National Aeronautics and Space Administration

As part of its Offfice of Earth Science, NASA supports various research programs in the Arctic that emphasize applications of airborne and space remote sensing to studies of the earth and space sciences. This issue focuses primarily on NASA's Program for Regional Arctic Climate Assessment.

NASA's Program for Regional Arctic Climate Assessment (PARCA) was formally initiated in 1995 by combining into one coordinated program various investigations associated with efforts, started in 1991, to assess whether airborne laser altimetry could be applied to measure ice sheet thickness changes. It has the primary goal of measuring and understanding the mass balance of the Greenland ice sheet, with a view to assessing its present and possible future impact on sea level. The lessons learned from this program will be applied to the more global assessment of ice sheet volume that will become possible after the 2001 launch of NASA's Geoscience Laser Altimeter System (GLAS), which has the primary goal of measuring changes in ice-sheet elevation at latitudes up to 86N. The main components of PARCA are as follows:


FY 96 FY 97

Polar Ice Interactions 5,500 5,300
Ecology 6,400 2,500
Solid Earth Science 500 460
Arctic Ozone 1,000 7,500
Clouds and Radiation 1,100 1,200
Sub-Orbital Science 700 700
Iono/Thermo/Mesospheric SR&T 600 600
FAST Auroral Snapshot 5,200 4,076
Magnetospheric SR&T 1,000 1,000
Solar Terrestrial Theory 400 400
Arctic Data Systems 14,600 14,300
Research Balloon Program 0 832
Sounding Rocket Program 1,000 955
Total 38,000 39,823

Field activities in 1996 and 1997 involved remeasurements along a GPS traverse to infer ice velocity across the 2000-m contour (closer to 3000 m along the eastern flank of the ice sheet), collection of several shallow (20- to 150-m-deep) cores to improve our knowledge of snow accumulation over the ice sheet, local measurements of ice thinning/thickening in shallow drill holes, and maintenance of several automatic weather stations (AWS). Also, the airborne topographic mapper (ATM) and ice-sounding radar were flown aboard the NASA P-3 to measure ice thickness along the GPS traverse and across selected outlet g laciers and to complete surface elevation and ice thickness surveys over the Jakobshavn catchment and over an area inland from the Humboldt and Petermann Glaciers.

Airborne Laser-Altimeter Monitoring of the Ice Sheet

Each year since 1991 the NASA P-3 aircraft, equipped with dual-frequency carrier phase tracking GPS receivers, a ring-laser gyro inertial navigation system, scanning and profiling laser altimeters, and (since 1993) a low-frequency radar to measure ice thickness, has flown over numerous transects of the Greenland ice sheet. In excess of 100,000 km of trackline have been mapped, covering all major drainage basins and characteristic geophysical regions in Greenland. The airplane location was measured precisely using differential GPS surveying techniques, allowing all altimetry data to be converted into measurements of ice surface elevation relative to the Earth ellipsoid. Analyses of these data indicate that ice surface elevations can be reliably measured with an accuracy of approximately 10 cm over baselines of more than seven hundred kilometers.

Mapping of the ice sheet by the ATM scanning laser onboard NASA's P-3 aircraft typically produces a set of surface elevations along a 150-m-wide swath. With an aircraft speed of 150 m/s and a laser pulse rate of 3000 per second, this results in an average of one surface elevation per 7.5 square meters, with each elevation measurement having some uncertainty due to measurement noise, aircraft pitch and roll errors, and GPS positioning uncertainty of the aircraft. To monitor elevation change the aircraft flies over the same groundtrack one or more years later, and the surface elevations are then compared.

All major drainage basins on the ice sheet were mapped with the ATM in 1993 and 1994, and the flight lines will be resurveyed in 1998 and 1999 to reveal any changes in surface elevation that have occurred during the five-year interim, providing an estimate of the change in volume of the ice sheet.

Data from the ATM have also been used to estimate surface velocity by tracking elevation features at the ice surface. During the 1997 field season, repeat flight lines were surveyed over four sections in the Jakobshavn drainage basin, with time separation between these flights ranging from two to six days. By interpolating the laser scanner data onto a 1- X 1-m grid and using cross-correlation analysis techniques, the movement of the elevation features, the most distinct of which are crevasses, were tracked. The results show that the surface velocities near the calving front are on the order of 7 km per year. This method of estimating ice velocity complements satellite-based techniques in that it offers a flexible platform for site-specific observations. When combined with the thickness data from the ice-penetrating radar, which is flown simultaneously on the P-3, these velocities can be used to estimate discharge fluxes. The method also offers an independent means of validating ice velocities derived from SAR and visible imagery.

