SOTC: Ice Sheets
Like a glacier, an ice sheet forms through the accumulation of snowfall, when annual snowfall exceeds annual snowmelt. Over thousands of years, the layers of snow build up, forming a flowing sheet of ice thousands of feet thick and tens to thousands of miles across. As the ice thickens, the increasing height of snow and ice causes the ice sheet to deform and begin to flow.
Unlike a glacier, which generally flows in one direction, an ice field flows outward in all directions from the center. If an ice field covers more than 50,000 square kilometers (20,000 square miles), it is defined as an ice sheet. Although ice sheets covered much of the Northern Hemisphere during a series of Pleistocene Ice Ages, the Earth now has just two major ice sheets, one on Greenland and one on Antarctica.
The Greenland Ice Sheet covers roughly 1.7 million square kilometers (650,000 square miles). The Antarctic Ice Sheet covers nearly 14 million square kilometers (5.4 million square miles), and is divided into three sections: the East Antarctic Ice Sheet, the West Antarctic Ice Sheet, and the Antarctic Peninsula.
Together, the Greenland and Antarctic ice sheets hold about 99 percent of the world's freshwater ice. If the Greenland Ice Sheet melted away completely, sea level would rise roughly 7 meters, or 23 feet (Gregory et al. 2004). If the entire Antarctic Ice Sheet melted, sea level would rise by about 57 meters, or 187 feet (Lythe et al. 2001). While this is unlikely for the foreseeable future, even a partial loss of these huge ice masses could have a significant effect on coastal areas. At present, both ice sheets are shrinking, but the rate is small (in terms of sea level contribution, on the order of about 1 millimeter per year).
Ice sheet structure, flow, melting, and fracture
The Greenland and East Antarctic Ice Sheets are roughly 3,000 to 4,000 meters (10,000 to 13,000 feet) high at their summits. The West Antarctic and the Antarctic Peninsula Ice Sheets are about 2,500 meters (8,200 feet) high.
Ice sheet flow is a function of surface slope and ice thickness. Near the summit of the ice sheet, where the slope is the lowest, flow speeds are generally a few centimeters to a few meters per year. Along fast-flowing outlet glaciers, ice speeds can reach hundreds of meters or even several kilometers per year.
Ice sheets flow outward from their dome-like centers, where they are generally thickest, and push ice outward until they encounter ocean, or where climate is warm enough to melt the ice faster than the combined flow rate and winter snowfall. In areas where summer surface melt exceeds winter snowfall, old interior layers in the ice sheet are exposed. The ice sheet becomes thin, meltwater runs off the surface of the ice, and the ice sheet may terminate on land. However, for much of Greenland and Antarctica, ice flow terminates at the ocean, as a tidewater glacier (not fully afloat) or an ice tongue or ice shelf (fully floating thick permanent ice above the ocean). In these areas, the location of the edge of the ice sheet is very sensitive to both ocean condition and the amount of ice fracturing (crevasses or rifts). Areas with some ocean heat can rapidly melt the floating ice from the underside, thinning the ice sheet and making it weaker. Stresses from ice flowing over bedrock or around islands causes fracturing, and at the front edge of the ice this fracturing leads to iceberg calving.
Climate, weather, and ice sheets
Ice sheets and cyclonic storms (low pressure cells) have a complex interaction. Most of the moisture and energy in a storm is in the lower part of the atmosphere. As a storm approaches an ice sheet, it encounters the steep slopes of the ice sheet edges, and the air is lifted and cooled. This leads to heavy blizzards and snowfall along the ice sheet margins. By the time the air masses reach the center of the ice sheet, they are stripped of most of their moisture. As a result, snow accumulation is typically very low near the summit of an ice sheet. In addition, the large bulk of the ice sheet, like a wall or building in the wind, can redirect storms around the ice sheet.
The high-elevation center of an ice sheet also plays a role in driving a peculiar kind of local weather created by the ice sheet itself. Over the center of an ice sheet, the air is typically dry, and skies are clear. Heat radiates to space from the ice sheet surface. This chills the surface and the layer of air just above the ice, creating an inversion of cold, dense air near the ice surface and warmer air above. Gravity then pulls the dense layer of cold air downhill. As it flows down the flanks of the ice sheet, the cold air layer picks up speed. By the time it reaches the coast, hurricane-force winds, known as katabatic winds, result. In contrast to coastal storms, katabatic winds can be bone dry. Cold, dry winds over the surface can lead to ice evaporation of snow and exposure of ice. "Blue ice" seen in many satellite images results from this process, as do the famous "dry valleys" of Antarctica, where the ice sheet has been completely evaporated away.
Thus, there are four ways that ice in the ice sheet may be lost: ablation (evaporation of the ice), surface melt, calving at the interface with the ocean, and melting from contact with the ocean. Mass gain occurs almost entirely by snowfall, although in a few areas rainfall on the snow can add a small fraction to the mass input.
Ice sheet mass balance
A key area of glaciological study in recent years is ice sheet mass balance. The mass balance of an ice sheet is the difference between its total snow input and the total loss through melting, ablation, or calving. So long as an ice sheet gains an equal mass through snowfall as it loses through melt, ablation, and calving from glaciers and ice shelves, it is said to be in balance. Because ice sheets contain so much ice and have the potential to raise or lower global sea level so dramatically, measuring the mass balance of the ice sheets and tracking any mass balance changes and their causes is very important for forecasting sea level rise. Scientists monitor ice sheet mass balance through a variety of techniques. No measurement method is perfect, however, and ice sheets' sheer size makes exact measurement difficult.
Measurement techniques for ice sheet mass balance
Scientists have adopted three general approaches to ice sheet mass balance measurement: comparing outflow and melt to snowfall accumulation (the mass budget method), observing changes in glacier elevation (volume change or geodetic method), and detecting changes in the Earth’s gravity field over the ice sheet (gravimetric method).
The study of ice sheet mass balance underwent two major advances, one during the early 1990s, and again early in the 2000s. At the beginning of the 1990s, scientists were unsure of the sign (positive or negative) of the mass balance of Greenland or Antarctica, and knew only that it could not be changing rapidly relative to the size of the ice sheet. Advances in glacier ice flow mapping using repeat satellite images, and later using interferometric synthetic aperture radar SAR methods, facilitated the mass budget approach, although this still requires an estimate of snow input and a cross-section of the glacier as it flows out from the continent and becomes floating ice. Satellite radar altimetry mapping and change detection, developed in the early to mid-1990s allowed the research community to finally extract reliable quantitative information regarding the overall growth or reduction of the volume of the ice sheets. By 2002, publications were able to report that both large ice sheets were losing mass (Rignot and Thomas 2002). Then in 2003 the launch of two new satellites, ICESat and GRACE, led to vast improvements in one of the methods for mass balance determination, volume change, and introduced the ability to conduct gravimetric measurements of ice sheet mass over time. The gravimetric method helped to resolve remaining questions about how and where the ice sheets were losing mass. With this third method, and with continued evolution of mass budget and geodetic methods it was shown that the ice sheets were in fact losing mass at an accelerating rate by the end of the 2000s (Veliconga 2009, Rignot et al. 2011b).
Case studies of ice sheet mass balance: mass budget method
Remote sensing data measuring the velocity of Jakobshavn Isbræ (a fast-moving glacier on the west coast of Greenland) showed that between 1992 and 2003, the glacier exhibited a large increase in velocity. The data indicated that large glaciers can alter their ice discharge at timescales less than a decade, much faster than previously thought (Joughin et al. 2004).
Researchers used RADARSAT synthetic aperture radar to map ice flow velocity over Greenland for the winters of 2000-2001 and 2005-2006. The data showed an overall pattern of accelerated glacier flow and terminus retreat between 2000 and 2006, although there were some glacier surges and slowdowns. Because the data were acquired at different times of year—September-January for the 2000-2001 measurements, and December-April for the 2005-2006 measurements—seasonal differences may have played a part in the changes observed (Joughin et al. 2010).
InSAR observations from 1992 to 2006 mapped the ice flow for most of the Antarctic coastline, and detected different patterns of ice flux into the ocean in East and West Antarctica. In East Antarctica, small glacier losses led to a near-zero loss of 4 ± 61 gigatons per year. In West Antarctica, more widespread glacier losses increased ice sheet loss by 59 percent over a decade. In 2006, the estimated loss was 132 ± 60 gigatons. Along the Antarctic Peninsula, losses increased by 140 percent, to 60 ± 46 gigatons in 2006 (Rignot et al. 2008).
