Over geologic time, sea level depends on the relative balance of the water stored in ice versus that stored in the oceans. Most of the Earth’s ice is stored near its poles where global warming has had the greatest effect thus far, and is projected to be most severe in the future. Thus predictions of sea level rise largely depend on what happens in Greenland and Antarctica over the next century, but the physics of ice sheet melting are poorly understood. The result is that the most reliable projections of sea level rise use linear extrapolations of ice sheet melting, if this term is incorporated at all. (image credit: JA Dowdeswell)
Late last week we found out that those linear extrapolations may be significant underestimates. Today’s Published Research Synopsis focuses on a paper in Science that provides comprehensive estimates of the rates of melting across Greenland over the last decade. Using combinations of satellite and aerially-sensed data, the velocities and depths of glaciers draining the bulk of Greenland were measured in 1996, 2000, and 2005. The researchers found that between 1996 and 2005, the rate of ice sheet melting has more than doubled across Greenland.
Citation
Rignot, E. and Kanagaratnam, P. (2006), “Changes in the Velocity Structure of the Greenland Ice Sheet”, Science 311(5763), pp. 986-990 [online at CiteULike.org]
Synopsis
In order to determine the rate of mass loss of ice across Greenland over the last decade, Rignot and Kangaratnam took a slightly different approach than other groups. More conventionally, hundreds of flights’ worth of radar-elevation data were combined to produce detailed topographic maps across large sections of the ice sheet. However, as the authors point out, the coverage using this technique has been incomplete, particularly across the southeast and northwest. Instead, the authors attempt to estimate the rate that ice flows out to sea at the outlets of the islands’ glaciers. The figure on the left shows how the ice sheet can be divided into ice drainage basins, each of which is drained by a single fast-moving glacier. To measure the rate at which Greenland is losing ice, then, the velocity of each of these outlet glaciers, and their depth must be determined.
To measure the velocity, the authors use a technique known as satellite radar interferometry in which radar measurements taken several weeks apart are compared to reveal the velocity of surface features. They estimate that their uncertainties in velocity are about 10 to 30 meters/year, so only relatively quickly moving areas of the ice sheet work with this method. This is why the colored velocity indicator in the figure on the left only applies near the edges of the ice sheet. The interior areas are indeed moving, just too slowly for the spacecraft to measure. What the figure does show is that velocities range from as little as 10 meters/year to several kilometers/year.
| Discharge | Mass Balance | Sea Level Change | |
| Year | km3/yr | km3/yr | mm/yr |
| 1996 | 56 | -91 | 0.23 |
| 2000 | 92 | -138 | 0.35 |
| 2005 | 167 | -224 | 0.57 |
The depth of glaciers is determined by radio echos from airborne surveys. Combining these measurements with the surface velocities and widths measured from space, the discharge rate of ice into the ocean at the outlet of each glacier can be estimated. The table on the right lists the authors’ estimates of total discharge of ice into the ocean each year.
But that is not the whole story, after all, the balance of ice snow accumulation, glacier melting, and iceberg calving is what ultimately determines sea level rise. The authors combine their rates of iceberg calving with rates of snow accumulation (which have been increasing somewhat) and ice melting. This result, the third column in the table above is what determines sea level rise, and it has been increasing dramatically since 1996. The effect this has had on sea level is shown in the fourth column.
Context
Even though Greenland seems to be melting and losing ice more quickly than before, the net increase in sea level has been small. However, glaciologists have discovered that increases in glacial meltwater can lubricate the bases of the glaciers and thus speed the calving of the glaciers into the ocean. Then, as the outlet glacier thins dramatically, the entire ice sheet behind it begins to flow more quickly toward the sea. The dynamics and stability of the entire system are very much an open area of research, but it appears that the process of ice sheet thinning suffers from positive feedback. In this case though, positive feedback is very bad for us because it means that the melting and thinning that occurs each year serves to increase that of the following year without any natural stopping mechanism.
Similar processes are, of course, occurring in Antarctica where most of the world’s ice is stored. However, the size and relative inaccessibility of the continent make estimates of rates of ice accumulation much more uncertain than those from Greenland. In Antarctica however, massive floating ice shelves appear to hold back the ice sheets behind them. We have witnessed rapid deterioration and breakup of many of those ice sheets over the last decade. What follows is then a rapid acceleration of the outlet glaciers behind; again, a “positive” feedback mechanism.
These observations, and others, are part of what climate scientists lump together in their discussion of a “tipping point” in temperature rise on the Earth. Above this point, say 2 degrees celsius, there will be no stopping what is known as catastrophic global warming. The exact point is not known with any real certainty, and to what degree the warming would be catastrophic is not known that well either. But what is certain is that the Earth is warming up, and Greenland is melting more quickly as it does.
General Explanations
Anatomy of a Frozen “Continent”
Greenland is nearly entirely covered by one, continuous sheet of ice that is several kilometers thick in most locations. This sheet of ice has taken many hundreds of thousands of years to accumulate. At the base of all of that ice, the pressures are enormous, causing ice to behave very differently than it does at the surface. Have you ever taken an ice cube and slowly crushed it between your molars? If so, you may have noticed that at a certain point it does not fracture but rather it slowly flattens. At high pressures, ice is a visco-plastic material that is able to flow very much like molasses; very cold molasses. This flowing ice is what defines a glacier, if the ice doesn’t flow it’s just a chunk of snow.
So Greenland’s many kilometers of ice are slowly flowing away from the center of the island where the ice is the thickest, and thus pressure the highest. In some areas, due to topography beneath the ice as well as variations in snowfall across the sheet of ice, the ice flows much more quickly. These rivers of ice, which I’ve called outlet glaciers above, are responsible for most of the ice leaving the ice sheet itself. When the outlet glaciers meet the ocean, they float out until ocean currents and melting are sufficient to break off the nose of the glacier as an iceberg, a process called calving. On Greenland, most calving occurs at the mouths of fjords, whereas in Antarctica, calving often occurs at the edges of floating ice shelves. Sea level rises only when an outlet glacier meets the ocean, not when an iceberg calves off.

I can’t imagine anything more boring than watching ice melt professionally!
Tom,
If there’s a lot of ice, and maybe a volcano thrown in for good measure it could be pretty exciting. Huge floods, massive cave-ins and so forth.
Ah. Well that sounds better than what I am doing!
Oh and note to other readers: it’s ANTHONARES’ BIRTHDAY TODAY. Happy birthday.
Thanks Tom! It’s been a pretty sweet birthday so far. Oh wait, no it’s been a pretty normal day. But hey, my car insurance just went down! That’s some birthday goodness I can take to the bank.
Hello, I found your site through blogger, and I love how you cited your sources. What do you think about Johannessen et al’s paper, here?
mac_davis,
Just looking at the abstract of Johannessen’s paper, it seems to be in general agreement with what Rignot and Kanagaratnam reported in their paper: snow and ice accumulation is increasing in the deep interior because of increased precipitation, while the ice sheet is thinning along its boundaries due to increased melting and ice calving.
Rignot and Kanagaratnam’s paper speculated that snow and ice over other parts of Greenland are melting and the water is flowing into the ocean. Citing other work by Hanna et al. (2004), Rignot and Kanagaratnam figured another 35 km^3 in 1996 and 57 km^3 in 2005 of ice loss occurred from surface melting bringing the total annual loss volume to 91 km^3 in 1996 and 224 km^3 in 2005.
This water and ice input leads to a sea level rise of 0.23 ± 0.08 mm/year in 1996 growing to 0.57±0.1mm/year by 2005. Not surprisingly, the reason that is given—presumed—for the melting ice and the rising seas is that temperatures are going up because of global warming.
This paper with no reference to Johannessen’s paper showing that Greenland is accumulating ice at a rate of about 5.4±0.2cm/year. Johannessen even used data from some of the same satellites. What’s more, Johannessen used real data and Hanna et al., cited by Rignot, used a model of surface melt.
Consider what would have happened had the latest study included the ice and snow gains observed by Johannessen (and ignored the losses modeled by Hanna et. al.). Johannnessen’s increase of 5.4 cm/year averaged over Greenland converts to about 75 km^3/year. Rignot and Kanagaratnam could have subtracted Johannessen’s gains. If they had done so, the total volume of ice loss from Greenland would only have become positive during the last 5 years, totaling 17 km^3 in 2000 and 92 km^3 in 2005. This translates to a sea level rise contribution of 0.04mm in 2000 and 0.23mm in 2005—values much less dramatic than those they published.
Mac_davis,
Thanks! I’ll be keeping an eye on this further, but it’s good to hear critical comments on published work based on science rather than supposition. I wonder why the referees did not make sure that the Johannessen results were accounted for. Perhaps this had something to do with the relative recency of the Johannessen paper?
[…] Greenland is melting. Faster. In now hosts some of the fastest moving glaciers in the world and looks to perhaps play a larger role in sea level rise than previously thought. […]
Today, RealClimate posted an article that directly mentions the disagreement that mac_davis highlighted between Rignot’s paper and that of Johannessen (2005). Basically, the results are not mutually exclusive and Real Climate suggests that Johannessen’s estimates are perhaps overestimates because of their inability to measure snow accumulation over certain parts of the ice sheet. Anyway, check out the entry.
[…] Interferometry (not for the faint of heart!) Since I’ve been talking about interferometry for the last few weeks (here and here) at Anthonares.net, I thought I would discuss the basic concept very briefly. Waves (like light waves, sounds waves, or ocean waves) travel right through each other, but when they are in the same location at the same time, they interfere. Interference results in the addition of the amplitude of the two waves. So, if two waves are in phase, they interfere constructively and the combined result has a greater amplitude than either wave did to start with. If they are out of phase, meaning the peak of one is aligned with the trough of another, then they interfere destructively and the result has a smaller amplitude than both did originally. […]
Very good reading. Peace until next time.
WaltDe
[…] Because of the enormous difficulty in measuring the mass of a continental ice sheet, Greenland’s ice mass budget has always been calculated indirectly. Some of the more exciting methods have involved using satellite radar interferometry (discussed here in this previous PRS) to measure the speed of ice sheet flow, but this requires several assumptions that can introduce significant errors. Other methods rely on spotty measurements of ice sheet thickness, snow accumulation, and meltwater discharge that also require lots of assumptions and extrapolation. […]