Coherent Radar Depth Sounding of the Greenland Ice Sheet

This research involves performing airborne ice thickness measurements using a coherent radar depth sounder. The radar operates at 150 MHz and is capable of measuring polar ice sheet thickness to about 4000 m in cold ice and to lesser thicknesses in temperate glaciers. The radar uses complementary surface acoustic wave devices for pulse expansion and compression. The system operates as an unfocused SAR.

Radar data, collected over the interior ice sheet, around the margin and over outlet glaciers, were tagged with geolocation information obtained from the onboard GPS receiver. The radar data were collected in conjunction with laser altimeter measurements of ice surface elevation. A next-generation coherent radar depth sounder (NGCORDS) using microwave monolithic integrated circuits was also developed and field tested in 1997. The ice thickness along the 2000-m velocity traverse and at Summit (up to about 3200 m) was measured, and the measurements are being used in conjunction with other data to estimate Greenland's mass balance.

Ice thickness measurements were also used to assess the accuracy of high-quality digital elevation model data that were used to estimate the thickness of floating glacier sections for comparison with the CORDS measurements. An extensive data set was also obtained for the Jakobshavn outlet glacier, which is considered to have the highest iceberg production of all Greenland glaciers and is a major drainage outlet for a large portion of the western side of the ice sheet.

Accumulation Estimates from Ice Cores

The primary objective of this work is to use multi-species chemical analyses of 20- to 150-m-deep ice cores to estimate annual accumulation at various sites on the Greenland ice sheet. This will improve our estimates of snow accumulation over the ice sheet for comparison with ice discharge and for use in models. The core data are also analyzed for the interannual variability of snow accumulation rates, which is a major cause of short-period variability of ice sheet elevation and must be understood before we can infer long-term trends in ice sheet volume from observed surface elevation changes.

shallow cores
Drilling shallow cores to establish the historical accumulation and elevation of the ice sheet over the past few hundred years.

During the summer of 1995, 150-m firn and ice cores were drilled to determine annual accumulation rates at two Greenland sites, 73.84N, 49.49W (NASA-U) and 78.53N, 56.83W (Humboldt Glacier). Annual layers were identified in the cores using multiple parameters: d18O and concentrations of dust hydrogen peroxide, ammonium, calcium and nitrate. Using all parameters together to define annual layers resulted in a 350-year record for the NASA-U core with no dating uncertainty. For the lower-accumulation Humboldt core, the dating uncertainty is about 5 years over the 852-year period of record, with no uncertainty over the past 200 years. Annual accumulation over the periods of record at the two sites averaged about 0.34 and 0.14 m of water equivalent, respectively. There was no statistically significant trend in the NASA-U annual accumulation rates over the period of record. However, the Humboldt data do show an increasing trend of about 1.3 ± 0.4% per century over the period of record. A set of 20-m firn cores drilled near the main 150-m cores showed that interannual variability of accumulation exceeded spatial variability at NASA-U. The Humboldt cores showed equal spatial and interannual variability.

PICO four-inch drill setup at the NASA-U site in northwest Greenland.

In 1996, 120-m cores were collected at GITS (77.14N, 61.04W) and TUNU (78.10N, 35.50W), with several 20-m cores also collected at TUNU. In 1997, shallow firn cores were collected at eight sites around the perimeter of the Greenland ice sheet at approximately the 2500-m elevation contour. The 4-in. cores were collected by two- and sometimes three-person teams using the "sidewinder" coring device, a mechanically operated hand auger developed by the University of Nebraska. At those shallow coring sites that were co-located with automatic weather stations, more than one core was collected in order to investigate issues of short-scale spatial variability in snow accumulation. Two sites in northwest Greenland (75-76N) had average annual accumulation values of 0.30-0.36 m of water equivalent, and two sites in west-central Greenland (71-72N) had values of 0.40-0.42 m of water equivalent. All four sites had values that were only 70-80% of those estimated from prior work and call into question the accumulation "ridge" in western Greenland that is apparent in older data.

