I know what Iowahawk does for a living and it involves crunching massive amounts of high-value data in near-real time.
I’d suggest that everyone read it…
–
I know what Iowahawk does for a living and it involves crunching massive amounts of high-value data in near-real time.
I’d suggest that everyone read it…
–
Iowahawk’s is a truly excellent post, and I recommend everyone interested in the AGW debate go read it. If you’re got the interest and minimal spreadsheet skillz, he shows you how to create your own hockey stick temperature model. There are also a set of ‘exercises for the reader’ that amount to a rudimentary sensitivity test for that model. Even if you aren’t going to run the tests yourself, it’s worth the read for how it puts the jargon into context.
What it does NOT do is get into causal modeling for climate – the so-called ‘forcings’ debate. There’s a lot more to be said about that, and about the positive/negative feedbacks argument. Here’s hoping he keeps the jester’s hat off for a bit longer and takes that path as well.
One thing that Iowahawk didn’t mention is that PCA (principal components analysis) is sensitive to the scaling of the proxies, i.e., they all need to be calibrated in the same units with similar gains. For example, you can’t just combine raw tree ring widths from different locations with different growing seasons and growth rates. Or measure one set in inches and the other in millimeters. This point tends to be overlooked.
Test…
[Well, I finally managed to sign back in to WoC, but only by dumping Typepad for Movable Type. If memory serves, this is the second time Typepad has taken a dump on me. There shall be no third. They may send their inexplicable error messages to someone else.]
I’m proud to say that I knew iowahawk back when we were both comment-creeps, during the heyday of lgf. My respect for him continues to hockey stick.
¡Olé! both for Movable Type and Iowahawk. He just read my mind!
As all this mess began, IMHO, because computers had grown powerful and widespread enough to allow scientists to spend their time running models that cannot be validated, the same may be used to show the average citizen what this issue is about. And macros in Excel or Open Office constitute the perfect way of doing this since, although inefficient, they can handle a model and be run everywhere.
As it already happened regarding UFO’s and Little Green Men, widespread access to computer tools (in that case, Photoshop) demystifies the issue. Regarding the theory of the Global Warming, brings back to the people what was discussed only in a priestly cast, as during the Ancient Egypt.
The interesting thing about the models and such is how simple they are, and how little data there actually is. We aren’t talking finite-element sims running on billion-dollar supercomputers. This certainly makes climate science look far simpler than old fashioned meteorology, which at least needs big fat HPC-class machines to run their sims.
Yeah, I’ve been looking for the global lattices and something-more-than-PCA-and-regressions models that I thought I’d find, and not finding them. Is this really all there is, or am I missing some treasure troves through not knowing the players?