End Times, Or: Global Warming!
The writers of South Park are creative geniuses. Their new season started this week with an episode mocking “The Day After Tomorrow” by having people run back and forth trying to flee what seemed to be a very physical, if invisible, global warming. Of course the boys are walking around, completely fine, while the idiot adults cower in fear.
Why is the debate on global warming so heated? Why, back in college, was my class asked to give presentations on global warming, only to see those who disagreed that it was occuring literally silenced and told they’re idiots by my professor? Why are we told that there is a complete consensus of scientists that global warming is occuring and that any dissension is foolish and solely motivated by corporations?
I’m thinking there’s some kind of white guilt/self hatred behind it. Primarily this is driven by a need to correct our world and to feel as if we’ve been doing things wrong and that we can and must take corrective action. We see our social ills and need someone to blame. Global warming is caused by humans because if we can’t blame ourselves, we can’t fix it. My problem with global warming is that much of it is junk science. Oh sure, some of the observations are solid. I’m not going to argue with localized observations of certain effects, or even satellite imagery. I’m not obnoxious enough to believe that the scientists involved are idiots. I do believe them when they say that they truly believe that CO2 levels are causing global warming.
My central problem with this assertion is that it assumes that climatologists have a fully integrated understanding of Earth’s climate. Earth is one of the most complex systems known to man. It involves countless interactions, and I have trouble believing that any one climatologist, even less so a panel of climatologists, has the knowledge or mental capacity to incorporate all that is going on in Earth’s climate and make the conclusion that the high CO2 levels are directly causing global warming. The climate models that attempt to predict the future effects of global warming invite even more doubt.
Let me divert in order to explain how I can back up my doubts. My fear is that there is a lack of statistical rigour being applied to these predictions. Much of what we know about global temperatures is based on data going back to the 1800’s. I believe much of this data. Temperature was recorded based on mercury thermometers. Mercury is a chemical and once enclosed in glass there really isn’t terribly much that a recorder could do to screw up the recording of the temperature data. Any analysis of a plot of this data would certainly account for minor variations due to anything that may have happened that particular year, as well as much smaller variations attributed to measurement. Most of this data, once plotted, shows an essentially flat line. However, this data is still based on a measurement of less than 200 years. Do me a favor. Hold a magnifying glass to one of the keys on your keyboard. Now, based on that, make assumptions about your entire keyboard. You see my problem.
200 years of data is less than a hiccup in trying to understand global climate. So now we look to indicators we can use. Observations of environments. Deserts, lakes, etc. Observations of rocks. Hey, let’s look at the ice cores. We’ve drilled ice cores from Antarctica, and by correlating the amount of deuterium in the ice with the global temperature, we can chart the climate for several thousands of years. What does it show? The damn thing looks like a seismograph. Rapid climate change is built into the Earth’s history.
But CO2 does cause warming, right? Yes! It can be demonstrated in a laboratory. You shine a light on CO2 gas molecules and they absorb it and you have increased heat. CO2 also traps hot air, and it is alleged that it creates a feedback loop with the Earth, raising surface temperatures. So yes, increased CO2 does cause a greenhouse effect. The real question, however, is a matter of concentration, and cause and effect. The biggest mistake that is made is believing that a correlation between CO2 levels and global surface temperatures implies cause and effect. Still, most scientists are intelligent enough to know the difference. So why do I say I have a problem with the statistical rigour of global warming analysis?
Regression analysis. Let’s say you’re trying to account for three different factors. Atmospheric CO2 levels, Atlantic Ocean effects (gulfstream, hurricanes, something like that), and an estimate of total coverage of forests based on satellite imagery. If you ran the test and could somehow magically turn the knobs of these conditions, you would have to have different combinations of the levels of these factors. Let’s say you do a simple experiment, two levels. So your experiment consists of CO2 levels of 200 ppm and 600 ppm, Atlantic Ocean effects of 10 hurricanes with 150+ winds and 20 hurricanes with 130+ winds, and 30 billion acres of forest and 10 billion acres of forests. The “runs” of your experiment would look like this: a combination of 200 ppm CO2, 10 hurricanes, and 10 billion acres of forests, then a combination of 200 ppm CO2, 10 hurricanes, and 30 billion acres of forests, and so on. The last combination (the eighth) would be 600 ppm CO2, 20 hurricanes, and 30 billion acres of forests. You’d then look for the main effects and the interactions. So if your output was global surface temperature, would the number of hurricanes be the main driver of global surface temperature? Or would it be the CO2 level, the acres of forests, and the interaction between those two factors? The interaction is seperate from the main effects. The equation would look like this: A + BCO2 level + Cacres of forest + DCO2 levelacres of forests. The units would be different on the interaction, it’s a different beast. Or, what if the measurement is mostly noise on the main effects and only the interaction itself is important? I’ve had it happen before in regression analysis.
