I really like the fact that there is still so much to discover about important parts of the climate system. The Bell et al paper in Science Express this week (final version in Science) reporting on the surprising results from airborne ground-penetrating radar studies of the Antarctic Ice Sheet is a great example. The ice sheets themselves are the biggest challenge for climate modelling since we don’t have direct evidence of the many of the key processes that occur at the ice sheet base (for obvious reasons), nor even of what the topography or conditions are at the base itself. And of course, the future fate of the ice sheets and how they will dynamically respond to climate warming is hugely important for projections of sea level rise and polar hydrology. The fact that ice sheets will respond to warming is not in doubt (note the 4-6 m sea level rise during the last interglacial), but the speed at which that might happen is highly uncertain, though the other story this week shows it is ongoing.
The radar results (shown on the right) picked up on some really weird looking features that look to be related to pressure-related freezing of basal meltwater as it is pushed uphill by the weight of the ice sheet above. If that sounds odd, it is because it is.
How can water flow uphill in the first place? This is a function of the pressure gradients. If there is a lot of ice above a valley, but it tapers off towards a mountain range, the pressure on any liquid water at the bottom of the valley will be greater than the pressure up the side of the mountain. This will force water uphill. Incidentally, there are many sub-glacial geomorphological features that show this effect in places affected by the LGM ice sheets.
The freezing point of water is also pressure dependent. With 3km of ice pressure above it, water freezes at about -2ºC (the change is -7.5*10-8 ºC/Pa). Water below 0ºC can therefore exist at the base of the ice sheet (and can also be seen emerging from under ice shelves). When the pressure forces this water upwards to lower pressure areas, this can promote instant freezing, and this seems to be the explanation for the structures seen in the radar.
The surprise is how large these structures are, one shown in the paper is 10’s of km long and 100s of meters thick – certainly large enough to be important in ice flow locally, though probably not at the continental scale.
However, at the continental scale, there is a new assessment of the net mass balance of Antarctica and Greenland. Rignot et al have updated results, including those from the GRACE gravity measurement satellite, to the end of 2010 and show that the downward trend in ice mass is continuing (stronger in Greenland than in Antarctica). The net rise in sea level associated with this decline is about 1.3 mm/yr, which will likely accelerate with further warming. Complementary analyses of the surface mass balance of Greenland (Tedesco et al, 2011) also show that 2010 was a record year for melt area extent.
This rate of melting is more than was figured into the tabulated IPCC AR4 estimates of sea level rise, and any further acceleration will obviously make the discrepancy worse. Indeed, even in the highest forcing A1F1 scenario, the IPCC calculated only a 0.3 mm/year contribution from the ice sheets averaged over the whole 21st Century! This was clearly a gross underestimate.
Extrapolating these melt rates forward to 2050, “the cumulative loss could raise sea level by 15 cm by 2050” for a total of 32 cm (adding in 8 cm from glacial ice caps and 9 cm from thermal expansion) – a number very close to the best estimate of Vermeer & Rahmstorf (2009), derived by linking the observed rate of sea level rise to the observed warming.
There is certainly more to learn about ice sheets, and more of a reason than ever to do that as fast as possible.