Rignot and colleagues published a high-resolution digital mosaic of Antarctic ice flow speed in 2011 (Rignot et al. 2011). Based in InSAR measurements acquired between 2007 and 2009 the mosaic was compiled from 900 satellite tracks and more than 3,000 radar data orbits. The map of ice flow speed revealed a complex pattern where fast glacier flow near the coast extended well inland in narrow tributary bands. The next year, Rignot and Mouginot published another comprehensive, high-resolution map of Greenland based in radar interferometry data from 2008 and 2009 showing that Greenland's 100 fastest glaciers drain 66 percent of the ice sheet area, and marine-terminating glaciers drain 88 percent of the ice sheet area (Rignot and Mouginot 2012).
West Antarctica has three major drainage basins where glaciers reach the ocean: the Ross Sea Embayment, the Weddell Sea Embayment, and the Amundsen Sea Embayment. A study of ice discharge from the Amundsen Sea Embayment used ice-velocity measurements derived from Landsat and radar interferometry, and previously documented ice thickness to estimate the total discharge from 1973 to 2013. The study found that ice discharge increased by 77 percent since 1973, half of that occurring from 2003 to 2009 (Mouginot et al. 2014).
Elevation change or geodetic method
Satellite radar altimetry, in which timing of a radar or laser beam return back to a satellite is used as a measure of surface elevation, enabled researchers to assess ice mass by examining elevation change over time.
From 1997 to 2003, volumetric methods showed that average loss of ice in Greenland was 80 ± 12 cubic kilometers per year. This is compared to roughly 60 cubic kilometers per year for 1993 through 1994. About half the increased ice loss was from higher summer melt. The rest of the loss resulted from the velocities of some glaciers outstripping those needed to balance upstream snow accumulation (Krabill et al. 2004). Later research showed Antarctica and Greenland have both lost overall mass at about 120 gigatons of ice per year. The suspected triggers for accelerated ice discharge on both continents include surface warning and melt runoff, ocean warming, and circulation changes. Over the 21st century, the team predicted, ice loss would counteract snowfall gains predicted by some climate models (Shepherd and Wingham 2007). Recently an improved radar altimetry study confirms and extends earlier measurements (Flament and Rémy 2012).
Laser altimetry from ICESat has now supplemented radar altimetry measurements for more detailed volumetric-based studies. In 2009, using ICESat, measurements of both Greenland and Antarctica found that dynamic thinning (ice loss resulting from accelerated glacier flow) now reached all latitudes in Greenland, and had intensified at key areas of Antarctica's grounding line. The study concluded that dynamic thinning lasts for decades after an ice shelf collapse, a situation that occurred several times in the late 1990s and early 2000s. Moreover, the thinning reached far inland. In other regions, warm sub-surface ocean water was shown to be responsible for thinning glaciers as they went afloat, resulting in rapid acceleration of a broad area of the glacier (Pritchard et al. 2009). ICESat data indicated that basal melting was also thinning floating ice shelves, reducing their ability to buttress the glaciers feeding them. Because Antarctica drains more than 80 percent of its ice sheet through floating ice shelves, accelerated glacier flow has the potential to affect ice sheet mass balance dramatically and raise sea level (Pritchard et al. 2012).
One augmentation of the altimetry method is to use satellite stereo images with satellite laser or airborne laser altimetry to map a complex region. This resolves large changes in outlet glaciers residing in narrow fjords, and the smaller changes in the interior ice sheet that drains into the outlet areas (Howat et al. 2007, Shuman et al. 2011, Berthier et al. 2012, Scambos et al. 2014).
The European Space Agency's CryoSat-2 mission has enhanced Antarctic ice sheet monitoring by including areas closer to the poles than earlier satellites, and by acquiring better data in moderately sloping areas, including ice sheet margins where most of the ice loss occurs. CryoSat-2 observations taken between November 2010 and September 2013 indicate annual ice sheet mass losses of 134 ± 27 gigatons in West Antarctica, 3 ± 36 gigatons in East Antarctica, and 23 ± 18 gigatons on the Antarctic Peninsula. The Amundsen Sea showed the largest signal of ice loss (McMillan et al. 2014).
NASA's Gravity Recovery and Climate Experiment (GRACE) has provided glaciologists with a new tool to study mass balance on both Greenland and Antarctica. GRACE measures changes in the strength of the gravitational force over the surface of the Earth, including changes driven by the accumulation or loss of ice.