Accumulation Rates from Microwave Remote Sensing Data

Because of the volume scattering characteristics of microwave radiation in polar firn, data from both active and passive microwave sensors hold information about the grain size and layering characteristics at and below the surface of the snow pack, which are related to accumulation rates. One way in which the accumulation rates in the dry snow zones are being studied is based on the polarization-dependent microwave emission characteristics at 4.5-cm wavelengths. The approach is based on the fact that the density layering in the snow is of greater impact on horizontally polarized emission than on vertically polarized emission. Thus, the differences between the two can be used to the estimate mean annual accumulation rates. Researchers have had considerable success with this approach in Antarctica, with results comparing favorably to field observations, and their effort is being extended to Greenland. The resulting maps broadly agree with previous compilations of field observations.

A second method that is applicable to the dry snow zones is based on microwave emission at a wavelength of 1.55 cm. Emissivity at this wavelength has been shown to be related to accumulation by a hyperbolic function. The relationships have been demonstrated for the dry snow zones in Greenland and Antarctica, and microwave-based accumulation maps are being developed. The relationships at this shorter wavelength, however, are somewhat complicated by snow metamorphism and the development of hoar layers near the surface.

antenna horns
Two antenna horns of a radar instrument used to monitor the backscatter properties of snow and ice at the equillibrium line altitude.

Third, 5.6-cm backscatter data from spaceborne scatterometers show promise for estimating accumulation rates in the dry snow and percolation zones. The backscatter characteristics show a direct relationship with known accumulation patterns. The data have also been used to identify facies boundaries and detect changes in these boundaries over time. The results so far show a clear reduction in the location and extent of the dry snow zone since 1978, with the largest changes occurring in the southwestern part of the ice sheet. These changes are consistent with the decadal warming trend and increase of more than 1C between 1979 and the present. Finally, accumulation data have also been obtained by correlating stable isotope ratio profiles (dD and d18O) from snow pits with passive microwave brightness temperature trends. These analyses yield the amount, rate and timing of density-corrected accumulation at locations around Greenland with sub-seasonal resolution.

Greenland Precipitation Estimated from Atmospheric Analyses

Observations of precipitation over Greenland are limited, refer to different time periods, and are of uncertain accuracy. However, the analyzed wind, geopotential height and moisture fields are available for recent years. The objective of this work is to develop a dynamic method for retrieving precipitation over Greenland from these analyzed fields twice a day. Precipitation and accumulation over Greenland from 1957 to 1995 and their seasonal interannual variations are being estimated by this approach. How the atmospheric general circulation and weather systems control precipitation and accumulation over Greenland is also being studied.

The mean precipitation for 11 years, from 1985 to 1995, shows maximum values of more than 100 cm/yr along the southeastern coast and the southwestern edge of Greenland, with a secondary band of relatively high precipitation along the western coast. A large area of very low precipitation (less than 20 cm/yr) dominates the northern interior region.

Retrieved precipitation amounts for 1985-1995 are being compared with published accumulation maps, as well as the accumulation rates from recent ice cores. Preliminary analysis shows an average difference of 10% for these comparisons. The accumulation/precipitation data sets are compared by deriving the spatial distribution of their differences. The total annual accumulation over the entire ice sheet and its major regions is also computed from the different data sets. The annual accumulation computed by the dynamic method for the 1985-1995 period is lower than the observed data over a significant part of the ice sheet. This difference may be real, in that most
of the ice core observations were acquired before this period, and a decline in precipitation amounts over Greenland for the last three decades has been inferred from a variety of atmospheric methods.

Ice Sheet Mass Balance

Ice velocity has been measured at approximately 30-km intervals, mainly along the 2000-m contour line, around the entire ice sheet. Repeated GPS data were collected at these sites, with more than 400 individual site occupations at about 180 stations. The sites were established by planting an aluminum pole vertically in the ice, with a flagged bamboo pole planted nearby to make the site easier to find for the next occupation. Spacing between sites was usually 30 km, with some areas of denser spacing on the western slope of the ice sheet and an area with 40-km spacing in the northwest. Prior to 1995 the sites were visited by snowmobile traverse, but from 1995 the work was done using a ski-equipped Twin Otter aircraft.

