, Linkoping University, Sweden.

Institute for the Study of Earth, Oceans and Space, University of New Hampshire, USA.

Snow radar surveys of accumulation, Dronning Maud Land

Snow layers have been mapped successfully using a ground-based snow radar along two traverses from the coast up to 3,000 m asl. The recordings show large local and regional variations in the accumulation pattern along the 2,300 km measured profiles. These traverses were a part of the Swedish contribution to the International Trans-Antarctic Scientific Expedition (ITASE). In order to verify the geographical representation of firn cores, a detailed net of snow radar profiles was surveyed at the firn core drill sites.

The radio-echo sounder is a continuous wave step-frequency radar, based on a Hewlett-Packard Network Analyser (8753C). The antennae used were a pair of AEL APN-106AA. The snow radar was operated in the frequency range 800–2,300 MHz giving a penetration depth of 12–14 m. GPS positions were recorded every 5 km, ie at the beginning and end of each data file.

Distinct snow layers are visible in the radar images. The concept is to follow a specific snow layer from the coast to the polar plateau and to register changes in depth. The speed of the radar signal is a function of snow density, and the depth-scale of the radar images is calculated from density data obtained from firn cores. By correlating the radar registrations to core data the snow layers can be dated; also, these correlations make it possible to deal with problems caused by regions with disturbed snow stratigraphy. A coastal section of the traverse follows a stake line over a distance of 160 km. The accumulation pattern obtained from the snow radar soundings is consistent with the results from stake measurements made at several occasions between 1987 and 1994.

2.5 Australia (Presented by J Jacka)

Australian Antarctic field activities concerning the ice sheet mass balance over the past few years have involved a tractor-train traverse from Mawson to the Larsemann Hills, approximately following the 2,500 m elevation contour around the southern end of the Lambert Glacier. The Lambert Glacier Basin Traverse measurements included ice surface and bedrock elevation using a 100 MHz ice radar, surface velocities (GPS), snow accumulation (canes placed every 2 km) and gravity. In addition, shallow cores were drilled for temporal accumulation variability studies, a pit was excavated for stratigraphy studies and six automatic weather stations were deployed around the basin.

With the completion of the Lambert Glacier basin project, Australian traverses over the past twenty years have (with the exception of one sector) completed surveys and examined the ice sheet mass balance along the 2,000 to 2,500 m elevation contour from Enderby Land to Porpoise Bay (eastern Wilkes Land). This survey extends from AI to near D, except for a gap from C halfway to CI, which will be completed in the near future. This missing sector is that between the Grove Mountains and a point inland of Mirny Station. Measurements (Sikorsky 76 helicopter-supported) for this sector are included in the current ANARE strategic plan and are expected to be carried out during the next five years.

During the 1995–96 austral summer, differential GPS measurements will be made to determine directly the mass budget of the Lambert Glacier – Amery Ice Shelf drainage basin by measuring elevation change since 1968 when measurements were made in the vicinity of the grounding zone. Measurements will also provide surface elevation data to ground truth profiles derived from satellite radar altimetry, reference locations for future airborne RES and accurate values of the geoid-ellipsoid separation.

Currently, the Australian programme does not have a digital ice radar suitable for airborne RES measurements. The present oversnow traverse ice radars are digital logging units based on a digital storage oscilloscope. These cannot sample the returned waveforms fast enough, receive only a fraction of return power and sample only an average over 10 m along track. This radar sampling is too slow to be used in aircraft at much higher speeds and rates of sampling. Development of an airborne digital ice radar system will be carried out over the next two years. With the recent introduction into the Australian programme of Sikorsky long-range helicopters and the possibility of fixed-wing aircraft availability over the coming five-year period, it is anticipated that an aerial ice radar capability will allow ice radar coverage of the grounding zone for the entire Australian sector over this period. Currently, the long-range helicopters operationally fly the legs Mawson – Davis, Davis – Bunger Hills and Bunger Hills – Casey, transporting personnel and equipment. It is anticipated that these flights may be utilized to carry out the ice thickness surveys, in the near future. There is a possibility of extending airborne RES coverage along the grounding zone in Terre Adélie and George V Land, toward DI, by using the Sikorsky 76 helicopters from a ship base. These helicopters would have a range of 600 km. It is unclear whether the Australian ship, RV Aurora Australis, will be available. It is more likely that a cooperative programme using a ship such as the US RV Nathaniel B Palmer will be required. It is envisaged that the airborne RES surveys could be flown in conjunction with ANARE/international geological programmes in this sector, that are proposed for 1998–99.