I’m doing a poor job of explaining the statistics, and I do apologize for that. I’m sure it strikes you that a true climate model would need to account for many more factors than these three. And I’m sure these factors have gone into the predictive models being run currently on supercomputers. The computer simulations cannot magically predict what will happen to the Earth’s climate. They must be fed predictive equations, which inherently may be confusing CO2 levels with the Earth’s surface temperature. The problem is that, again, we only have so much data. On the global time scale of the Earth, this is nowhere near enough data for us to predict climates. Any attempt to predict the climate is confounded by this correlation. Solar radiation levels actually correlate decently well with a rise in surface temperatures. I would imagine it would be very difficult to statistically determine which one of these is the primary cause of global warming. I don’t dispute that the temperature of the Earth is rising, that much is obvious.
Again, the problem has to do with the assumption that we can make predictions about the Earth’s climate. I read an article anouncing that global dimming is occuring, which I do believe because of the increase in SO2 and NOx, which is why new technologies such as recycled-stream coal are important to filter this crap out of our smokestacks. Then the article claims that although there’s some study saying that the Earth today is less sensitive to CO2 rises than in the past and the effect on temperature is less than we thought, that analysis may have been thrown off by global dimming, and therefore global warming is worse than we thought. Yeah, that caused a question mark to appear above my head, too. I’m thinking that what they meant is that it would be worse, if global dimming wasn’t occuring. I must again raise the point that it is next to impossible to un-confound all of this data. Essentially, any analysis of the rise in Earth’s surface temperature must begin with an assumption: is this caused by CO2 levels, global dimming, or solar radiation?
The problem with most of the climate models is that, without a fuller understanding of all the factors and interactions in the Earth’s climate, acceptance of these models must be based on our assumptions and what we can immediately observe. I get from wikipedia that the acceptance criteria for whether or not a model can predict climate is whether it accounts for specific observed effects, such as El Nino. The problem here is that it doesn’t get us away from basing our predictions on our assumptions. Any good scientist must make assumptions, sometimes even wild ones, in order to begin their work. However, the output data that confirms the predictive models is again dependent on our assumptions. Allow me to clear this up a bit by going through an example:
We make an assumption. CO2 levels are causing global surface temperatures to rise.
We run a statistical analysis on global surface temperatures and find that only the main effect of CO2 levels is significant.
We take the predictive equation from this analysis and perform a simulation.
We see if our model accounts for El Nino. It does!
We conclude that our model is correct and then look at our prediction for global surface temperatures.
OH MY GOD! OCEAN LEVELS WILL RISE BY LIKE A BILLION FEET!
I’m simplifying and poking fun, but you see the inherent flaw in this analysis: we assume that because our model accounts for certain effects, such as El Nino, it must also be accurate in predicting the effect of CO2 levels on global surface temperatures. Now I’m pretty sure those two aren’t related. Other factors built into our simulation, such as volcanic activity or something, may be accounting for El Nino. I know that scientists must have some criteria for determining the strength of a model. However, in order to successfully predict an output the confirmation of a model must measure that output. So, again, any model will run into the CO2 levels being confounded with the rise in temperatures.
Personally I find solar radiation to be much more convincing. However, that goes with a huge grain of salt. I won’t deviate from my assertion that it’s all confounded as hell. So why are so many scientists trying to stick with the CO2 theory? Honestly I do believe CO2 levels have some effect on global warming. You’re seeing more of it now: the idea that this much of the warming is attributable to solar radiation, this much is attributable to CO2 levels, this much attributed to the number of rice patties in China. I’m serious, methane is a greenhouse gas. Not that I’ve heard any scientists blaming the rice patties. I feel kinda racist even saying the phrase “rice patty”.
I’m not naive enough to believe that people with Ph.D’s in statistics haven’t already worked on the global warming stuff. However I strongly believe that a rigorous statistical discussion is being shut out of the debate for political reasons. Global warming has become a movement now, with a life of its own, and as with any idea with enough backers, it will bite back. It will suppress facts and logic all in the name of proving itself true. The problem, I believe, has to do with the way the human brain stores information. The brain will rationalize certain thoughts in order to block conflicting information. I think someday we’re going to find that writing new thoughts into the brain requires some stupid chemical and the brain doesn’t have enough of it so it secretes it sparingly, which is why we so often get angry when our ideas are challenged. This issue is important, which is one factor making the argument so charged. Another factor is that we’ve all generalized the opposition. Those who support the CO2 theory must hate corporations. You saw it at the beginning of this post, I reasoned that those who think we’re the primary cause of global warming must necessarily be white liberal self-haters. You can also generalize my side of the issue, saying I’m an evil corporate fatcat American who cares nothing for the world or the future or mother Earth or the marijuana fields. See, I did it again, I generalized that those who support the CO2 theory are liberal idiot potsmokers. The fact is many idiot Republicans smoke pot too. Also lots of idiot Liberterians. They all smoke pot too.
So what should the discussion really be about? Well, we can figure out a better way to figure this one out, or continue the slap-fest until we decide that one side is right. We’ll all still probably be totally wrong.