Skip Navigation

May-June 2016

ResizeResize Text: Original Large XLarge

Weather Queries

Is the Antarctic ice sheet losing mass or staying the same?

Bill Haaf

Kennett Square, Pennsylvania

The Antarctic continent as a whole, unlike Greenland, is probably gaining ice mass, and thereby not contributing to sea-level rise. Parts of the Antarctic ice sheet are gaining mass, while other parts are losing it. See, for example, Figure 1, a NASA photo from West Antarctica, which is losing ice to the sea at an accelerating rate. Just a few years ago, it was thought that Antarctica as a whole was losing mass. Determining whether this continent is gaining or losing ice mass is anything but straightforward. How scientists draw their conclusions makes for an interesting narrative.

Changes in ice sheet mass are estimated in three very different ways:

  • 1. Mass budget: This method keeps track of annual precipitation (the gain), and estimates losses to the atmosphere by sublimation (conversion of solid ice to vapor), and to the ocean by strong winds blowing snow offshore and, much more important, by huge glaciers flowing into the sea.

  • 2. Radar and laser altimetry: Changes in elevation of the ice sheet can be measured accurately from satellite. Radars bounce microwaves off the surface of the ice and time precisely how long it takes for the signal to make the round trip from satellite to the ice surface and back. The travel time determines the distance from satellite to ice sheet. The Global Positioning System locates the satellite very precisely, and this allows estimation of surface elevation. If surface elevation changes with time, gain or loss of ice is indicated. Lasers send pulses of visible and infrared light earthward and operate on the same principle as radar.

  • 3. Gravimetry: The acceleration due to gravity at the Earth's surface is usually given as g = 9.8 meters per second per second (m s−2). But g varies for a number of reasons:

    • Altitude: g varies as 1/r2, where r is the distance from the Earth's surface to its center. Thus g decreases with altitude.

    • Latitude: The shape of the earth is not quite spherical but somewhat flattened, having the greatest diameter at the Equator and the least at the poles. The poles are thus closer to the center of the earth than the Equator. g increases with latitude, being a maximum at the poles and a minimum at the Equator.

    • Distribution of mass within the earth: The density of the solid earth is not uniform, and therefore gravity varies slightly depending upon where you are, even at sea level.

Satellites can measure minute variations in g, very much smaller than 1%. Temporal changes in g at a given location on an ice sheet can indicate a gain or loss of ice mass.

The acceleration of a falling object is affected by centrifugal force exerted by the spinning Earth. This force depends on latitude (zero at the poles, maximum at the equator), but is here excluded from the definition of g because it is not a gravitational force.

Let's examine these three methods in more detail.

Mass Budget

Antarctica occupies 74% more land surface than the Lower 48 United States, and yet, because the Antarctic climate is so cold and harsh, there are probably no more than 100 surface weather stations there, and most of them are automated. Their numbers are insufficient to permit any but the most crude estimates of annual precipitation across Antarctica. Increasingly, scientists use global reanalyses, estimates of atmospheric conditions at regular intervals derived from all sources of global observations, supplemented by prediction models that can fill data voids by extrapolation in a physically consistent way. Estimates of annual precipitation from reanalyses compare well with accumulated station precipitation.

Glaciers, mountains, an ice shelf, and a frozen sea surface with large embedded icebergs are visible in this NASA aerial photo from West Antarctica, taken at low sun on October 29, 2014.

Most of the mass loss from the Antarctic ice sheet occurs where huge glaciers descend the steep margins of the continent and empty into the sea. Many of these glaciers are impeded from flowing straight into the ocean by ice shelves—huge blocks of ice, hundreds of feet thick—that float on the ocean surface. The place to measure mass loss is not at the edge of the ice shelves, because floating ice does not contribute to sea-level rise. (Actually, it can have a very small effect on sea level when it melts because the meltwater is fresher than the saline seawater and has lower density.) Rather, it is at the grounding line, the place where glacier ice still rests on solid ground, but immediately adjacent to where sea water lies under the ice. To measure the ice loss at the grounding lines of all large glaciers in Antarctica, scientists must estimate the cross-sectional area of the ice over the grounding line and the speed at which the ice is moving. Airborne radars have estimated the ice thickness for many glaciers. Repeated surveys of markers on the ice or high-resolution images from satellite overpasses help to estimate glacier movement over the grounding line.