2.6 Italy (Presented by G Orombelli on behalf of Dr M Frezzotti, not attending)

The Italian programme has a SPRI digital ice radar which will be installed in a Twin Otter aircraft in the 1995–96 season. Flight lines are planned along the coast south from Terra Nova Bay to Mackay Glacier, and north to Rennick Glacier, along the western Ross Sea. Flights are also planned inland toward Dome C, in support of the EPICA deep drilling site selection. Future plans include transects of the main outlet glaciers and possibly further surveys around Dome C. It is likely that continuous coverage from McMurdo Sound to Cape Adare can be provided. It is logistically difficult to work west of Cape Adare. The Italian programme may have a Twin Otter in future years for logistic support, that may also be available for additional radar flights.

Recent mass balance data from northern Victoria Land

Snow accumulation

A shallow (10–40 m) firn core drilling programme was conducted at seven sites in northern Victoria Land. Snow accumulation ranges across these sites from 130–270 kg m-2yr -1 and was interpreted using oxygen isotope (δ16O/18O), hydrogen peroxide (H2O2), and methyl sulphide (MSA) seasonal signals. The sites are located at: McCarthy Ridge 74°33'S, 163°01'E; Styx Glacier 73°55'S, 163°45'E; Hercules Névé 73°07'S, 164°58'E; Priestley Névé 73°38'S, 160°38'E; Rennick Glacier 73°15'S, 162°29'E; Pilot Glacier 73°15'S, 165°30'E; Niggli Névé 72°41'S, 166°58'E; Bowers Glacier 72°44'S, 168°50'E).

Snow temperature, at 10 m depth, was measured at four sites (Hercules Névé 73°07'S, 164°58'E; Priestley Névé 73°38'S, 160°38'E; Pilot Glacier 73°15'S, 165°30'E, Drygalski Ice Tongue 75°30'S, 165°18'E)

Ice-velocity measurements

Surface velocity was measured by GPS for:

David Glacier-Drygalski Ice Tongue (12 stations*), Priestley Glacier (3 stations), Reeves Glacier (2 stations*);

and by remote sensing for:

Mackay Glacier (OG*), Mawson Glacier (OG*), Marin Glacier, Harbour Glacier Tongue, Cheetham Ice Tongue, Clarke Glacier, David Glacier (OG*), Larsen Glacier (OG*), Reeves Glacier (OG*), Priestley Glacier (OG*), Campbell Glacier*, Random Hills Glacier, Tinker Glacier Tongue, Glacier Tongue, Parker Glacier Tongue, Icebreaker Glacier, Fitzgerald Glacier, Wylde Glacier, Suter Glacier, Ridgeway Glacier, Mariner Glacier Tongue, Borchgrevink Glacier Tongue, Tucker Glacier, Ironside Glacier, Moubray Glacier, Lillie Glacier Tongue, Barber Glacier*, Gannutz Glacier*, Rennick Glacier (OG), Pryor Glacier, Suvorov Glacier*, Gillet Ice Shelf, Tomlin Glacier* (OG) Outlet Glacier.

*Measurements close to grounding line.

GPS and remote sensing surface velocities differ by less than 2% in valley glaciers or near nunataks and lower than 10% on ice tongues and on ice shelves without reference points and with different path and row satellite orbits.