Ice thickness was also measured along the velocity traverse using the CORDS airborne low-frequency radar. These two sets of measurements have been used to estimate the total volume of ice discharged across the traverse, and to compare this with total snow accumulation within the ice sheet catchment area inland from the velocity traverse, to infer the mass balance, or rate of thickening/thinning of the ice sheet. Initial results indicate that, taken as a whole, this interior portion of the ice sheet is almost in balance but with localized regions thickening or thinning by 10 cm or more per year.

This analysis is complete for most of the traverse, with a major gap in the southwestern quadrant of the ice sheet, where ice thickness has yet to be successfully measured. Deep ice in this region is warmer than elsewhere, severely reducing the radar penetration. Nevertheless, planned improvements to the CORDS depth sounder should resolve this problem during the 1998 field season.

The accuracy of the estimated mass balance using this approach is primarily determined by the accuracy of the assumed snow accumulation rates. Consequently 1998 plans for PARCA include a continued program of shallow coring to further improve our estimates of snow accumulation and its temporal variability over the entire ice sheet.

Localized Ice Thickness Changes

The objective of this project is to determine local rates of ice thickness change at various sites in Greenland. The results will be used to identify regions of the ice sheet where large changes are occurring, the causes of which can be investigated in future studies. The results will also be helpful for interpreting elevation changes detected by repeat airborne and satellite altimetry.

Precise measurements of vertical velocity are compared with the accumulation rate at the same location; if the two quantities differ, the ice sheet must be either thickening (when accumulation exceeds downward velocity) or thinning (when the reverse is the case). Vertical velocity is obtained from repeated GPS surveys of markers anchored at several depths in the firn or ice. By placing the markers at depth, vertical motion due to variations in snowfall and firn compaction is avoided. In most cases, markers are placed in hot-point-drilled holes, at depths ranging from 5 to 25 m. If markers can be anchored in solid ice, then the correction for firn compaction does not need to be made. At two sites, NASA-U and Humboldt, markers were installed in ice at the base of 150-m-deep holes from which other PARCA investigators have recovered cores.

Twelve sites have been installed on the Greenland ice sheet, and as of 1997, five sites have been studied in detail. Thickness changes range from +13 cm/yr (thickening) to -54 cm/yr (thinning).

Ice Velocity and Discharge Flux from Interferometric Synthetic Aperture Radar

The objectives of this research are to measure the ice discharge of the Greenland ice sheet as close as possible to the grounding line or calving front of outlet glaciers and to compare the results with mass accumulation and surface ablation in the interior to estimate the mass balance of the ice sheet.

The approach involves using multiple-pass ERS radar interferometry data to measure the velocity, topography and grounding zones of the outlet glaciers, in combination with surface topography derived from satellite altimetry. At the grounding line, ice thickness is derived from the glacier surface elevation, assuming that the ice is in hydrostatic equilibrium. Where no grounding line is present, ice thickness must be measured by other means, such as using ice sounding radar. Mass accumulation is obtained from published data, and surface ablation is calculated using a degree-day model.

Interferometric SAR observations of the Greenland ice sheet gathered by the European Space Agency's Earth Remote Sensing (ERS-1 and 2) satellites were used to map the grounding line and ice velocity of northern Greenland glaciers. Combined with an existing digital elevation model of north Greenland, limited radar echo sounding data and prior data on mass accumulation, the interferometric results were used to measure the grounding line ice discharge of 22 glaciers and to compare these estimated fluxes with the total mass accumulation in the interior, corrected for losses by surface ablation calculated from a degree-day model. The results suggest that basal melting is a major form of mass loss to the ocean from the northern sector of the Greenland ice sheet. Basal melt rates inferred assuming steady state conditions are ten times those recorded on Antarctic ice shelves. The mass balance analysis, recently complemented with new glacier additions, continues to indicate that the northern sector of the ice sheet may be slowly thinning at present. It has also been found that the grounding line of Petermann Glacier retreated by several hundred meters between 1992 and 1996, consistent with the suggested thinning trend of north Greenland glaciers.

A similar approach, combining ERS interferometry data from 1995-1996 and ice sounding radar measurements to be collected in 1998, is being used to characterize the ice discharge of the entire coast of East Greenland.

