Ozone depletion has been tied to flurocarbons, mainly, and laws are in place to restrict their use. Global warming is not impacted by ozone depletion any more than by methane or CO₂ for that matter. The key is the impacts of water vapour, precipitations, and their link to the sun’s energy. Weather people cannot predict the weather’s precipitation one week out, because of the severe complications of precipitations, rain, snow, on weather systems often miles from YOUR current weather. However, political weather folks want YOU to believe they can tell you what will happen DECADES ahead. How naïve.

You are right on target with your comments. The concept of P and R²Â is a bit much for most to handle. R² seemed more intuitive, alone, though flawed. I write about nonsense correlationships in my applied stats book, so I understand what you are saying.

The whole arguement about correlationship of CO₂ to Temp change is also flawed and I should leave it alone. Since water in the atmosphere overwhelms the idea of CO₂, such information about CO₂ and temperature change is nonsense to start with as an analysis basis. Thanks for your positive comments.

The ozone layer issue has never been posited to have any major effect on global temperature, as far as I can tell. UV radiation does not transmit very much heat, it’s the wrong wavelength. Even though its energy eventually gets transformed to heat, it has not been accused of being a major contributor. Interestingly, the ‘hole’ was discovered right after we gained a means of measuring it for the first time. In all the excitement, I have not been able to find any evidence for there not being holes in the past. In other words, the assertion is made that the hole is man-made, and nobody asks whether it could be a natural phenomenon that was periodically there before we could detect it.

The ‘psycho-historical analysis’ is right there in the media for anyone to analyze for themselves. I remember reading of “ozone holes over Kennebunkport, Maine” while Bush Sr. was president. The advocacy of an issue by researchers using the press to pressure politicians for public funding soon becomes transformed into government money being used to fund only research that promises evidence to support the advocated issue. Soon, the new orthodoxy purchases an avalanche of data in support of the hypothesis, scare stories get ratcheted up, evidence to the contrary is suppressed, and the few men of integrity who stand for the truth are subject to ridicule and attack. Consensus and public opinion completely and overwhelm reasoned arguments contrary to the orthodoxy. This ‘psycho-historical’ phenomenon is as old as history. Environmentalism is simply the new state religion. Will it be successful in creating a new dark age?

–John Danforth–

Yours Truly,
Pil

Thanks for posting this great article by Mr. Baldwin.

I am going to pick a couple of nits before I offer my praise.

Mr. Baldwin is correct when he states that higher values of the coefficient of determination (R-Squared) tend to show a stronger relation between a predicted variable and its explanatory variable. In this case, the predicted variable is recorded temperature and the explanatory variable is the CO2 level.

First nit picked: a high R-Squared says nothing about causation – it says only that there is a high degree of correlation observed between the two variables. Lots of folks get this wrong. To be deliberately ridiculous, we may observe a R-Squared of 1.0 between, say, the winner of a horse race and the “coolness” of the winning horse’s name as voted by the audience, yet clearly there is no causation there. There may in fact be causation in other less ridiculous scenarios, but the R-Squared never proves it. Yes, I’m a nit-picking actuary geek.

Second, there are instances when a R-Squared of 0.21 actually accompanies a good model. Yes, 0.21 is low as far as R-Squareds go, but a more useful (in my opinion) diagnostic is to look at the “P-Value” of the explanatory variable. The lower the P-Value, the better. The P-Value says: what is the chance of observing the result I observed if in fact there is no underlying relationship? Thus, the P-Value of the CO2 explanatory variable may have been, say, 0.35, which would mean: there is a 35% chance of observing what I observed even if there is no true underlying relationship between CO2 and temperature. Statisticians generally like to see P-Values under 0.10, and under 0.05 are even better. My belabored point: it is possible for a model to have an R-Squared of 0.21, but a P-Value under 0.05.

OK, no more nit-picks. His idea to use the SCPI measure is brilliant, and this idea deserves greater exploration and, assuming it delivers good statistical correlation, as I gather it does, then this idea deserves the same level of attention as the blowhard Gore.

Well done!