Ice tongue and ice shelf thickness were evaluated at the ice front with GPS:

Nansen Ice Sheet, Hells Gate, Larsen Glacier Tongue, Campbell Glacier Tongue, Random Hills Glacier Tongue, Tinker Glacier Tongue, Aviator Glacier Tongue, Suvorov Glacier Tongue*, Tomlin Glacier Tongue*.

*Measurements close to grounding line.

Preliminary RES data have been obtained for local glaciers.

A preliminary evaluation of the mass balance for the sector from Cape Crozier 77.5°S, 169°E (E of Giovinetto and Bentley, 1985) to Williamson Head 69.2°S 164°E (D' of Giovinetto and Bentley, 1985) suggests a positive value (M Frezzotti, in preparation). Surface elevation was measured by GPS at Tarn Flat and the Strandline Glaciers (1993) and along the traverse from Terra Nova Bay to the East Antarctic plateau 280 km (1993).

Italian Antarctic Programme 1996-1998

The following measurements will be conducted in 1996–98. Mass balance measurements will be made on the ITASE traverse in 1996–97, from Terra Nova Bay to Talos Dome, about 450 km inland. They will also be made in 1997–98 on a traverse from Terra Nova Bay to Dome C, about 1,300 km inland along the David Glacier drainage basin. Remote sensing measurements will include the determination of ice front fluctuations and ice flow from Williamson Head 69.2°S, 164°E (D' of Giovinetto and Bentley, 1985) to point west of Dibble Glacier Basin 66.1°S, 134°E (D of Giovinetto and Bentley, 1985). RES will be carried out in 1995-96 and the following seasons in northern Victoria Land and the Dome C area.

2.7 USA: Field surveys (Presented by C R Bentley)

The ability of the US Antarctic Program (USAP) to contribute to the planned RES of the Antarctic Ice Sheet grounding zone is problematic. Although the capability exists the operating time of the SOAR remote sensing Twin Otter aircraft is fully committed for the next year or two. Furthermore, it would be logistically difficult for the Twin Otter to cover the Amundsen and Bellingshausen Coasts (F to HI) because they are so far from any support base. He stated that he will pursue further the possibilities for covering the entire grounding zone of the Ross Ice Shelf. He will also investigate whether the SOAR aircraft could do a small amount of sounding in the Terra Nova Bay for comparison of results with those of the Italian airborne radar sounding to be initiated in the 1995–96 field season.

C R Bentley also discussed the opportunity raised by the potential availability of the long range Lockheed P-3 aircraft that has been used for many years by the US Naval Research Laboratory (NRL) for Project Magnet. The P-3 has a range of 5,000 km at a suitable low altitude flight height (7,000 km at high altitude), and thus in principle could be used to cover not only the FHI sector but also much of the rest of Antarctica, from a base at McMurdo Station. The P-3 is not ski-equipped. The challenge for using the P-3 is to find the financial support required for its operation, which must be found from sources outside the US Navy. USAP cannot support it alone, although a contribution to an internationally-supported activity might be possible. He circulated a paper giving more details on the P-3 and its potential with a 150 MHz radar antenna system installed, that he had received from Dr J Brozena, head of the Geodetic and Geophysical Systems Section, Marine Geoscience Division, NRL. The P-3 aircraft would potentially be available for RES in Antarctica between 1996 and 1998.

The members of the Task Group expressed an interest in pursuing a cooperative, internationally funded RES programme using the P-3 aircraft. Countries without an RES capability may be interested in contributing. Dr Brozena is a geophysicist, who is principally interested in simultaneous aeromagnetic and aerogravity surveys for which the P-3 is equipped. A combined aeromagnetic, aerogravity and RES survey may be a selling point to the national programmes. It was decided that C R Bentley will work with Dr Brozena to develop a proposal to be circulated to all SCAR countries as an attempt to develop an internationally supported activity.