Ice Flow in the Northeast Greenland Ice Stream

SAR interferometry controlled by GPS field surveys is providing us with a detailed picture of ice flow in the major ice stream draining much of the northeast Greenland ice sheet, which was first detected in SAR imagery. The flow patterns in the onset area are more complicated than had previously been reported, with an apparent second tributary entering from the south. Downstream, the flow changes character and seems to deviate from what would be expected from simple balance flux estimates. This may be due to problems with the data used for estimating the balance, or it may be reflecting a variability in the system related to the surging behavior of Storstrommen Glacier.

In addition to the ice flow information, newly acquired radio-echo sounding profiles and altimetry data are helping to define the character of this large feature. Multiple-azimuth photoclinometry using AVHRR data is also improving our understanding of the surface topography that is produced by the rapid ice motion. Initial results from this work show that the margins of the stream are topographic troughs and that the undulation field on the stream is strong enough to produce closed basins in several areas.

Greenland Ice Surface Elevation Changes from Satellite Radar Altimetry

Estimates of the overall mass balance and seasonal and interannual variations in the surface mass balance are obtainable from time series of ice surface elevations measured by satellite altimetry. Although satellite radar altimetry has significant limitations in coverage and accuracy over sloping surfaces, information on ice sheet surface elevation changes has been derived for central parts of the Greenland ice sheet.

One group of investigators re-examined elevation change estimates for the Greenland ice sheet by incorporating a global method for analyzing altimeter orbit error present in the ice sheet data sets. Because the predominant radial orbit error is a long-wavelength signal concentrated at the circular frequency of the orbital period, they used a global analysis of ocean altimeter data sets. For the Geosat-Exact Repeat Mission (ERM: 1987-1988) and Seasat (1978) mission, crossover data sets were created with respect to a reference (mean) ocean surface. The radial orbit error was identified by passing the sea-height crossover residuals through a stochastic filter designed to estimate the time-varying amplitudes of the sinusoidal orbit error function and the measurement system bias that may be present in inter-satellite comparisons.

The resulting orbit error corrections were applied to the Seasat and Geosat-ERM Greenland data sets, and the elevation change was analyzed by dividing the average change in elevation by the average time interval using all crossovers. A spatial analysis reveals large geographic variations in elevation change from -15 to +18 cm/yr. After isostatic adjustment, a spatial average of 32,283 ERM ¥ Seasat crossovers yields a growth rate of 1.7 ± 0.5 cm/yr from 1978 to 1988. This growth rate is less than 10% of that calculated in earlier analyses, primarily because of improvement in our knowledge of the satellite orbits. Given the large spatial variations in elevation change, this averaged growth rate is too small to determine whether or not the Greenland ice sheet is undergoing a long-term change due to a warmer polar climate.

Elevation changes obtained between Seasat (1978) and Geosat (1985-1989) are affected by the short three-month period of Seasat and interannual variations in the seasonal cycle, and in a separate analysis, the ice thickening rate derived from four years of Geosat data (1985 to 1989) for the area south of 72N was found to be about 7 cm/yr. Seasonal variations in elevation, caused by variations in snowfall, firn compaction and melting, are also observed. A consistent set of improved satellite orbits and altimeter corrections have been used in this analysis. Ice thickening is observed in the southern portion of the ice sheet by about 10-20% of the mass balance. Seasonal variations in the surface elevation are also observed, ranging from a 14-cm peak-to-peak amplitude cycle with a minimum in July at elevations of 2200-3300 m to about 2 m with a minimum in late September in the upper ablation zone.

hand auger
Drilling a 10-meter hole witha hand auger to insert a thermistor chain for snow and ice temperature recordings. This is done at each automatic weather station of the Greenland Climatic Network (GC-Net).

Preliminary analysis of ERS-1 radar altimetry data (1992-1996) above 76N suggests a thinning of 11 cm/yr above 2700-m elevations on the northeast part of the ice sheet and a thickening at lower elevations with a 10-cm/yr maximum at 1200- to 1700-m elevations. On the northwest part of the ice sheet, the data suggest a thinning of 20 cm/yr above 2200 m and a thickening at lower elevations, with a 24-cm/yr maximum at 1200- to 1700-m elevations. These results from ERS-1 data are very preliminary and are not confirmed by other PARCA investigations.

The analysis of radar altimetry data to measure changes in elevation is fraught with problems. Beginning in 2001, NASA's Geoscience Laser Altimeter System (GLAS) will significantly improve the range accuracy, orbit accuracy and spatial coverage for measurements of seasonal and long-term changes of ice sheet elevations.

