This really grinds my gears. Experts are all like this. The Knowledge Problem has been known now for at least eighty years. Being an “expert” doesn't make one any better a planner than the market is, in fact, their limited knowledge makes them worse almost a priori. They don't even think they “know” that underestimating is more dangerous. It's blindness to their own ignorance, blindness to the possibility that overestimating warming is more dangerous than they can conceive…or, considering the political implications of their belief that the should decide which information is useful for planning, it's Far-Left politics.

That quote is, to those of us “in the know” the most revealing look at RC's ideology one can get.

Fun fact about Greenland:

At the current rate of ice loss (.4% per century) Greenland won't melt entirely for 25 thousand years!

http://wattsupwiththat.com/2009/03/21/recent-oc...

This concern is well placed. I won't get much into the data quality issues (Buckets and Inlets etc.) But the physical interpretation of a relationship between SST and sea level is pretty tricky. From the surface temperatures, one could not tell if the bulk of the ocean increased in heat content or just a shallow layer-the thermal expansion depends on the total heat content of the oceans, not the temperature at the surface. Over long time scales this may not be a big issue. But that begs the question of how quickly heat diffuses from the surface to lower layers, or vice versa. My understand is that this is tied into lots of other climate issues, including sensitivity.

http://4.bp.blogspot.com/_TDdxO2FHJPs/SQ2wsVEvo...

Intuitively, it should not be linear because expansions of the oceans are not just upwards but sideways as well. I have not seen an estimate of the various volumes of water that would cause successive 1 mm rises on addition, taking into account the shape of the basins to be filled.

Then there is the problem of the whole ocean. Strictly, a thermal increase in volumes requires all of the volume, on average, to be heated. It is not adequate to show that the top x metres have heated. It is also required to show that the rest of the oceans have not cooled. I suspect that the data are too sparse to model whole ocean temperatures for expansion, given that currents are altering grid cell temperatures all the time; and that some currents have too few data points, like those that flow under ice.

Then there is lag. I have not seen a paper that uses buoy data to estimate thermal lag with depth, though it should be possible to get some idea from day/night changes. There might be papers, I just have not seen them. If you have a delta T at the surface, how long does it take to propagate to X, Y, Z, m depth? And then the reverse movement as it cools? Surely one needs a handle on this to try to dissect causation.

Finally, I have grave uncertainty about the meaning and accuracy of SST. Is it merely a parameter that has no great significance but has a convenient shape with time as Andrew notes?