Dr J Sievers is doing repeat image velocity studies using ERS-1 SAR images and has asked the Germans in the field to emplace radar reflectors for geodetic control this season. Control using reflectors or nunataks is very important, as otherwise the accuracy of position determination between images is only about ±50 m. Since features tend to disappear after two years there is a 25 m yr -1 limit on the accuracy of velocity measurements. The geodetic control problem is more severe with ERS-1 radar images than with Landsat images because the smaller image size (100 km on a side vs 200 km) makes the chance of including a nunatak on each image less likely. Other expeditions should be urged to put out radar reflectors where possible. Other scientists doing ice-velocity studies from repeat image analysis are B K Lucchitta (USA), M Frezotti (Italy) and N Young (Australia). It is important to encourage similar remote sensing studies by other workers, since the United States Geological Survey (USGS) programme may be terminated before reaching its goal of complete circum-Antarctic coverage. The problem of a lack of geodetic control exists for interferometric SAR also. In principle, overlapping images could be extended to the nearest fixed point but this would be very expensive. ERS-1 and ERS-2 are in tandem mode with a one-day delay; the tandem mission will last a total of six or nine months. There will be a workshop in October on obtaining the greatest benefit from the ERS SARS generally (not just in Antarctica).

2.8 USA: Remote sensing (Presented by B K Lucchitta)

Establishing velocities of glaciers, ice streams, ice sheets, and ice shelves is essential to calculate the discharge of the mass balance equation. Velocities can be obtained on the ground, by satellite, or by a combination of the two. Ground measurements are time consuming and difficult logistically; nevertheless they exist already for many glaciers and ice streams that have been studied in the past. A need exists to have these measurements assembled in a readily available data base. Satellite-based velocities are obtained by tracking features that move with the ice (mostly crevasses) on repeat images covering the same area on the ground. A new satellite technique is radar interferometry, in which repeat images need to be closely aligned in space and time. Measuring velocities by a combination of satellite and ground efforts involves, for instance, tracking satellite radar images by corner reflectors placed on the ground, and tracking moving stations by GPS. Considering the difficult Antarctic environment, the vast area involved, and the cost of ground expeditions, the workshop members felt that obtaining velocities by satellite is the only feasible method for mass balance determinations within the near future.

Landsa

B K Lucchitta showed examples of velocity determinations made by Landsat along the Marie Byrd Land coast. Most of the measurements were made on images dating from the early 1970s to the late 1980s, spanning time intervals of as much as twenty years. As the early images are MSS (Multispectral Scanner) having 80 m resolution, only large crevasses in floating ice (shelves and tongues) can be tracked. However, regression lines on plots of velocity versus distance, when projected toward the grounding line, give approximate velocities at this line. B K Lucchitta gave the velocities along the Marie Byrd Land coast (projected to the grounding line) only to the nearest 0.1 km yr -1 because of errors introduced by irregularities in the internal geometry of early Landsat images and uncertainties in orbit determinations. (B K Lucchitta has shown in publications that velocities as accurate as ±0.02 km yr -1 can be obtained under optimum conditions, including the registration of MSS images to later and more stable Landsat TM (Thematic Mapper) images of 30 m resolution.) Velocities were obtained for the Dotson Ice Shelf, for five areas along the Getz Ice Shelf, and for three areas on the shelf in the Sulzberger Bay. Some areas could not be covered because suitable Landsat images without cloud cover were not available. Projected to the grounding line, velocities along the Dotson and Getz Ice Shelves are around 0.2 km yr -1. Velocities on the shelf near the DeVicq Glacier increase westward to near 0.5 km yr -1 and reach 0.6 km yr -1 at the glacier. The Land Glacier, at 1.6 km yr -1, is exceptionally fast for a glacier with a small drainage basin. Ice in the Sulzberger Bay, on the other hand, is very slow at less than 0.1 km yr -1.

