The further to the left or the right you move, the more your lens on life distorts.

Tuesday, February 21, 2006

Runaway Models

When I was a young engineer, I developed a sophisticated mathematical model to predict the wear characteristics of a metal cutting tools (I know, not very exciting). My computer-based model considered all important variables, produced sophisticated tabular data, and generated graphical charts that wowed the uninitiated. The results of many computer runs made an argument for substantial redesign of the cutting process. There was only one problem, the mathematical model was incorrect!

Models, no matter how sophisticated, can be wrong – dead wrong. And therefore, the results that they produce are useless.

The old saying “garbage in, garbage out” is applicable to computer-based models that are correct. But when software (and its underlying model) is incorrect, even “good data in” will invariably result in “garbage out.”

The argument for global warming or extreme climatic change is based solely on mathematical models. Climatic models use regression analysis and other sophisticated statistical techniques to predict global temperatures, sea levels, and greenhouse gases, not over the next month or year, but for time periods extending to 100 years! Global climate is an extremely complex system, and as such it is extreme hubris -- on the part of environmentalists and researchers -- to claim that these models are (1) inherently correct and (2) accurate over a 100 year period. (It’s worth noting that we currently have trouble predicting the weather more than 10 days in advance). And yet, based on data derived from these models (and selective present day measurements) environmentalists declare global warming a “fact.” It isn’t.

I don’t know if human-induced climatic change is occurring, and I would submit that no one else does either. The global warming hypothesis is worthy of study, but it does not yet justify broad-based legislation or international treaties. Maybe someday it will, once the hypothesis is proven. Then again, maybe the hypothesis is dead wrong, and the human intervention mandated by legislation and treaties will have severe and unintended side effects.

In an off-topic comment posted at The Belmont Club, Wretchard states:

The world has been warming and cooling for a long time. What Global Warming asserts isn't that the world is warming or cooling but that it's warming and cooling according to a particular model. Near as I can tell the Global Warming models are regression equations. Michael Crichton makes the point that complex systems are not reliably modeled in this way and that in the recent past, the same type of regressions were used to assert Global Cooling.

The Global Warming models may be completely wrong, partially right or right in some way. But as I understand it, the assertion is that we have enough confidence in this regression model to climate-engineer the other way. That's a very large claim. The Kyoto advocates are calling for caps, which actually translate to job losses, which translate to statistical deaths from hunger, inability to afford health care, etc. to pay for this climate engineering experiment. You really couldn't convict a single person of murder on the standard of proof the Global Warming folks have, yet we seem willing to bet a whole lot of farms on it.

Yesterday I spoke to someone who said that the Andaman Islands were sinking because the seas were rising. I asked that person to go down to the Sydney Opera House and tell me if it had gone under. "Why?", the question came back. "Because the sea level in the Andaman Islands will not rise separately from the sea level in the Opera House", I answered. But as you can guess, the answer was not convincing. Deep down some people think the Andamans are going under while their beachfront property isn't just because the environmentalists say so.

But many might argue that we’d be better off safe than sorry – honor the environment just in case. What’s the harm?

Consider Wretchard’s further comment on the Kyoto Protocol, an international effort to respond to the predictions of global climatic models:

One of the things that has always bothered me about Kyoto Protocol is that they never examined the downstream effects of poverty far enough. Putting caps on US and EU industrial output may cut back on the production of "Greenhouse gases" but it also reduces the export markets of the developed world, which results in job losses in poor countries.

That pushes people into subsistence livelihoods. They leave the factory and eke out a living [by] farming in the country. They burn down brushland, plant a few crops of spuds, ruin the soil, move on further into the forest. And it does this to people in the millions and tens of millions. Taken far enough it is an extremely destructive environmental practice. That's an effect in a complex system, a system whose characteristics are not well or wholly modeled in the Global Warming carbon climate models which has reduced the variable of interest to a single one which they are going to adjust with UN treaties.

Of course, the MSM spends no time with these subtleties, instead breathlessly reporting that Arctic ice caps are melting (while failing to report that Antarctic ice is increasing) and predicting that the coastal US will be underwater in 100 years plus or minus. That’s what some models predict, but they’re models, not crystal balls.

I guess what bothers me is the sanctimonious attitude of many environmentalists. They cherry pick the data, promulgate selected 100-year models that portend disaster, and forget they every action they suggest has consequences that go far beyond their objectives. This might be okay if the models were provably accurate and the data were scientifically representative of global climate. But they aren’t.