Antarctic ice – more accurate estimates

Last week’s Science magazine had a News Focus article on estimates of ice loss in Antarctica. It quietly discussed a paper published in May by two NASA scientists:

H. Jay Zwally & Mario B. Giovinetto (2011) Overview and Assessment of Antarctic Ice-Sheet Mass Balance Estimates: 1992– 2009. Surv Geophys DOI 10.1007/s10712-011-9123-5  (note this is Open Access)

Estimates of Antarctic ice net variation vary widely.  This is in part due to the different methods used, but the magnitude of the change might surprise you.

“Mass balance estimates for the Antarctic Ice Sheet (AIS) in the 2007 report by the Intergovernmental Panel on Climate Change and in more recent reports lie between approximately +50 to -250 Gt/year for 1992 to 2009. The 300 Gt/year range is approximately 15% of the annual mass input and 0.8 mm/year Sea Level Equivalent (SLE).”

The paper set out to investigate the various estimates, assessing previously published results that used three standard methods which the Science article conveniently summarises:

“Each of the three methods has its foibles.

• In the first, altimetry, researchers bounce laser or radar signals off the ice to measure its height and thus its volume. The method has been used to survey most of the continent, but converting changes in volume to changes in mass raises major uncertainties.
• The second technique, gravity, employs the two satellites of the Gravity Recovery and Climate Experiment flying over the ice in tight formation to measure mass directly. But that record goes back to only 2002, and data analysis is tricky.
• Finally, the input-minus-output method (IOM) works by subtracting ice flow into the sea from total snowfall. Both numbers are huge, however, and the mass of snow falling on East Antarctica is especially hard to gauge.”

Zwally and Giovinetto’s reassessment also included a challenge to some assumptions, substituting field measurements and making ‘preferred estimates’. These took account of the uncertainties inherent in the various techniques. Their reanalysis provides much lower estimates of net change in ice, ranging from +27 to -40 billion tons per year.  For 1992 – 2001 they are prepared to go even further, estimating a loss of only 31 billion tons per year.  These still sound like huge numbers, but to put it in perspective, 2400 billion tons of snow falls in Antarctica each year, so we’re dealing with a gain or loss in the range +1.1 to -1.7%.

The same techniques applied by the authors in a previous paper (Zwally HJ et al (2011) Greenland ice sheet mass balance: distribution of increased mass loss with climate warming. J Glaciol 57(201):88–102) brought a significant convergence to estimates produced by ICESat altimetry and the GRACE gravity signal in Greenland, however, while the Greenland ice sheet continues to grow inland and thin at the margins, overall recently it has been losing mass.

What I find most refreshing is the revision and quantification of uncertainty. What is shocking is not the magnitude of the possible loss, but the short timescale on which the estimates are based and the lack of knowledge of historical data.

“The new analysis is a “perfectly reasonable reinterpretation,” glaciologist Andrew Shepherd of the University of Leeds, U.K., says. “The paper’s main contribution is a very convincing argument that one needs to account for uncertainties in a consistent way.” “

The Science article ends by hinting that this may make no more than tiny ripples in the consensus.

“Getting more than a feeling for what Antarctic ice is doing to sea level will take more than one group’s reassessment of the published literature, researchers agree. The Intergovernmental Panel on Climate Change (IPCC) is already working on an ice balance assessment for its report due in mid-September 2013, but researchers say more must be done to focus scientists’ attention on the problem.”

Does that sound like a dismissal? It does to me.