Radar and Laser Altimetry

Over the last two decades, satellites providing data from radar or laser altimeters have been launched by the European Space Agency (ESA) and the National Aeronautics and Space Administration (NASA). Here are the ESA satellites, all of them carrying radar altimeters:

  • ERS–1, the European Remote Sensing satellite, operated from 1991 to 2000.

  • ERS–2 operated from 1995 to 2011.

  • Envisat, successor to the ERS satellites, operated from 2002 to 2012.

  • CryoSat–2, studying the Earth's cryosphere (especially ice sheets), went into orbit in 2010. (CryoSat–1 failed at launch in 2005.)

The NASA satellite, ICESat (Ice, Cloud, and Land Elevation Satellite) was launched in 2003 and produced good data through fall of 2009. ICESat carried a laser altimeter.

Both radar and laser altimeters look straight down, but they differ in several important ways:

  • 1. The footprint is the area surveyed on the ground by an instrument aboard a satellite. Radar altimeters have a footprint of 2 to 20 kilometers across, depending on the beam width of the antenna and the altitude of the satellite. The footprint of laser altimeters is much smaller, about 70 meters across, giving them a much better chance of resolving surface elevations in steeply sloping or rough terrain.

  • 2. Radar altimeters operate in the microwave part of the electromagnetic spectrum. Their pulses can easily penetrate clouds. Laser altimeters operate at much shorter wavelengths in the visible or infrared parts of the spectrum. Their pulses cannot penetrate clouds—a significant disadvantage.

  • 3. Lasers give very good along-track resolution with high precision. However, tracks are widely separated, especially at lower latitudes. Radar altimeters give about 75% coverage over most of the Antarctic ice sheet.

Radar and laser altimeters can estimate annual changes in ice surface elevation to within a few centimeters. Much of the elevation change is caused by accumulated snowfall, but some may be caused by glacial rebound. When ice sheets slowly thicken over thousands of years, their weight depresses the earth's surface underneath. As the ice sheet thins and recedes, glacial rebound causes the rocky underlayment to rise slowly toward its original equilibrium elevation. This process is called glacial isostatic adjustment (GIA). The effect of GIA is estimated to be 0.2 to 0.6 centimeters of uplift per year in West Antarctica, on the Antarctic Peninsula, and toward the coasts of East Antarctica, but the Earth's crust may still be sinking in the interior of East Antarctica. The map in Figure 2 distinguishes these three regions and shows the major glacial drainage basins as well, as determined from ICESat surface elevation measurements.

The depth of bedrock beneath the Antarctic ice sheet is fairly well determined. Because the surface elevation of the ice sheet is accurately measured at regular intervals by satellite, it is possible to estimate temporal changes in the volume of the ice sheet, with adjustments for GIA mentioned above. Changes in the mass of the ice sheet are what matter for sea level rise, and the problem is how to convert from volume changes to mass changes.

Firn normally pertains to the layer of fallen snow on a glacier that may partially melt during the summer, but it neither runs off—it refreezes—nor does it morph into glacier ice. Antarctica is so cold that melting is rare (nonexistent in the interior), and so Antarctic firn refers to the intermediate layer between fresh snow and glacial ice, having densities between about 350 and 900 kilograms per cubic meter (kg m−3). (920 kg m−3 is the density of pure ice with no air bubbles). As the fallen snow ages, its density increases, but the rate of metamorphosis depends on temperature, wind speed, and depth of burial. At the South Pole, where the climate is exceptionally cold, seldom windy, and dry, the firn layer is nearly 100 meters thick. At the Antarctic coast, where it is warmer, windier, and more precipitation falls, the firn layer may be 40 m thick. Ice scientists have devised sophisticated models that use analyzed atmospheric conditions to estimate how the density changes within the firn layer and how quickly. Such models compare favorably with density measurements obtained from core samples at dozens of locations. Firn densification models make it possible to convert volume changes of the ice sheet to mass changes, and thereby to estimate the contribution of mass loss to sea-level rise.