Greenland Network of Automatic Weather Stations

The Greenland Climate Network (GC-Net) was initiated in 1995 with the goal of monitoring climatological and glaciological parameters at various locations on the ice sheet. Stations have been added each year, so that the present network includes 14 stations and monitors 350 parameters. So far the GC-NET database contains more than 14 station-years of measurements that have been quality controlled and calibrated.

The tasks of calibration and quality control have been updated for the majority of GC-Net stations. Documentation of the data sets and the description quality control methods have been generated. Satellite-transmitted data are currently being used to extend the GC-NET record to the present. The quality control task is complicated by the fact that up to 20% of the data are not transmitted. This set of procedures was developed to optimize the data record using interpolation techniques. Once a station is revisited, continuous data can be retrieved to replace the transmitted data. Quality control procedures are applied to all AWS data sets.

An interesting cross correlation was found between wind speed and surface temperature, with a coefficient of better than 0.5. The wind speed increase does precede the temperature increase by approximately 6-10 hours. This relation was observed mainly during winter and has important implications for microwave satellite remote sensing and for the interpretation of shallow ice cores. Because of the warmer air and the wind-pumping effect of strong wind, the snow cover temperature increases near the surface. If the temperature gradient is sufficiently large, depth hoar layers form in the middle of winter.

A mean monthly lapse rate was derived along the Jakobshavn-Summit profile of 0.6-0.7C per 100 m for spring and summer and 0.8-1.0C per 100 m for fall and winter. The katabatic wind along the same profile showed an increase in directional constancy towards the coast. The interannual variability of snow accumulation at various GC-Net sites, as well as the radiative and turbulent energy exchange, has been studied in detail.

Multiple snow precipitation events were identified using AWS data. Typical accumulation events are characterized by higher than normal temperatures, wind speeds increasing with the approach of a storm and then diminishing as soon as snowfall begins, an increase in downwelling long-wave radiation flux, an increase in boundary layer stability as reflected by positive Richardson numbers, an increased surface albedo, and decreasing pressure. Humidity was observed to approach saturation during snow events yet was preceded by a short dry phase just before snowfall began, indicating the invasion of a different air mass. For the Humboldt AWS, the wind direction during snow events is westerly, indicating that Baffin Bay is an important moisture source region. At TUNU-N, snow events are associated with winds from the north and east.

Estimates of Ablation Rates on the Greenland Ice Sheet from Passive Microwave Observations

The primary objective of this project is to estimate the ablation rate of the Greenland ice sheet using SMMR (scanning multichannel microwave radiometer) and SSM/I (special sensor for microwave/imaging) data. Passive microwave satellite sensors have been used to map the spatial extent and frequency of snowpack melting on the ice sheet; a continuous time series is available since October 1978. The derivation of melt frequencies from microwave data is based on the increase in microwave emissivity as liquid water is introduced into a previously dry snowpack. A simple radiative transfer model is used to estimate the SMMR and SSM/I 37-GHz, horizontal-polarization brightness temperatures associated with melting snow and ice. The modeled brightness temperatures are used as thresholds. If the observed brightness temperature exceeds the modeled brightness temperature, melt is said to have occurred that day.

Time series of brightness temperatures have been extracted from the Pathfinder SMMR data at 100-m elevation increments from 900- to 2700-m elevations near long-term climate monitoring stations at Qamanarssup sermia (64N) and Patiksoq (70N) in western Greenland. Microwave-derived annual melt duration (in days) was compared to the melt duration estimated from mean monthly surface temperature data for the Qamanarssup sermia and Patiksoq transects. The same time period was used to calculate average annual melt durations from the surface temperature data. Mean monthly temperatures from long-term climate monitoring stations, augmented by coastal weather stations, are used to calculate the number of days with melt. The number of days with temperatures greater than 0C, based on an assumed normal temperature distribution, is calculated for each month and then summed to arrive at the annual melt duration. Along the Patiksoq transect there is no more than a three-day difference in annual melt duration at any elevation between the satellite and surface temperature estimates of melt duration. A snow and ice melt model is used to account for different degree-day factors for snow and ice surfaces and to account for the formation of superimposed ice. Mass balance calculations give an overall average annual accumulation of approximately 530 km3/yr and a mean annual ablation rate of 230 km3/yr. The ablation rate shows a large range during 1988-1996 from 100 to 500 km3/yr, but most years experienced 200-300 km3/yr. Approximately 13% of the ice sheet experiences net ablation, but this varies from 8% (1992) to 23% (1995).

