Very clear article. So clear you can actually draw additional points from it (not called out by the author). One of the problems of using regression on time series is the model is much better at modeling a grand mean than getting the variation right. So if you take an recent upward trend of something you are measuring and then use regression to get data for past points you don't know- you are likely to predict a fairly flat past, almost no matter what the nature of the original data. You need to (at the very least) cross validate by holding out a large contiguous interval of the past where you know the temperature and the proxies to see if the model tracks.
I have translated the procedures from OpenOffice to R (didn't bother bounding the date-range where the fitting is done, so I get worse fit and bit more variation; you can only replicate so many mistakes before you feel dirty): http://www.win-vector.com/blog/2009/12/cru-graph-yet-again-w...
Fascinating stuff. I'm not a professional statistician, but I was mostly able to follow along. Here are the two most interesting statements from the piece to me:
> Foul play is assumed and the hunt is on for a culprit; a natural suspect is man made CO2, which has increased coincidental with temperatures. It's a small part of overall global greenhouse gases (5.5% if you don't include water vapor, 0.3% if you do), but maybe -- just maybe -- the atmosphere is in a delicate, wobbly, equilibrium balance.
It never occurred to me that human-produced carbon-dioxide was such a small percent of the greenhouse gases.
Then there was this:
> But a funny thing happened on the way to Copenhagen: a couple of Canadian researchers, McIntyre and McKitrick, found that when they ran simulations of "red noise" random principle components data into Mann's reconstruction model, 99% of the time it produced the same hockey stick pattern. They attributed this to Mann's method / time frame for selecting of principle components.
That's just a wow. Incredibly damning of that model, and that model was one of the biggest calls to urgency in the field. I'm happy this sort of information is coming out and being thought about.
> It's a small part of overall global greenhouse gases (5.5% if you don't include water vapor, 0.3% if you do), but maybe -- just maybe -- the atmosphere is in a delicate, wobbly, equilibrium balance.
Considering what we know of CO2 levels and temperature in the geologic past, we're far away from any potential tipping point.
damn it. everytime i hear about the hockey stick i first think of an actual hockey stick. i guess i'm not as familiar with the greenhouse debate as i thought i was. or maybe it's just that time of year. i make my own underwater hockey sticks, and have wanted to give a go at the frozenwater kind now that the ponds are about to freeze...