In 2002, NASA launched two satellites in identical orbits at an altitude of 460 km, one lagging behind the other by about 220 km. This is part of the GRACE mission. GRACE stands for Gravity Recovery and Climate Experiment. These satellites can measure minute changes in the Earth's gravitational field in the following way. If the leading satellite in orbit encounters slightly greater gravitational pull, it accelerates slightly, increasing the distance between it and the trailing satellite. The distance can be measured with extreme precision—down to a few millionths of a meter, considerably less than the width of a human hair. If the leading satellite encounters slightly less gravitational pull, it slows down, decreasing its distance from the following satellite. By keeping tabs on the separation distance, NASA scientists can infer variations in the earth's gravitational field, which GRACE can map globally about once every 30 days.

Because of the spacing of the satellites, the GRACE mission can measure only smoothed variations in the gravity field with a resolution of roughly 300 kilometers. An advantage of this measurement is that it is directly sensitive to the distribution of mass on and beneath the earth's surface. Temporal changes in the gravitational field reveal changes in the distribution of mass. The challenge is to separate changes in the mass of ice from changes occurring beneath the ice.

The maximum density of ice is about 900 kg m−3, whereas the density of the bedrock below ranges from 3400 to 4000 kg m−3. Geodesists apply models of glacial isostatic adjustment (discussed earlier) to isolate the changes in the mass of the Antarctic ice sheet.

Estimates as of 2012

In November 2012, a paper by Andrew Shepherd and 46 coauthors (truly a group effort) appeared in Science magazine. It attempted to reconcile all types of measurements described above to arrive at a best estimate of the ice mass balance in Antarctica. The authors divided the continent as shown in Figure 2.

Figure 3 summarizes the results of this amalgamation of many previous studies. The top half of the figure pertains to Antarctica alone. The vertical axis gives the cumulative gain (+) or loss (−) of ice mass in gigatons (Gt) since 1992. (A gigaton is a billion metric tons, and a metric ton is 1,000 kilograms. A gigaton is also 1.0 cubic kilometer of water and 1.1 cubic kilometers of ice.) The top curve (green) is for East Antarctica, which has actually gained mass since 1992. The Antarctic Peninsula (orange curve) maintained ice mass through 2000, but has been slowly losing it since about 2000. West Antarctica has been losing mass since about 1996 and, in recent years, at an accelerating rate. Note the vertical axis on the right. This gives the contribution (in millimeters) to sea level rise from melting ice or sea level reduction when water becomes locked up in the ice sheet, as in East Antarctica. If 362 Gt of ice is lost to the ice sheet, sea level rises by one millimeter.

As a side note, the bottom of Figure 3 shows the cumulative loss of ice mass from all of Antarctica (purple curve), Greenland (blue curve), and both together (black curve) since 1992. The loss from Greenland is about twice the loss from Antarctica. The colored shading in Figure 3 represents the margin of error (the uncertainty) in the estimates. It is fairly large.

Estimates as of 2015

A very recent revised estimate of the Antarctic ice mass balance was published in the Journal of Glaciology late in 2015 by H. Jay Zwally and five coauthors. This estimate was based on some of the same data as considered in the paper by Shepherd and others, but the conclusion is substantively different: that the overall mass of the Antarctic ice sheet is growing. Zwally and coauthors tried to improve several aspects of the data processing in order to reduce the uncertainty of the estimates. This included using meteorological reanalysis and firn compaction modeling to obtain refined estimates of firn and ice densities, needed to convert measured elevation changes to estimates of mass changes. They also discuss why previous estimates of mass change differ from theirs.