ERS

B K Lucchitta showed velocities obtained for the Pine Island and Thwaites Glaciers as examples of data that can be obtained with radar images (ERS-1). An automated cross-correlation technique was used for tracking patterns in the ice, and hundreds of points were obtained for each glacier both above and below the grounding line. (The 25 m ground resolution of ERS permits tracking patterns above the grounding line.) Findings include:

  1. The velocity of the Pine Island Glacier increased across the grounding line (as located by radio-echo sounding by Crabtree and Doake, 1982) from about 1.5 km yr -1 to about 2.6 km yr -1, an increase of about 1 km yr -1.
  2. The velocity of the Thwaites Glacier also increased across the grounding line from about 2.0 km yr -1 to about 3.0 km yr -1, a similar increase.
  3. Few trackable points were picked up by the computer programme in the region of both grounding lines, perhaps reflecting a grounding zone or movement across a steeper slope, where patterns were not preserved.
  4. Neither glacier is buttressed by an ice shelf, and movements are exceptionally fast. The observations show that future RES bedrock traverse profiles near the grounding line have to be placed with care for fast-moving glaciers without shelves, because the velocities may change rapidly. The best place to locate the radar profile would be at the upstream side of the rapid velocity increase. Ice streams that merge with shelves, on the other hand, may show a velocity decrease, as pointed out by C S M Doake for the Rutford Ice Stream, or show no noticeable acceleration, as noted for Ice Stream E by Scambos (NSIDC, Boulder, Colorado, USA, not attending, personal communication). M Frezzotti (ENEA, Rome, Italy, not attending, personal communication) noticed an increase in velocity of about 0.1 km yr -1 where the David Glacier (northern Victoria Land) crosses the grounding line and becomes the Drygalski Ice Tongue. He thought that, for slow-moving glaciers (on the order of 0.5 km yr -1), such an increase is insignificant and does not warrant special care in locating the grounding line.

Coastal Maps

B K Lucchitta mentioned that the Marie Byrd Land velocity measurements are part of a larger project by R S Williams, J G Ferrigno (USGS, Woods Hole and Reston, USA), C W M Swithinbank (SPRI, Cambridge, UK), and that B K Lucchitta is to make an inventory of the position of ice fronts and grounding lines of the entire Antarctic coastline, based on Landsat images of the 1970s and again of the 1980s. The results are to be published as twenty-four 1:1 000 000 scale maps covering the coastline. A prototype map of the Bakutis Coast has been compiled. The velocities are shown as scaled velocity vectors on the maps.

