Lots of press has been devoted to four papers in this week’s Science, on the topic of ice sheets and sea level.
We’ve already discussed the new evidence that Greenland’s glaciers are speeding up. What is new this week is an effort to evaluate the impact of future warming on Greenland by looking at what happened to it last time it got very warm — namely during the Last InterGlacial (LIG) period, about 125,000 years ago. The same group of authors looked at this in two ways, using NCAR’s Community Climate System model (CCSM) coupled to a state-of-the-art 3-D ice sheet model.
First, in a paper by Otto-Bliesner et al. they ran simulations for the Last Interglacial, and took a look at what happened to the ice sheets. They find that most of the icefields in Arctic Canada and Iceland disappear, and that the Greenland ice sheet is reduced to a steep ice dome in central and northern Greenland. These results are in very good agreement with the available ice core and other paleoclimate data evidence, which indeed show that the Canadian ice sheets disappeared during the LIG, and strongly suggest that much of southern Greenland was deglaciated.
Second, in a paper by Overpeck et al., they examine the implications for past and future sea level rise. The results show that the Greenland and other Arctic ice sheets probably did not contribute more than 3.4 m to the LIG sea level rise. However, data from coral reefs exposed above sea level today, and other evidence, point to an LIG sea level at least 4 m and possibly as much as 6 m greater than today. This suggests that the balance came from the Antarctic ice sheet. This is turn implies a strong sensitivity of the Antarctic ice sheet to sea level rise and climate warming — an idea that goes back to John Mercer (1976) but that had until recently fallen out of favor in much of the glaciology community.
Projecting forward in time, the implication is that our future will also see 4-6 m of sea level rise, and that — given the recent evidence for accelerated flow of both Greenland and Antarctic glaciers — this may occur much faster than we expect. In the model simulations, Greenland may already be warmer in 2100 than it was at the height of the LIG. The rate of sea level rise associated with the warming into the last interglacial was probably greater than 10 mm/yr* while current sea level rise is roughly 3 mm/yr. To the extent that the LIG is a good analog for our future, sea level rise is therefore rather likely to accelerate.
Also in this week’s Science are two articles that further strengthen the case that ice sheets are quite sensitive to warming climate. A paper by Göran Ekström et al. shows that the increased speed of Greenland glaciers occurs in distinct lurches (observed as micro “ice-quakes”) that are strongly seasonal, with the greatest number occuring in late summer. This provides evidence that meltwater plays an important role in the acceleration of Greenland’s glaciers. Essentially, the idea is that surface melting that occurs in the summer can make its way quickly down to the glacier bed, lubricating the bed and allowing the glaciers to slide more rapidly. The “ice quakes” occur because the rough bedrock surface causes the glaciers to stick; they only accelerate when enough hydraulic pressure has built up to help float the glacier over the bumps. This is strong evidence that climate, not merely “internal ice sheet dynamics”, has contributed to the recent increases in Greenland’s glaciers. Indeed, a doubling of the rate of quakes has occurred over the past five years, just as the aerial extent of surface melting has increased.
Finally, in a very nice bit of work Velicogna and Wahr use data from the “Gravity Recovery and Climate Experiment” (GRACE) satellites to show that the Antarctic ice sheet has been losing mass at a rate of 150 +/- 80 km3 each year since 2002. That’s equivalent to about 0.4 mm of sea level rise each year. This is about twice other recent estimates, while IPCC 2001 actually gives negative 0.1 mm/yr. What is especially nice about Velicogna and Wahr’s study is that by using gravity measurements they have measured mass changes directly, avoiding the problem of virtually all previous measurements of ice sheet mass change, which usually measure either input (snowfall) or loss (calving, melting, or thinning of the ice), but not both at once.
What does all this news mean in practice? Reading the editorials in Science, and quotations from various researchers in newspaper articles, one might be under the impression that we should now expect “catastrophic sea-level rise” (as Science’s Richard Kerr writes). Of course, what is catastrophic to the eye of a geologist may be an event taking thousands of years. In the Otto-Bliesner et al. simulations, it takes 2000-3000 years for Greenland to melt back to its LIG minimum size. And while we don’t advocate sticking with the typical politician’s time frame of 4 or 5 years, the new results do not require us to revise projections of sea level rise over the next century or so. This is because even with Arctic temperature continuing to rise rapidly, there will still be significant delay before the process of ice sheet melting and thinning is complete. There is uncertainty in this delay time, but this is already taken into account in IPCC uncertainty estimates. It is also important to remember that the data showing accelerating mass loss in Antarctica and rapid glacier flow in Greenland only reflect a very few years of measurements — the GRACE satellite has only been in operation since 2002, so it provides only a snapshot of Antarctic mass changes. We don’t really know whether these observations reflect the long term trend.
On the other hand, none of the new evidence points in the direction of smaller rates of sea level rise in the future, and probably nudge us closer to the upper end of the IPCC predictions. Those who have already been ignoring or naysaying those predictions now have even less of a leg to stand on. Coastal managers, real estate developers, and insurance companies, at the least, would be wise to continue to take such predictions seriously.** As Don Kennedy and Brooks Hanson write in the lead Editorial, “accelerated glacial melting and larger changes in sea level should be looked at as probable events, not as hypothetical possibilities.”
*Note that we don’t actually have good constrains on the rate of sea level rise from the penultimate glacial period (~140,000 years ago) to the last interglacial (LIG, ~125,000 years ago). However, we have very good data on the more recent glacial-to-interglacial transition, between about 14,000 and 7,000 years ago. During that time, sea levels rose at an average rate of about 11 mm/year, and at rates much higher than that for short intervals.
**Consider for example, that 1 m of sea level rise would change the frequency of what are now 100-year floods in metropolitan New York to once in every four years events. (See here and Rosenzweig, C. and W.D. Solecki (Eds.). 2001. Climate Change and a Global City: The Potential Consequences of Climate Variability and Change – Metro East Coast (MEC). Report for the U.S. Global Change Research Program, National Assessment of the Potential Consequences of Climate Variability and Change for the United States, Columbia Earth Institute, New York. 224 pp)