by Rasmus Benestad, Eric Steig and Gavin Schmidt
In a recent paper in Science, Eric Rignot and Pannir Kanagaratnam present new satellite observations of the speed of glaciers of Greenland, and find that they are sliding towards the sea almost twice as fast as previously thought. Additionally, between 1996 and 2005, they detected a widespread glacier acceleration and consequently an increased rate of ice discharge from the Greenland ice sheet. However, previous papers have recently noted an increase in snow accumulation in the interior (i.e. Johannessen et al., 2005), so how do these different measurements fit into the larger picture of Greenland’s net mass balance?
The measurements by Rignot and Kanagartnam were made with interferometers which measure the movement of the surface horizontally, and so is complimentary to the altimeter data published previously (which measures the absolute height of the ice). Overall, they found widespread increases in glacier speeds, and increases of about 30% in ice discharge rates. (Note that the satellite image shows that the glaciers in the east tend to slide far into the sea whereas on the western coast that happens less).
The higher velocity of the ice is thought to be related to higher temperatures causing increased melt-water which can penetrate to the base of the glacier and hence reduce the ground friction. However, this accelerated movement is not necessarily tied to an increased rate of melting of the Greenland ice, although it can be related. Surges of ice streams from the ice sheet can also occur due to increased accumulation at the head of the glacier. However, when the increased ice velocity is matched to a decreasing thickness that can be sign of net mass loss. These ideas are consistent with observations of surface melting which had a record extent in 2005, and has been increasing steadily (though with significant interannual variability) since 1993. Using the analysis of Hanna et al (2005) (based on the reanalysis datasets) for the surface mass balance, Rignot & Kanagartnam estimate that Greenland is on balance losing mass, and over the period of their study the ice sheet mass deficit (the amount of ice lost to the sea) has doubled increasing from 90 to 220 km3/year (an increase of 0.23 to 0.57 mm/yr sea level equivalent – SLE).
In the earlier Science paper, Johanessen et al. found increased snow accumulation on the top of the interior Greenland ice sheet between 1992 and 2003. Above 1500m a.s.l in much of the interior Greenland they estimated an increase of 6.4 ± 0.2 cm/year and below 1500m they observed a decreasing trend of -2.0 ± 0.9 cm/year. Hence, growth in the interior parts and a thinning of the ice nearer the edges. However, Johanessen et al. were not able to measure all of the coastal ranges. Indeed, the thinning of the margins and growth in the interior Greenland is an expected response to increased temperatures and more precipitation in a warmer climate. These results present no contradiction to the accelerated sliding near the coasts, but both will affect the ice/snow (fresh water) mass estimate. Whereas the finding of Rignot and Kanagaratnam suggests a larger sink of the frozen Greenland fresh water budget (the ice is dumped into the sea), the snow deposition in Greenland interiors is a source term (increases the amount of frozen fresh water). It does not matter for the general sea level in which form the water exists (liguid or solid/frozen) when it is discharged into the sea: The same mass of liquid water and immersed ice affect the water level equally (Archimede’s principle).
A third relevant study is a recent paper in the Journal of Glaciology by Zwally et al. (2005) on the ice mass changes on Greenland and Antarctica. They use the same satellite obsevations (ERS 1 and 2) as Johanessen et al. and again find that the Greenland ice sheet is thinning at the margins (-42 ± 2 Gt/year = -46 ± 2 km3/year below the equilibrium-line altitude – ELA), but growing in the inland (+53 ± 2 Gt/year = 58 ± 2 km3/year). The mass estimates have been converted to volume estimates here, assuming the density of ice is 0.917 g/cm3 at 0°C, so that the mass of one Gt of ice is roughly equivalent to 1.1km3 ice*. This means that the Greenland ice has an overall mass gain by +11 ± 3 Gt/year (=10 ± 2.7 km3/year) which they estimated implied a -0.03 mm/year SLE over the period 1992-2002.
The critical point for Greenland is whether the increased rate of glacier motion more than compensates for the greater accumulation on the surface. While the broad picture of what is happening is consistent between these papers, the bottom-line value for Greenland’s mass balance is different in all three cases. Looking just at the dynamical changes observed by Rignot & Kanagaratnam, there is an increased discharge of about 0.28 mm/year SLE from 1996 to 2005, well outside the range of error bars. This is substantially more than the opposing changes in accumulation estimated by Johannessen et al and Zwally et al, and is unlikely to have been included in their assessments. Thus, the probability is that Greenland has been losing ice in the last decade. We should be careful to point out though that this is only for one decade, and doesn’t prove anything about the longer term. As many of the studies make clear, there is a significant degree of interannual variability (related to the North Atlantic Oscillation, or the response to the cooling associated with Mt. Pinatubo) such that discerning longer term trends is hard.
The largest contributions to sea level rise so far are estimated to have come from thermal expansion, with the melting of mountain glaciers and icecaps being of second order. Looking forward, the current (small) imbalance (whether positive or negative) of the Greenland ice sheet is not terribly important. What matters is if the melting were to increase significantly. Ongoing observations (most promisingly from the GRACE gravity measurements, Velicogna et al, 2005) will be useful in monitoring trends, but in order to have reasonable projections into the future, we would like to be able to rely on ice sheet models. Unfortunately, the physics of basal lubrication and the importance of ice dynamics highlighted in the Rignot & Kanagaratnam results are very poorly understood and not fully accounted for in current ice sheet models. Until those models include these effects, there is a danger that we may be under-appreciating the dynamic nature of the ice sheets.
Hanna, E; Huybrechts, P; Janssens, I; Cappelen, J; Steffen, K; Stephens, A (2005) J. Geophys. Res.Vo. 110, D13108, doi:10.1029/2004JD005641
Johanessen, O.M; Khvorostovsky, K; Miles, M.W; Bobylev, L.P. (2005) ScienceVo. 310 no. 5750, pp 1013-1016
Ringnot, E; Kanagaratnam, P (2006) ScienceVo. 311 no. 5763, pp 986-990
Velicogna, I; Wahr, J; Hanna, E; Huybrechts, P. (2005) Geophys. Res. Lett.Vo. 32, L05501, doi:10.1029/2004GL021948
Zwally, H. Jay; Giovinetto, Mario B.; Li, Jun; Cornejo, Helen G.; Beckley, Matthew A.; Brenner, Anita C.; Saba, Jack L.; Yi, Donghui (2005), Journal of Glaciology, Volume 51, Number 175, December, pp. 509-527(19)
*Update: Correction for arithmetic error in orginal post in converting Gt to km3, see comments.