For example, both satellite altimetry and gravimetry require corrections for motion of the bedrock underlying the ice sheets. Zwally and coauthors used essentially the same models of bedrock motion as Shepherd and others, but they note that gravimetry is about six times as sensitive to errors in the bedrock models as altimetry. The bedrock models are very dependent on the history of the ice loading/unloading that depresses the earth's crust or allows it to rebound. Zwally and coauthors argue that, because of ice loading over the past 10,000 years, the bedrock could still be depressing at a rate of 2.65 millimeters per year. A depression rate of less than this, only 1.6 millimeters per year, is all that is needed to increase the gain in mass in East Antarctica as estimated from altimetry by 15 Gt per year and to increase the gain as estimated from GRACE gravimetry by 90 Gt per year, bringing them in agreement at 150 Gt per year for East Antarctica. This small change in the modeled GIA is enough to tip the balance for the entire continent from negative to positive for the GRACE results, as well as for the altimetry.

Figure 4, taken from the paper by Zwally and coauthors, shows more detailed results. Drainage basins are numbered as in Figure 2. The top panel is for 1992–2001; the bottom panel is for 2003–2008. Latitude circles are numbered in increments of 5° from −65 to –85°S. The annual gain (+) or loss (–) in ice mass over the continent is color-coded as indicated in the color bars at the bottom of each panel. The mass balance dM/dt is given in units of Gt a−1 (gigatons per year) for each panel. The geographic designations are defined in Figure 2. Mass balances for each region appear in the four corners of each panel. “All” at the upper right pertains to the balance for the entire continent.

The interior of East Antarctica, where it is very cold and dry, is gaining mass very slowly, but this is happening over a huge area. More rapid gains in mass occur near the coastlines (orange, pink, and red colors), where precipitation is greater. West Antarctica is divided into WA1, where ice is being lost to the ocean at an increasing rate (green and blue colors), and WA2, where accumulation outweighs losses. The Antarctic Peninsula (AP) is losing much more ice in the later period (−29 Gt a−1) than in the earlier period (−9 Gt a−1), perhaps not surprisingly, because this region has experienced more rapid warming in recent years than most other places in the world.

The overall mass balance is 112 ± 61 Gt a−1 from 1992–2001 and 82 ± 25 Gt a−1 from 2003–2008. The uncertainty (the number after the ±) is a substantial fraction of the value itself.

Note the substantial mass losses near the coast (green and blue colors), especially in WA1. A few large, floating ice shelves have disintegrated in recent years, partly because of bottom melting at the base of the ice, which is in contact with relatively warm ocean water. Ice shelves tend to retard the movement of glaciers descending from the interior, but, when they break up, glaciers accelerate to many times their former speed, and mass loss to the ocean increases correspondingly.

It may come as a surprise that reputable scientists, working with the same observational data, conclude in 2012 that the ice mass balance in Antarctica is negative, and in 2015 that it is positive. The history of research on the ice sheet is one of estimates that are slowly refined with each passing year, as data processing and interpretation become more sophisticated and new data become available. Still, the uncertainties are substantial enough that we have probably not heard the last word on the subject.


I am searching for twenty-four specific Weatherwise back issues to complete my collection. If you have any from volumes 1 through 4 (1948 through 1951) in good condition, please e-mail me at: I will pay $15 per issue.

Weatherwise Contributing Editor THOMAS W. SCHLATTER is a retired meteorologist and volunteer at NOAA's Earth System Research Laboratory in Boulder, Colorado. Submit queries to the author at, or by mail in care of Weatherwise, Taylor & Francis, 530 Walnut Street, Suite 850, Philadelphia, PA 19106.

I thank Jay Zwally, of NASA's Cryospheric Sciences Laboratory, for help in answering this question.       

In this Issue

On this Topic

Privacy Policy

© 2018 Taylor & Francis Group · 530 Walnut Street, Suite 850, Philadelphia, PA · 19106