Rowan Sutton: on the amplification of warming over land. That the land warms more than the oceans is well know; but as RS points out the *why* is somewhat less well known. I would have said, unthinkingly, its because of the ocean heat capacity. But… you get the same effect (or similar) in equilibrium runs so its not that, or not all that. He says its because the oceans are wet and so can lose heat through evap, which is harder for the drier land, and this seems likely. Its interesting how well-known-things like this turn out not to have been much investigated.
Some talk of precipitation and its variability between different predictions… keyed to the usual water stress will be a problem in future stuff. Which brings me back to the Australian drought and the stick I got in comments by belittling it. But doesn’t help in knowing what is to come.
Fichefet – LOVECLIM simple climate model – total melting in 3kyr (depending on scenario) with half gone in about 1kyr. Which isn’t very fast (probably slower than the hadcm3 estimates. Mikolajewicz, later, gets 3.4m after 600 y but I think he is using 4*CO2). 6 mm/yr at max. In Q’s: is your model reliable: its quite simple? Maybe. Could the recent glacier speed-up remove it faster? A (from Huybrechts): no. Climate and melting is more important.
Gregory: wot the IPCC did re SLR from ice sheets. Mostly the obvious: that Ant will gain mass and Gr lose (ablation at margins bigger than gain in interior). But high-res models are needed to do this well. Thermal expansion is biggest, followed by small ice caps. But recent Gr/Ant speed-up isn’t modellable, so in the absence of knowing what to do with it, it gets added in as a term scaled with temperature. This then roughly balances the gain of Antarctica. Points out that AR4 ests v sim to TAR, but with reduced error range due to greater certainty.
Rahmstorf think we are underestimating future sea level rise. He has an empirical model simply correlating T to SLR over the 20th C, and gets 3.4mm/yr per 1 oC above pre-ind. Applying this to future T projections gets him a higher rise than IPCC – quite a lot higher, 50-150 cm by 2100. It will be out in Science soon. He does note that obs SLR is higher than the GCMs currently predict. Q: when will the difference between you and IPCC estimates become clear? A: about 20 years. Most of the Q’s focus on individual components, and he answers with “but I’m not looking at individual components”. Best Q: mountain glaciers will run out as a source, so the linear assumption is wrong. Slightly unsatisfactory answer is that other things will take over.
Nathan Gillett: detecting changes in humidity.
Treat of the day is to be Singer, in the anthro session. Wot will he come up with this time? His first point is unclear, possibly even to him… something about pattern correlations apparently taken from IPCC ’95. The second point is back to the trop temperature trends… the tropical problem. Which appears to amount to cherry-picking from the CCSP report (as someone who was familiar with the report and its review points out in the questions). He is pushing the mis-match between the models (warming inc with altitude in the trop) and the obs (not really in the tropics). But to do this he has to ignore all the stuff about problems found with the sondes, and indeed the conclusions of the report itself. So, unsurprisingly, he doesnt actually address any of the detection and attribution stuff at all.
Now its lunchtime. A little chat with that Alistair McDonald. I fear we still disagree about radiative param and LTE.
van Oldenborg: why does Europe warm more than the ar4 runs say it should? Globally, the ar4 models are on the obs trend. But in Europe, they are below it, by 3 SD. And especially over the one Netherlands grid point which they care about (is tis pushing local stuff too hard?). And the suggestion is that this may be due to problems with the THC: which is too southwards (and too deep in the oceans?) in the ar4 runs.
wmconnolley: scienceblogs.com/stoat archive 