Between April 2002 and April 2006, GRACE data uncovered ice mass loss in Greenland of 248 ± 36 cubic kilometers per year, an amount equivalent to a global sea rise of 0.5 ± 0.1 millimeters per year. The ice mass loss rate increased by 250 percent between April 2002 to April 2004 and May 2004 to April 2006. The increase was due almost completely to increased ice loss rates in southern Greenland (Velicogna and Wahr 2006a). Between 2003 and 2005, the Greenland Ice Sheet lost 101 ± 16 gigatons per year, with a gain of 54 gigatons per year above 2,000, meters and a loss of 155 gigatons per year at lower elevations. The lower elevations showed a large seasonal cycle: mass losses during summer melting, and mass gains from autumn through spring. The ice mass loss observed in this research was a change from the trend of losing 113 ± 17 gigatons per year during the 1990s, but was smaller than some other recent estimates (Luthcke et al. 2006).
In 2010, a study using GRACE and Global Positioning System (GPS) measurements from three long-term sites on bedrock near the ice sheet found that the ice loss already documented over southern Greenland was spreading along the northwestern coast. The acceleration of loss likely started in late 2005. GRACE data gave a direct measure of mass loss averaged over scales of a few hundred kilometers, and the GPS data observed crustal uplift resulting from ice mass loss. Uplift observed by both sources showed rapid ice acceleration in southeast Greenland in late 2003, and a modest deceleration in 2006 (Khan et al. 2010).
In the Southern Hemisphere, GRACE measurements indicated a significant ice loss in the Antarctic Ice Sheet from 2002 to 2005. Ice sheet mass decreased at 152 ± 80 cubic kilometers of ice per year, equal to 0.4 ± 0.2 millimeters of sea level rise per year. Most of the mass loss came from the West Antarctic Ice Sheet (Velicogna and Wahr 2006b).
A 2012 study (Shepherd et al. 2012) combined satellite altimetry, interferometry, and gravimetry data from the same regions, time spans, and models to examine ice sheet balance. The study found reasonable agreement between the different satellite methods, and arrived at the following best estimates of mass balance changes per year for 1992 through 2011: Greenland: lost 142 ± 49 gigatons; East Antarctica: gained 14 ± 43 gigatons; West Antarctica: lost 65 ± 26 gigatons; Antarctic Peninsula: lost 20 ± 14 gigatons. The study also found that, since 1992, polar ice sheets contributed to sea level rise by an average of 0.59 ± 0.20 millimeters per year—a total of 11 millimeters since 1992. (A 2014 study by McMillan et al. examining CryoSat-2 data more than doubled the estimated rate of Antarctic ice sheet contribution to sea level. Shepherd et al. estimated the annual contribution rate at 0.19 ± 0.15 millimeters over a 20-year period; McMillan et al. estimated the rate at 0.45 ± 0.14 millimeters per year between 2010 and 2013.)
Surface Melt and Ocean Impacts
Surface melt on an ice sheet not only directly reduces the ice sheet mass but also can accelerate ice flow and even leads to further melting. Surface meltwater can penetrate through cracks in the surface, and force them open, allowing large amounts of water to drain to the bed and spread out across the base of the ice sheet, lubricating it (Zwally et al. 2002). This effect led to initial concern about rapid acceleration resulting directly from warmer air over the ice sheet.
Recent studies have found, however, that there is a limit to the effect of surface meltwater penetration. The relative speedup of outlet glaciers is small in most years, less than 15 percent (Joughin et al. 2008). Satellite observations of southwestern Greenland in the 1990s documented ice flow development, showing how it changed in years of differing melt rates. In the first half of the summer, the flow rates were similar in all years, but flow rates differed in the second half. Surprisingly, the flow rate was 62 ± 16 percent, lower in warmer years, and the period of fast ice flow lasted only a third as long. The data suggested that, like mountain glaciers, melt-induced glacier acceleration actually stops in years of intense melting once subglacial water erodes through the sediments and creates channels for water flow (Sundal et al. 2011).
In February and March of 2002, the Larsen B Ice Shelf on the Antarctic Peninsula underwent rapid disintegration (Scambos et al. 2003). Warm summertime temperatures led to the formation of melt ponds on the ice surface. Some of this meltwater infiltrated cracks in the ice, slicing through the shelf. The increased amount of fracturing, and possibly changes at the ice shelf margins (loss of connection with the coastline) and wave action (flexing the shelf a slight amount), led to the break-up of the shelf. Similar events have occurred before (Larsen A Ice Shelf in 1995, Larsen Inlet Ice Shelf in 1986 or 1987) and since (northwestern Wilkins Ice Shelf in 2008), but so far these events are limited to the Antarctic Peninsula. The Peninsula is the fastest-warming part of the continent.
Ice shelves generally act like brakes on the glaciers upstream, but once an ice shelf disintegrates, the glaciers can accelerate. In the year and a half following the 2002 Larsen B disintegration, the glaciers feeding the Larsen Ice Shelf accelerated substantially some of them moving several times their previous speed (Scambos et al. 2004, Rignot et al. 2004). Glacier flow speeds subsequently dropped, but remained quite high compared to what they had been prior to the ice shelf breakup. While the Larsen B glacier system is of only moderate size (about 10 gigatons per year) the process is an example of what might occur on a larger scale. In this region, ocean effects are thought to be minor, because little warm ocean water reaches the Larsen ice front.
Similar patterns of ice calving and retreat leading to rapid glacier acceleration have also been observed in Greenland. In southeast Greenland for example, two large glaciers named Helheim and Kangerdlugssuaq lost parts of their floating ice tongues in 2003, and underwent a rapid acceleration to approximately three times their earlier speed. (Howat et al. 2007). These glaciers have gradually slowed in the following years, but calving and mass loss from other glaciers on the southeastern Greenland coast and the western coast continues.
The impacts of ice shelf collapse and ensuing glacier acceleration are substantial, but in general, the effects of ocean melt are proving to be far more important in controlling ice sheet mass balance. Warm ocean water plays a significant role in melting glacial ice from below, and a better mapping of Antarctica’s and Greenland’s landforms beneath the ice suggests that ocean melting of the glacier fronts may play a more significant role than previously thought as the ice sheets retreat (under a global warming scenario).
When a glacier fills a coastal valley, the elevation of the valley floor relative to sea level is significant. As a deep-keeled glacier retreats, seawater extends inland into the emerging fjord, and can continue to melt the remaining ice at the retreating glacier front. Where the valley floor rises above sea level, seawater cannot reach much of the remaining ice, and the pace of ice loss may slow (since only surface melt processes are in play at that point). A study of Greenland topography has found widespread, deeply incised glacial valleys with elevations well below sea level extending much farther inland than previously thought (Morlighem et al. 2014). In Antarctica, a new compilation called Bedmap2, produced by the British Antarctic Survey, merges multiple data sources to map the seafloor and sub-glacial bedrock elevation.
Multiple studies of Antarctica indicate growing ice sheet instability, especially in the Amundsen Sea Embayment, where the Thwaites, Pine Island, Smith, Kohler, Pope, and Haynes Glaciers drain the central West Antarctic Ice Sheet. A study using Earth Remote Sensing satellite radar interferometry (EERS-1 and -2) observations from 1992 through 2011 finds "a continuous and rapid retreat of the grounding lines of Pine Island, Thwaites, Haynes, Smith, and Kohler" Glaciers, and the authors conclude that "this sector of West Antarctica is undergoing a marine ice sheet instability that will significantly contribute to sea level rise in decades to centuries to come" (Rignot et al. 2014). Bedrock mapping combined with a numerical model shows that early-stage ice sheet collapse is potentially underway in the Thwaites Glacier Basin, largely driven by subshelf melt. The model forecasts that rapid collapse could occur within 200 to 900 years (Joughin et al. 2014).
The Thwaites Glacier "drains the so-called weak underbelly of the West Antarctic Ice Sheet" (Alley et al. 2015) and this glacier was identified in 1981 as the most likely conduit for collapse of the ice sheet. Multiple studies (Alley et al. 2015, Pollard et al. 2015, Feldmann and Levermann 2015) highlighted the vulnerability of the West Antarctic Ice Sheet to collapse, indicated that the collapse could happen in a matter of decades once it began, and suggested that the threshold or trigger point for the collapse might have already been passed, though rapid changes might not occur for centuries.
Last updated: 9 November 2015