Absolute Gravity and GPS Measurements in Greenland

Ongoing changes in the distribution and volume of ice in Greenland could cause vertical crustal motion of up to several millimeters per year or more around the edges of the ice sheet. By measuring this motion, it should be possible to learn about those changes in ice. However, the viscoelastic response of the earth to past changes in ice loading could cause vertical motion rates that are of the same order. These viscoelastic effects must be known before crustal motion observations can be used to help constrain the mass balance of the ice caps. Theoretical arguments suggest that this can be done by combining vertical motion measurements with simultaneous observations of time variation in gravity.

Multi-year simultaneous measurements of absolute gravity and GPS crustal motion are being made at two bedrock sites at the edge of the Greenland ice sheet: Kangerlussuaq and Kulusuk. Continuous GPS measurements have been made at Kangerlussuaq since July 1995, and three two-week occupations with an absolute gavimeter have also been carried out. GPS results indicate that the crust is subsiding at a secular rate of about 6-8 mm/yr. The absolute gravity measurements are consistent with the GPS measurements. These preliminary results suggest that the ice sheet may be thickening at the rate of a few tens of centimeters per year averaged over a few hundred kilometers from Kangerlussuaq; however, a longer data span is required to support this interpretation.

Atmospheric Chemistry and Ozone

NASA supports a number of tasks related to measuring and understanding chemical and dynamic processes in the Arctic atmosphere from the upper troposphere through the stratosphere and lower mesosphere. Such studies include the measurements and subsequent data analysis using space-, aircraft-, balloon- and ground-based instruments. These activities are accomplished through NASA's Upper Atmosphere Research Program (UARP) and Atmospheric Chemistry Modeling and Analysis Program (ACMAP), as well as a number of space flight missions. A particular focus of the work supported by these programs is on the seasonal, annual and long-term changes in Arctic stratospheric ozone and the atmospheric constituents and processes that affect this change.

The 1997 Photochemistry of Ozone Loss in the Arctic Region in Summer (POLARIS) aircraft campaign primarily used the NASA ER2 and balloon platforms based in Fairbanks, Alaska, to measure selected species within the reactive nitrogen (NOy), halogen (Cly) and hydrogen (HOx) families, aerosols and other long-lived species in the lower and middle stratosphere. The POLARIS campaign included a total of 30 ER2 flights and three balloon flights during three deployment periods in 1997: 17 April to 15 May, 24 June to 13 July, and 3 September to 27 September. These measurements, along with computer models of the atmosphere, meteorological data and satellite data, are being used to understand the seasonal behavior of Arctic stratospheric ozone as it changes from very high concentrations in spring down to very low concentrations in autumn due to chemistry and transport at high latitudes. This behavior has been attributed to an increased role of NOx catalytic cycles for ozone destruction during periods of prolonged solar illumination, such as occur at high latitudes during summer. Thus, the detail with which current photochemical models can describe this large natural signature in ozone serves as an indication of how well the role of increased stratospheric NOx from anthropogenic sources, a possible consequence of future supersonic transport, can be quantified.

Measurements of the distribution of total ozone over the Arctic are measured with the Total Ozone Mapping Spectrometer (TOMS) series of instruments. Two TOMS instruments have provided data in the last two years. One, launched aboard NASA's Earth Probe (EP) satellite in July 1996, has been operated routinely since launch. Another, aboard the Japanese ADEOS satellite, operated from shortly after its launch in August 1996 until the failure of the ADEOS satellite in June 1997. Although they were launched into different orbits, which affected their spatial coverage in the tropics and midlatitudes, both satellites have provided full daily coverage of the sunlit Earth at high latitudes (poleward of approximately 50o).

The TOMS observations have proved useful in showing unusual ozone distributions over the Arctic in the winter of 1997. In particular, there was a region of large ozone depletion over the Arctic that winter, prompted largely by the presence of low temperatures throughout much of that winter and a lack of wave activity that allowed for a stable polar vortex to last well into March. At that time of spring, there is enough sunlight for photochemical ozone-depleting processes, such as those that take place over the Antarctic during September, to occur over the Arctic. The lowest ozone column amounts seen over the Arctic in the winter of 1997 were as much as 40% below comparable amounts seen early in the TOMS record (1979-1980), when the Nimbus 7 satellite provided TOMS data. The region of ozone depletion was centered over the pole for most of its existence, which lasted until the breakdown of the polar vortex in the stratosphere. In 1997 this did not occur until April.

