Thursday, September 22, 2011

Climate change: An accounting problem

The latest RealClimate post, which describes the latest ice melt data from Greenland, features this really important figure. It illustrates an issue that arises every semester in my climate change course, and is, in a sense, fundamental to understanding to biogeochemical cycles and issues like why carbon is accumulated in the atmosphere.

The figure shows model-based annual anomalies (thanks ED) of snowfall (reddish-orange), water loss through surface melt and runoff (yellow) and net accumulation of mass (blue), all in Gt/yr. The key point is that an ice sheet shrinks not simply because it is melting (yellow), but because the loss of water through melt (yellow) is greater than the gain through snowfall (red). The difference is the change in mass (blue). It is an accounting problem; like your bank account, you need to look at the debits and the credits to know whether the balance is changing.

This is the same fundamental concept that  underlies carbon dioxide accumulation in the atmosphere, and as MIT management expert John Sterman has shown, befuddles most people. Carbon dioxide is accumulating in the atmosphere not simply because we are burning fossil fuels and clearing land, but because the flux in to the atmosphere from those sources is greater than the flux out (to land, and the oceans).

As an aside, in my climate change course, one of the many ways we discuss these points is by watching the infamous "CO2 is life" advertisements created a few years back by the Competitive Enterprise Institute. The "Glaciers" video cites scientific evidence for snowfall-driven growth of ice sheets in the interior of Antarctica to suggest that ice sheets around the planet are not shrinking. The mistake in the ad is that in order calculating whether an ice sheet is shrinking, on net, you need to do the full accounting of all inputs (from snow) and all the outputs (from melt), not just cherry pick one part of the ice sheet, or one flux in or out.

The CEI ads, by the way, are hilarious. If you've not seen them, you really must. They are well worth two minutes out of your day; the Onion News Network couldn't have come up with fake ads that funny.


Ed Davies said...

The figure shows model-based estimates of snowfall (reddish-orange), water loss through surface melt and runoff (yellow) and net accumulation of mass (blue), all in Gt/yr.

No, the figure shows the anomalies in these estimated rates. The difference is important as without knowing the absolute magnitudes of the rates (even approximately) it's hard to know if variations in the anomalies are likely to be significant.

A back of the envelope calculation shows that these annual fluxes are about 1/10'000th of the mass of the whole Greenland Ice Sheet so it's not hard to wonder if the change is well within the limits of year-to-year variations.

More information is needed. I understand that measuring the absolute rates would be harder than measuring the anomalies but still some order of magnitude estimates would help.

Simon Donner said...

Ah, you're right, I neglected to say the data on the figures are expressed as anomalies. Thanks for that. While that does matter for determining whether the change in the total ice mass of Greenland is
statistically significant given the year-to-year fluctuations in the fluxes over time, this post is simply showing the ice mass balance to discuss the accounting principles which underlie key problems in climate change. The statistical interpretation of these modelled ice mass changes in Greenland is another subject (one we actually deal with in the same course!).

If you want to calculate the change in the mass or volume of a reservoir, whether an ice sheet or the atmosphere, from year to year, you need to contrast the annual inputs and outputs. That seemingly basic concept is actually a point of confusion for many, who simply hear screams that "ice sheets are melting" and "carbon dioxide concentrations are increasing".