Limitations and Errors

Landsat images have limitations due to the coarse resolution of the MSS images, high cost of the TM images, current absence of a functioning system for Antarctic acquisitions, and obscured images due to cloud cover. Radar images from ERS and RADARSAT lack these limitations and they will probably provide most of the satellite-based velocity data in the future. However, ERS images also have limitations. The region south of about 80°S is not covered, and currently only two receiving stations are operational. In addition, they are manned only for short-time periods. The German station at O'Higgins on the Antarctic Peninsula covers the coastal regions from about 120°W eastward to about 0°, and the Japanese Station at Syowa from about 60°W eastward to about 110°E. The O'Higgins station has been manned once or twice a year for about one month each, providing good acquisition opportunities. The Syowa station has been less reliable, and scheduled acquisitions generally have not been received. The US McMurdo Station is now operational and covers the Antarctic coastline from about 70°E eastward to about 70°W. The station will be manned for the entire year, so that acquisition opportunities will be increased markedly. The memorandum of understanding (MOU) between the European Space Agency (ESA) and the United States Alaskan SAR Facility is about to be signed, and images received by McMurdo will be available in the near future. ERS images have a nominal ground-location accuracy of 50 m (Roth et al, 1993). This claim was verified by B K Lucchitta in her investigations. Where outcrops are available, images may be co-registered to near-pixel accuracy (12.5 m). Fast-moving glaciers tend to lose trackable points in the grounded part when time intervals increase beyond one year, whereas slowly moving ice readily preserves trackable points over a two-year interval. B K Lucchitta showed examples of measurements made on the grounded part of the Bailey Glacier in East Antarctica, which moves at about 0.1 to 0.2 km yr -1. Assuming location errors in the images of 50 m, a glacier moving at 1 km yr -1 traced for one year would have a velocity error of ±5%, whereas a glacier moving at 0.1 km yr -1 traced for two years would have a velocity error of ±25%. For images co-registered by outcrops the possible errors would be reduced by one fourth to near ±1% for fast-moving glaciers, and near ±6% for slow-moving glaciers. At this point the Chairman raised the questions whether such potential errors would invalidate the entire effort to measure velocities with satellite images, because a 50 m yr -1 error around the entire continent amounts to some 15% of the total output flux. J Jacka objected to this assessment because 80% of the Antarctic ice is discharged through fast-moving ice streams, and approximate velocities for the remaining 20% would suffice. For instance, five major ice streams cover 80% of the outflow from the Amery Ice Shelf to the Napier Mountains. W F Budd added that ice sheets flowing out onto shelves move at a rate on the order of 10 m yr -1, whereas ice streams may move at 800 m yr -1, demonstrating that not all of the coast is important in terms of discharge and that efforts should be concentrated on the ice streams. B K Lucchitta further noted that slow-moving ice sheets generally show few trackable points, so that velocity measurements can only be made in selected places or on adjacent shelves. Thus velocities for the ice sheets will have to be projected from locations nearby. For slow velocities, SAR interferometry may help.

SAR Interferometry

ERS-1 and 2 are currently in a tandem orbit, specifically designed to permit interferometry measurements. The spacecraft trail each other by one day, and their orbital tracks give an interferometry baseline of 50–600 m. This configuration is to be maintained through mid 1996. A special effort, spearheaded by F Carsey (ASF, Fairbanks, Alaska, US) and J Sievers (IfAG, Frankfurt, Germany) is underway to ensure that Antarctica will be adequately covered during the tandem-mission period (mostly using McMurdo Station). C S M Doake pointed out that interferometry may yield velocities accurate to 1 cm per day (4 m yr -1), but the calibration of the data (including separation of topography from velocity effects) remains a problem. Accurately located corner reflectors may help to calibrate the interferometry fringes. Even though acquisition of suitable images for interferometry appears assured, the question of who will exploit the data and how they will be exploited remains open.

RADARSAT

Accurately located corner reflectors (using GPS) are needed also for the calibration of the RADARSAT orbit. The satellite is to be launched in October 1995. A dedicated Antarctic Mapping Mission phase is planned for 1996 and 1998 and includes rotating the satellite so that the radar beam views south, thus permitting coverage of the entire Antarctic continent. Eighteen days of data acquisition are necessary to cover Antarctica utilizing "Standard Beam 2" (similar in geometry to ERS but looking south) and additional beams with greater look angles near the pole. The data are tape-recorded and can be downloaded at ASF in Alaska, so that Antarctic receiving stations are not essential. The RADARSAT Antarctic Mapping Mission seeks to produce mosaics controlled to better than 100 m. To accomplish this accuracy, good ground control is needed around the perimeter of the continent and south of 85°S. Corner reflectors, existing Digital Elevation Models (DEMs), and interferometry will be used to improve ground-location accuracy. The extent to which the RADARSAT maps will be useful for detailed velocity measurements remains to be seen.