The lowest ozone column amount seen in the Arctic in 1997219 Dobson Units (DU, 1 DU = 1 milli atm cm, or corresponding to 2.67 ¥ 1016 molecules cm-2)are still much above the corresponding amounts seen over the Antarctic in September/October, and the region of low ozone seen in the Arctic is appreciably smaller (roughly threefold) than that seen in the Antarctic (which has lower ozone amounts). Demonstration that this ozone depletion takes place in the lower stratosphere was provided by data from the Upper Atmosphere Research Satellite (UARS), as well as from balloon-borne ozonesondes and ground-based lidars. The lowest ozone values observed in 1996 (for which TOMS data were not available, but which were obtained using the solar backscatter ultraviolet instrument aboard NOAA's NOAA-9 operational meteorological satellite) were also appreciably lower than in prior years, but not nearly so low as in the winter of 1997 (maximum depletion some 24%, rather than 40% as in the latter year, and centered in the region northeast of Greenland rather than over the pole as in 1997). Ozone levels in the Arctic winter in 1998 were not nearly so low as in 1997 because of the very different meteorological conditions those two years.

Evidence for downward descent in the wintertime northern hemisphere polar vortex has been obtained from analysis of spaceborne data, especially that from the Atmospheric Laboratory for Applications and Science (ATLAS) and UARS. For example, analysis of data obtained during the second ATLAS flight (April 8-16, 1993) showed that in the winter of 1992-93, average descent rates ranged from 0.8 km/month at 20 km to 3.2 km/month at 40 km. These descent rates are comparable to those observed over the Antarctic during the third ATLAS mission (November 1994), but the overall descent distance in the Arctic is smaller because the vortex does not persist as long there. In the ATLAS-2 observations, a clear distinction could be seen in the distribution of long-lived tracers inside and outside the vortex in the middle and lower stratosphere.

Measurements of trace chemical species in the Arctic stratosphere are made regularly using the UARS. A particularly important set of measurements are those of the microwave limb sounder (MLS) instrument, which measures O3 and ClO. MLS measurements have been very useful in pointing out the abundance of high levels of ClO in the Arctic. Record high ClO and low ozone values were recorded at high northern latitudes in 1997. Because of slightly higher temperatures and less persistent confinement, Arctic ozone losses have not been as severe as in the Antarctic. Interannual variability in the dynamical conditions of the Arctic vortex is in general greater than in the Antarctic, and ozone distributions in the winter and spring will therefore have much greater interannual variation than in the Antarctic. A slight cooling of the stratosphere due to increasing concentrations of radiatively active gases has the potential to exacerbate ozone depletion in the Arctic through the formation of increased amounts of polar stratospheric clouds.

Ground-based Fourier transform spectrometer (FTS) measurements at Sondre Stromfjord, Greenland, (67.02N, 50.72W) had been made by researchers from the National Center for Atmospheric Research from October 1994 to April 1995. A new instrument (Bomem 120-M) was installed in the summer of 1997 at Thule, Greenland, (76.53N, 68.74W) and has begun obtaining data. Conducted under UARP support, these measurements are a component of the international Network for Detection of Stratospheric Change (NDSC) and have yielded column abundances for HCl, HF, O3, HNO3 and N2O. The results have been compared with similar observations made during the Second European Stratospheric Arctic and Mid-latitude Experiment (SESAME) and have been presented in a special issue of the Journal of Atmospheric Chemistry. Another recent NDSC Arctic activity was the ozone and aerosol lidar intercomparison held in Ny Ålesund, Spitsbergen, (78.92N, 11.93E) during the winter of 1997-98.

NASA's UARP and ACMAP continue to support measurements and multidimensional models for atmospheric chemistry and transport needed to study the global atmosphere. The mission results presented above underscore the necessity to include the full range of dynamical, radiative and chemical processes in models that are used to predict future ozone losses given prescribed scenarios for emission of CFCs, their substitutes and other chlorine- and bromine-containing compounds.