IanC::
I had a chance to read the OP link tonight.. It's THIS graph (the 450,000 yr record) that caught my attention..
On that scale, the most surprising behaviour is NOT the "which comes first" proposition which I've weakly seen sometimes on other timescales. But more consistently -- it's the "which goes last" proposition that seems to suggest some darn thing. On almost every major FALL in temp -- the temp precedes the "supposed" forcing function. Kinda violates the concept of a forcing function on that principle alone doesn't it?
Either the lags in the proxies are wrong and some axis diddling is required or the current CO2 theory would have to say that the feedbacks somehow go mostly negative at some point in the CO2 rise..
<edited for a 3rd possibility> OR the CO2 ice proxies are somehow "filtered" themselves to add delay. Perhaps liquid mixing? Or limits on the dating resolution when the cores are taken.
Forcing functions should predict BOTH the positive AND negative slopes of temperature.. Eh?
That's a good point and one I had missed. Goos catch.