3.08.2009

Global warming: a stance

I read a couple articles this week that were well-thought about the issue of global warming; something that one does not see much. The first was by a Princeton physicist, saying that the amount we know about the climate is not adequate to trust computer models. Having had the pleasure in the last several months to interpret the results of linearized partial differential equation modeling from CFD and FEM programs, I agree wholeheartedly with this assertion. I can't even trust a simulation of the stress on a beam in static equilibrium, a modified and controlled case of a simple real-world breakage test I did in sophomore year of university.

Most people haven't had the joy of tangling with non-linear PDE's directly, but everyone has watched a weather report generated from them. We know to take the proclamations therein as a mere probability. Now, it is a simplistic statement that invites simplistic rebuttals to compare weather forecasting to climate modeling. The inputs are completely different. But the math involved is the same, and it is the same math that is used to predict the airflow over an airplane, and the breakage of a truss, and the interaction of plasma in a fusion reactor. All are based on deceptively simple-looking equations which have no simple solutions. All these problems are made tenable by certain simplifying assumptions like, "turbulence can't be predicted, but it can be averaged" or "when this angle is really small it is approximately the same as this distance," or "this feedback mechanism is really a curve, but if we look at a really small part of it it appears to be a line," or "well, we know how to solve this really well, but if we did it really well it would take all the computers in the world a million years, so we'll just do it kind of well on what we have."

Depending on the assumptions, resources available, how well they are calibrated by real results, and how well you can confine the problem, these nonlinear equations are correct to a certain degree. I can be reasonably certain that an FEM calculation I run on a structure will predict breaking point to within 50% if I set it up right. I can go on intellicast.com and be fairly confident in the forecast for the next 3 days or so. This "correctness" is quantifiable against real results. What is more, it is predictable by studing the limitations of the model - in other words, a mathematician should be able to look at any nonlinear model and say something like, "well, this will be correct within 10% approximately 90% of the time." I have yet to see a comprehensive discussion of this nature in the popular science media, and I must admit I have not had the time and motivation to look in the scientific press. I suspect from my limited experience with these things that the numbers for time-dependent non-linear feedback model, calibrated by a reliable sample size on the order of 10^-7 the total dataset (billions of years of Earth vs. hundreds of years of data), with "success" determined by variations in output of about 1% (i.e., 303K vs. 300K), are vanishingly small.

My critique of the professor is that, beyond models, there is a large body of evidence that certain things are happening in certain parts of the Earth. But with the current sample size, it is difficult to rationally blame them on higher levels of CO2. Hurricane frequency and ferocity is somewhere between completely random and long-term cyclical. Desertification is mostly a product of farming practices and much larger forces than anything we have yet brought to bear. Even melting glaciers, ice caps, and rising temperatures remain on a trendline from the end of the last ice age. They don't really mean anything right now, but eventually it might. Someone should measure these things. Being a tragedy of the commons issue, I wouldn't even put up much of a fight if government insists on being the one to do it.

The second article was up on boston.com, and it was mainly notable for the comments at the end.

I wouldn't call myself a global warming denier. I'm prepared to admit that humanity, for the first time in history, is at a point where it can alter the climate on a global or at least multi-regional scale. We are in the first stages of a Kardashev level 1 civilization - we are gaining technological power (and responsibility) on the level of an entire planet. I am not ready to say that we understand how or how much we are altering the climate, and I am certainly not ready to say that it is an unmitigated disaster. Certainly, our previously-attained abilities to control plants, animals, metals, and electricity have had locally mixed but globally positive results; and probably began with similar ineptitude and alarm.

Anyway, to those comments. freddysavage writes:

Regardless, I don't understand why anyone wouldn't be in favor of limiting the pollution we dump into the environment. A study released a couple of weeks ago attributed a half-year increase in the average life span to the curbs on pollution over the last 20 years. What dopes are in favor of polluting, other than the companies that own the coal plants and oil refineries?

...says a person who, I am willing to wager money, heats his house, drives a car, eats food that was not attained through hunting and gathering, shits, and therefore pollutes. The hypocrite probably hasn't even invested in solar panels for his roof; if he did he would begin to understand the concept of "30 year return on investment." "Give a hoot, don't pollute" is a fine sentiment, if you're in second grade.

For those of us who live in the real world, CO2 is not a pollutant. Sure, enough of it will kill you, but so will water. CO2 is a potent fertilizer and drought resistance agent for plants. It is a macronutrient. It's direct effect on the capture of solar energy is negligble, mostly because it absorbs in a similar range as water vapor, and there's a whole lot more dihydrogen monoxide out there (and its a better greenhouse gas). All the hullabaloo about CO2 rests on possible feedback cycles, like slight warming will cause oceanic outgassing, which will cause more warming. Except that slight warming also causes more clouds which causes less warming. So no one really knows, and until we have a really good data set to look at we're relying on models, which I've already disgussed. The main point is that it's not going to kill us all. Volcanic periods have raised CO2 levels far, far higher than they are today and the Earth's climate still stayed in a habitable range through all the fluctuations. So, fearmongering aside, business as usual, but with agressive monitoring of the factors that will improve our theories, is probably the most sane course of action.

Ender3rd1 writes:

Little JJ [the article's author] may not even read his own posts. Although it may ppear to be a text from a human, JJ is a T-1000. You see a web developer generates entire websites by changing user names in a database. He can rollout several hundred a day. JJ is no more real than Sarah Conner.

It probably lowers the level of discussion here to even include this, but there were a few comments on the lines that this is all a massive conspiracy by oil companies. This is a line of thought I usually associate with UFOlogists and paranoid schizophrenics. However, I've heard this from both sides. Al Gore is bent on global domination. Exxon killed cold fusion.

On a slightly more sane level, mouthing along to Obama campaign dogma, ReasonedReply writes:

Consider the players and the real struggle to control the message. The dominant energy industries don't want to change. Change is needed and very do-able, and whole new enterprises will emerge that need trained professionals to run them.

This is irony. I see someone decrying one perspective of global warming as "fear of change." But what is global warming? All but the most pessimistic, Venus-type runaway predictions show a world that has undergone some change, but is certainly distinguishable as Earth. Clearly this man or woman believes that all change is not bad, so why must global warming be bad? Step back from the white man's guilt and ask yourself: what is honestly worse - a slightly warmer world, or force-feeding a change through political means for which we do not yet possess the capabilities technologically or economically? Looking at the history of history, I'd say the former is probably good, while the latter is almost certainly bad.

No comments: