On the up
GPS antenna (right) at a site in the Antarctic Peninsula and the power system (left) required to run it through the winter
9 November 2012 by Matt King and Pippa Whitehouse
Antarctica is still on the rebound after shedding huge volumes of ice. Matt King and Pippa Whitehouse explain how they're tracking its slow movement, and why.
If you've ever walked a long way with a heavy rucksack on your back, and then felt the enormous relief as that load was lifted off, you know how Antarctica feels.
20,000 years ago, during the last ice age, Antarctica was weighed down by a much larger ice sheet, but since then it has generally been losing mass. Vast tracts of ice, hundreds of metres thick, have flowed into the ocean - and this has lifted a great weight from its surface. The Earth beneath responds to these changes; it is now rising, just as it once sank when the ice sheet grew.
This process, known as glacial isostatic adjustment (GIA), doesn't happen instantaneously - the land can take thousands of years to fully rebound, and it's still moving today. This presents both an opportunity and an obstacle to understanding the history of the Antarctic ice sheet.
The opportunity is that measuring this ongoing response can tell us how the thickness and extent of the ice sheet have changed over time in response to past climatic changes. This is important, since it sets the context for present-day changes, which will have major implications for how fast sea levels rise over the next century.
The crust's vertical motion due to GIA is typically only a few millimetres a year, so to measure it we need a very accurate and precise technique. Global positioning system (GPS) receivers attached to rock outcrops poking through the ice sheet are the answer.
Solar panels get shattered by flying rocks, wind turbines get ripped off their bearings and cables slowly vibrated off their threads.
The information these provide is important because conventional methods of determining past ice extent, such as looking for glacial trim lines (bathtub rings etched in the landscape by the ice), or dating boulders left behind as the ice retreated, give limited data for most regions and periods.
This dearth of information has led scientists to propose several very different reconstructions of Antarctica's ice history. If we can relate measurements of present-day crustal uplift to the history of the ice, we can test these competing ideas.
To link ice history to present-day crustal uplift we need to know how the solid Earth responds to ice being added to, and then removed from, its surface. The characteristics that matter for GIA relate to the strength of the top 100km or so of the solid Earth and the viscosity of the mantle beneath.
These control both when and where uplift happens. Unfortunately they are not well defined - not only is the ice-loading history poorly known, but the Earth's response to that loading is uncertain.
To make matters worse, there are trade-offs between these two unknowns - a lack of uplift today could be because the ice didn't change much in the past, or because the ice changed a lot but the solid Earth adjusted quickly and has already finished rebounding. The picture is complex, but every new measurement brings it into sharper focus.
Melting ice, rising rock
The obstacle presented by continuing GIA referred to earlier, is that it obscures the changes in gravity caused by present-day changes in ice mass.
Since 2002, the Gravity Recovery & Climate Experiment (GRACE) satellites have been mapping Earth's gravity field every 30 days, letting us track mass as it moves on, or below, Earth's surface - the more mass there is in an area, the stronger its gravitational pull on the satellites.
The GPS antenna will uplift as the solid Earth responds to glacier thinning in the past and present
But GRACE can't tell the difference between ice mass, atmospheric mass, ocean mass or the mass of the solid Earth. So to calculate changes in ice mass accurately, we need to account for these other factors. And this is where models of GIA become critical - it turns out that the average rate of mass change measured by GRACE over Antarctica since 2002 is close to zero. So the overall amount of mass is hardly changing.
To GRACE, GIA uplift looks like an increase in mass, and this is cancelling out the fall in mass due to ice melt. If GIA is small then the loss of mass from melting ice must also be small. But if GIA's contribution is large, then the ice sheet must be losing a similarly large amount of mass. Our uncertainty about the ice sheets' history and the Earth's response led to an uncertain GIA model, and this may mean that previous estimates of ice-mass change derived from GRACE data are wrong. We need to know how wrong.
And that brings us back to GPS measurements of crustal uplift, which have been attempted in Antarctica since the mid-1990s. The challenges in collecting these data are numerous. The first is the lack of rock outcrops - less than 2 per cent of the continent is permanently ice-free. Getting to a single outcrop requires flying very long distances from the Antarctic bases - sometimes more than 1000 miles.
The second challenge is providing a steady 3 watts of power to the GPS receivers. This is a fraction of the amount needed by a typical light bulb, and in summer solar power generates it easily, but winter presents serious difficulties.
The third challenge is the climate - particularly the merciless wind. Solar panels get shattered by flying rocks, wind turbines ripped off their bearings and cables slowly vibrated off their threads.
However, with recent equipment advances, Antarctic GPS receivers are now running almost continuously. And the uplift rates we are observing are a revelation.
Pippa Whitehouse in the field
Comparison of GPS velocities with predictions from GIA models has revealed major differences. The models predict uplift of more than 15mm a year in West Antarctica; GPS velocities aren't even half this. Recent evidence from glacial geology suggests that existing ice-history models predict too much ice loss in the past; unless the Earth-response models used in the GIA predictions are hugely wrong, the low GPS uplift measurements appear to add weight to this argument.
To address these problems we have produced a new model of how the Antarctic ice sheet retreated since the last ice age, reconstructing its extent every 5000 years, from 20,000 years ago up to the present, by trying to find the closest possible match to the sparse data.
The Earth's properties are tuned so that predicted relative sea-level changes match the available observations at 15 sites around the coast. Our uplift predictions define our new best estimate of present-day GIA, and our model fits the critical GPS data more than twice as closely as the previous best efforts.
It is an important advance and will affect estimates of ice-mass change from GRACE dramatically, moving us even closer to the truth of how much Antarctica is contributing to present-day sea-level change.
Yet our new model is not perfect. GPS rates suggest we still overpredict the crust's rebound in parts of West Antarctica. Sensitivity tests show that understanding ice-sheet changes in the past few thousand years - even the past few hundred years - may be critical to modelling many regions accurately. More data and perhaps new methods are needed before we can truly validate GIA models.
This is especially true in the vast expanse of East Antarctica, where we still aren't even sure whether the crust is rising or sinking beneath the mysterious 4km-thick ice. It's going to be some time before this burden is lifted off the backs of those of us studying GIA.
Dr Matt King is a reader in polar geodesy at Newcastle University where he applies geodetic techniques, such as GPS and GRACE, to problems relating to the polar ice sheets.
Dr Pippa Whitehouse is a postdoctoral research associate at Durham University where she uses numerical models to develop our understanding of past changes in the ice sheets and sea level.