|Turned up to eleven: Fair and Balanced|
Thursday, March 28, 2002
Skeptic magazine, which contains excerpts from E.O. Wilson's and Bjorn Lomborg's new books, along with a rebuttal to Lomborg by one of the scientists he criticizes, Dave Pimentel (of Cornell). I will post my thoughts on these three pieces, which I found very interesting, in a few days when I have time for an extended treatise.
Returning to the topic of Global Warming, let me first say that I think it is likely to be a real phenomenon, but I agree with the skeptics that we have very poor knowledge of what exactly the anthropogenic part of it are, what the extent of it will be, or what its effects will be. Jay links to an article about the psychology of global warming politics, which is enlightening in many ways. I get the sense that Jay is very skeptical of GW, and thinks that we will learn we were wrong. I agree in a limited sense, and I think that what we will learn is that linear regression analysis is incredibly dangerous, because it leads to incredibly poor predictions. This may seem like a non-sequitur, but here is my reasoning. Scientists trying to study this or many other phenomena will try to build models, and one of the most common modeling tools in regression analysis. Every lab student has taken a bunch of data and tried to fit a curve to it. Linear regression analysis is the simplest form of this, where we just use a straight line, of the form y=mx+b. More complex analysis may use an sinusoidal, exponential, or polynomial function, but in essence, what we are trying to do is predict the future from the data. As the commercials for mutual funds used to say, past performance does not predict future results. When a system is very large or "complex" (in the mathematical sense), the regression tools have not been figured out yet, so the predictions are poor. This is why we had predictions that humans might bring on an ice age, and now we have predictions that everyone in CA will drown. Who's right? Who knows. The smart scientist will withhold judgment on that issue, and stick to the available facts. Short term temperature, especially regional temperatures, is erratic to say the least, and for a while, idiot newspaper reporters are going to stick GW into every story about an unusually warm summer somewhere or a mild winter somewhere else, oblivious to the notion that we are talking about Global Warming, not local warming. Anyway, the same fallacy is present in budget projections, projections of poverty, inflation, stock market growth, and any number of other things. So, my advice, assuming anyone wants it, is to take any projections of what anything is going to be like in 50 years with a huge grain of salt.
Still, common sense dictates that we should be thinking about this sort of thing. Rand Simberg has been discussing the potentiality of an asteroid or comet impact with me and others, and while I am a bit skeptical of some of the numbers thrown around by the people who worry a lot about this, I am in favor of efforts to quantify the risk by improving our survey of the near-Earth Solar System, and plans for avoiding a catastrophic impact should certainly be made. At the same time, on the GW front, just because we aren't sure it will be a problem, we shouldn't disregard it. As scientists improve their climate monitoring and modeling, better predictions will be possible (hopefully) and the scope of the problem, large or small, will be more precisely pinpointed.