ASTER

The problem of ground control is not unique to RADARSAT. Future spacecraft, such as the Japanese-American ASTER (Advanced Spaceborne Thermal Emission and Reflection radiometer), to be launched in June 1998 as part of the Earth Observing Mission (best ground resolution 15 m, stereo capability, fourteen spectral bands in the visible, near, and far IR), also need good ground control. This mission has a dedicated glacier monitoring programme that includes Antarctica. Multiple acquisitions are planned to minimize the effect of clouds. ASTER's along-track stereo capability can produce digital elevation models with 15–25 m vertical resolution where surface contrast (eg crevasses, nunataks) is available. One point per image is sufficient because the images are internally stable. To facilitate future ground control needs, H Kieffer (USGS, Flagstaff, USA) is promoting the establishment of a databank that lists all available Antarctic ground control points, documented by photos, satellite images, and digital information on pixel location within the images. Ideally, each Antarctic research party using GPS should thus document their measured locations. Even though such a databank would be very useful, it most likely will not be available in the near future. Other points of interest were discussed.

Miscellaneous

Other thoughts and concerns with respect to velocity requirements were expressed throughout the meeting:

H Oerter commented that German research groups have made measurements on the Foundation Ice Stream (0.5 km yr -1), but that it is difficult to locate the grounding line because no high-resolution satellite images exist for the region and seismic and radar-echo investigations do not give the same results for the location of the line. In addition, flying aircraft into the region is difficult because of problems associated with fuel supply. He commented that the availability of a US C-130 aircraft would be very helpful for this region.

B K Lucchitta asked C S M Doake whether the UK is systematically measuring glacier velocities around the Antarctic Peninsula. C S M Doake replied that systematic efforts are being undertaken in the Ronne-Filchner ice shelf region but not on the peninsula, where velocity measurements and coastal monitoring is spotty. He mentioned that J Sievers of IfAG, Germany, may have systematic plans in that region.

J Jacka gave some data on ground measurements in the Lambert Glacier region. The Australians measured surface velocities by GPS every 30 km along the 2,500 m contour around the Lambert Glacier basin. On the SE side, inflow into the Lambert Glacier is 54 m yr -1, on the NE side about 10 m yr -1. Errors are within tens of centimetres.

N Young (Antarctic CRC, Hobart, Tasmania, not attending) is obtaining velocity measurements of several major outflow glaciers in the "Australian" sector of Antarctica. The Australian programme will concentrate on obtaining longitudinal velocity profiles on the coastal glaciers, in concert with radar sounding. Velocities of all major glaciers have been obtained. It also proposes to obtain some velocity-depth profiles on coastal glaciers to calibrate the surface velocities on outflows.

G Orombelli mentioned that the Italian groups are obtaining velocity values by GPS. In addition, M Frezzotti (not attending) measured velocities of most glaciers along the northern Victoria Land coast based on all available aerial and satellite images. This region may serve to calculate the mass balance of small glaciers, to assess whether such glaciers should be included into discharge and mass balance calculations for the Antarctic ice sheets. The results from such a study may also apply to the Antarctic Peninsula, where many small glaciers exist. However, on the peninsula minor glaciers are most likely important because of the high precipitation rate in this region. I D Goodwin expressed concern that the ice input data for mass balance calculations may not be registered temporarily with the output data. He suggested that mismatch in time may result in errors of as much as 25%. Velocities, however, most likely do not vary by that much in the time intervals covered by the anticipated mass balance investigations. Careful monitoring of velocities of selected ice streams may show to what extent velocities do indeed change.

B K Lucchitta encouraged all workshop participants to include in their reports the names of glaciers for which velocity measurements have been made by their respective countries, in order to get an idea to what extent the Antarctic coastline has been covered.

2.9 Japan (No report received)

The Chairman informed the Task Group that there was no information about the Japanese RES capabilities.

3. Accumulation Update and Modelling Overview

W F Budd reported that modelling of atmospheric moisture fluxes over Antarctica using all available observational data on and around the continent, including buoys, and a diagnostic atmospheric circulation model, can now yield values for small-scale average accumulation rates that are good to ±10%. Data exist to carry the annual average analyses back as far as 1988. W F Budd proposed that the ISMASS group set up a GLOCHANT databank comprising raw data and 10 km gridded data, including accumulation, ice thickness and velocities. He proposed that this databank be produced jointly by the Antarctic CRC and the SCAR Global Change Programme Office in Hobart. C S M Doake and H Oerter agreed with the proposal and recommended that it be a joint SCAR and EISMINT databank. I D Goodwin urged members to encourage field programmes to include shallow firn coring in preference to stake networks, to give a broad picture of temporal variation in snow accumulation. W F Budd suggested that national fieldwork programmes should aim to determine long-term estimates of accumulation rates from the stratigraphic use of a common bomb horizon. This would allow the determination of long-term accumulation rates for a common period across Antarctica.

4. Action Items

The Chairman and I D Goodwin will prepare a proposal to be sent to national glaciology representatives and logistics operators (COMNAP) about seeking interest and support for an international cooperative RES and aerial geophysics programme, using the Lockheed P-3 aircraft. Countries that do not presently have an RES capability should be particularly encouraged to contribute financial support. The proposal should be distributed to key people including the respective SCAR National Committees, Managers of National Antarctic Programmes, and national logistics operators.

GLOCHANT-ISMASS plans for RES surveys around the grounding zone of the entire Antarctic Ice Sheet should be distributed to national operators to ensure they understand the importance of supporting this programme.

W F Budd will prepare a proposal for the establishment of a GLOCHANT/Antarctic CRC (Australia) data set, comprising raw data and 10 km gridded data on accumulation, ice thickness and surface velocity.

H Oerter and C S M Doake will encourage Weddell Sea regional workers to coordinate their RES efforts and their fuel depot requirements. B K Lucchitta and the Chairman suggested that there is a need to evaluate the contribution of small coastal glaciers to mass flux, especially in Victoria Land (E-DII) and along the Antarctic Peninsula.

The Chairman and I D Goodwin will arrange for Dr A Golynsky of PMGRE in St Petersburg to be invited to become a member of the ISMASS Group.

B K Lucchitta identified that there was a large area where no RES coverage existed in Dronning Maud Land, since the Swedish survey area does not extend from A to AI. It was suggested that the Japanese programme should be asked to make a contribution to this area.

J Jacka will arrange liaison between the Australian and Russian glaciology programmes to determine cooperative work in the sector between Mirny and Davis stations. The Russians should be encouraged to conduct RES parallel to the coast and along the grounding zone in this sector.

Australia will develop an aerial ice radar facility and complete RES surveys along the Lambert Glacier and Amery Ice Shelf system, and conduct two parallel surveys along the grounding zone of the ice sheet between Enderby Land and King George V Land.

B K Lucchitta agreed to prepare a summary on the accuracy of satellite-image analyses of ice-surface velocity determination. The purpose is to establish the minimum velocities that can be determined with reasonable confidence. She also agreed to prepare a general overview of the locations where various satellite image techniques are being used for velocity measurements.

All measured velocities are to be submitted to a GLOCHANT databank, listing latitude, longitude, magnitude, and direction of selected measured points. Current databanks exist at BAS (Dr D N Vaughan, UK), in Australia, and at the SAR facility at the USGS in Reston, Virginia, USA. No decision was made concerning the location of the envisioned databank for velocity values.

The use of satellite passive microwave data should be encouraged for the interpretation of accumulation data across the ice sheet. Research on the application of continuous snow layer profiling and accumulation rate interpretation from snow radar should be encouraged and used on all future ITASE traverses, in conjunction with shallow firn coring at regular intervals to enable the calibration of the radar data, with a known stratigraphic horizon such as a bomb horizon. It was recommended that the ISMASS Group should merge with ITASE on the planning of future traverses. The Chairman agreed to contact Dr P Mayewski and discuss future arrangements.

W F Budd recommended that measurement of velocity-depth profiles should be made, near the coast where possible, in future fieldwork programmes.

List of Acronyms and Abbreviations