Time-series analysis is applicable to greenhouse gas formation and climate change.
I have a great deal of familiarity with the techniques being used; the usual list of statistical mistakes common in environmental research doesn’t seem to be made here.

What they have done is construct a 3 variable VAR with non-stationary data, and bootstrapped the distributions of causality tests. In short, they’ve done relatively standard analysis that a macroeconomist would be comfortable with. I’ve only spent about 10 minutes with the paper, but I’m OK with what I see so far.

And they conclude that global temperatures decoupled from solar radiation over the last 50 years, but did not decouple from greenhouse gases.
The full article is in Environmental Research Letters.

No comments: