New Antarctic Glacial Melt Study Slightly Increases IPCC Rate Estimate

There is a new study by a French/English team looking at the rate at which Antarctic glaciers might contribute to sea level rise, due to global warming, between now and 2100 and 2200 AD.

The study produces several estimates, but suggests that glaciers in Antartica might contribute as much as 30 cm by 2100 and 72 cm by 2200. That is a large amount of sea level rise, but it is actually less than other studies that rely more on paleoclimate evidence have suggested. I personally have something of a bias towards paleo evidence; Good paleo evidence is evidence of what actually happened, suggesting that contradictory results form modeling that does not make direct use of paleo data is suspect.

The new study, by Catherine Ritz, Tamsin Edwards, Gaël Durand, Antony Payne, Vincent Peyaud, and Richard Hindmarsh came out today in Nature, and is called "Potential sea-level rise from Antarctic ice sheet instability constrained by observations."

The results of this paper raise key IPCC estimates of sea level rise by a tiny bit, which is conservative, as the IPCC estimates are probably low (again, coming from my paleo perspective).

This study looks specifically at marine-ice-sheet instability (MISI). This is the very difficult problem of how ice sheets that are grounded on bedrock sitting below sea level deteriorate. The full-on collapse of such ice sheets has not been directly observed, and it is a very difficult process to model. I liken it to trying to solve the following problem.

An engineer, a theoretical physicist, and a paleoclimatologist are at a wedding. There is a ice large sculpture of a swan on a flat topped table, for decoration. The three start a betting pool on how long it will take for the entire swan, which has already started to melt, to end up on the floor.

The engineer notices some of the meltwater dribbling off the back of the table. She places a set of beer mugs under the streams of water, and records how long it takes for a measured amount of liquid to accumulate. She uses this to generate a graph showing melting over time, estimating the volume of the swan by looking it up in his manual on Ice Sculpture Specifications, and suggests that it will take eleven hours.

The theoretical physicist estimates the volume of ice by assuming a spherical swan, measures the air temperature, and calculates the rate of conversion from ice to water using thermodynamics. He comes up with a different estimate, because the engineer forgot to account for density differences in ice vs water. He estimates that the swan will be entirely the floor in eight and a half hours.

The paleoclimatologist disagrees, and says, "It will take between one and three hours for that swan to be on the floor."

"Why do you think that, you are clearly an idiot, and I am clearly a physicist, so I must be right!" says the theoretical physicist.

Just as the paleoclimatologist is about to answer, the already melting neck of the swan breaks, and the upper part of the neck and head fall backwards, knocking off one of the large wings. All of those pieces slide off the table and crash on the floor. The stress of the impact causes the second wing to break off, but it stays on the table, but it begins to slowly slide toward the edge, clearly about to fall off as well.

"Because," the paleoclimatologist says. "Last wedding I went to, that happened."

I think you get the point.

Ritz, Edwards, et al. try to address the problem by using what they claim to be a better approach to modeling of ice sheet disintegration. From the abstract:

...Physically plausible projections are challenging: numerical models with sufficient spatial resolution to simulate grounding-line processes have been too computationally expensive to generate large ensembles for uncertainty assessment, and lower-resolution model projections rely on parameterizations that are only loosely constrained by present day changes. ...Our process- based, statistical approach gives skewed and complex probability distributions ... The dependence of sliding on basal friction is a key unknown: nonlinear relationships favour higher contributions. Results are conditional on assessments of MISI risk on the basis of projected triggers under the climate scenario A1B..., although sensitivity to these is limited by theoretical and topographical constraints on the rate and extent of ice loss. We find that contributions are restricted by a combination of these constraints, calibration with success in simulating observed ASE losses, and low assessed risk in some basins.

Nonlinear relationships. That is the swan's head falling off.

Like another recent paper on Antarctic ice sheets, other studies as well as the paleorecord conflict with the present study enough that this study has to be reviewed carefully before we can assess its contribution to understanding Antarctic ice sheet melting. It may be right, and that would be good news in comparison to some of the higher estimates. However, ice sheet deterioration is very complex, and it is possible that this modeling effort does not account for enough of the important variables, and may not be detailed enough to be reliable. The authors note some of these problems.

It will be interesting to see how other scientists working on this problem respond. I'll keep you posted.