Friday, June 2, 2017

- Climate Change Groundhog Day

Climate change is the perfect liberal boogie man, and it's back again like Jason from the Friday the 13th movies. It's pervasive, tenuous connected to everything, and therefore justifies making every single moment of every single life on earth political. Even better is that the facts are complex enough so that it's more or less unknowable to the vast majority of ordinary humans.

But it isn't unknowable by guys like me.

I’ll come clean here, I am absolutely one of those people who denies that reducing man’s impact on climate change is the only thing that will save us. It’s nonsense. It’s exactly like that idiotic movie ‘The Day After’ which featured good looking people from Hollywood running as fast as they can from the glaciers and they thundered south on 5th avenue. It takes a slow process and makes it seem fast, critical and scary, when the truth is that we will be able to easily adapt to any changes it may bring.

Upon what do I base my conclusions? I don’t know, maybe the main cause of my unbridled arrogance in this space is my 25 years as a data scientist trying to tease causality and correlation out of big complex data sets, and my actual experience – successful experience – at building models that successfully predict the future with that data… successfully.

When Terrence Mann’s famous hockey stick data was made public I downloaded and analyzed it. Myself. Personally. I didn’t farm it out to the slide rule geeks to tell me what they think, I am that slide rule geek. So I looked at the data, and in 15 minutes I knew it was fake. A lot of other people did the same thing that I did. And when the world's 'climate scientists' learned of our copious rejection of it, the great leaders of the global warming ‘science!’ debate instantly knew what they did wrong, and stopped releasing data to the public.

Explaining how I knew sounds like so much cryptic jargon because data science is hard. You can’t learn it from a class in probability and statistics. You can learn what the tools are from that class, but that isn’t enough to learn all the possible mistakes you can make with them. So let me just tell you the thing I learned that had me marching around the Caxton trading floor 15 minutes after I downloaded the data proclaiming the hockey stick and Terrence Mann both as utter frauds.

The earth’s weather is a closed, mean reverting dynamic system, and Mann’s hockey stick showed a 5+ standard deviation moment. His model predicted a 9+ standard deviation moment.

Even if this were absolutely correct, to extrapolate a 9 standard deviation moment from a 5 standard deviation moment is ridiculous. It’s the kind of mistake a high school freshman wouldn’t make. It’s not only an incorrect assumption, but an assumption which assumes that everyone else is an absolute imbecile. Closed mean reverting systems don’t do that because as the number gets higher, so too does the pressure on the system to revert to the mean. Simply stated, the more it goes up, and even more important, the faster it goes there, the more it wants to go down.

There is a causal basis for this in weather. As temperatures and CO2 rise, plants have a longer growing season, and consume CO2 more rapidly. The facts are complicated in this case because it’s mostly sea dwelling plants that add O2 to the air and land plants that add it to the water, but on a large enough scale that’s neither here nor there. Both do better in average warmer temperatures and higher C02 concentrations on average. And that's the way you have to treat it if you’re talking about a planet wide phenomenon.

So the temps and CO2 rise, more plants grow, and that drives CO2 and temps down and O2 up.

It might be different locally. You could for instance have the temperatures rise so high in the gulf of Mexico, already known for it’s warmer water temperatures, that plant life can no longer survive there or that gives rise to things like red tides which have a negative effect on plant growth. But on a global scale, the falling amount of plant life in the Gulf of Mexico will be more than offset by the rising amount of plant life in the gulf of Alaska.

If you can read between the lines, you begin to see the outlines of 2 errors that inexperienced data scientists often make. Mistake one is extrapolation. Just because something deviates from the mean doesn’t mean it always will, like Mann specifically said it would at the time.

The second is a scaling problem which says that just because a feature of data is true on a small scale doesn’t mean it will be true on a larger scale. Larger scales include many more variables, and all need to be taken into account before you scale from small to large, and back. It’s for this reason that lower average black IQ’s don’t automatically mean that all black men are idiots. You can’t go from the specific to the average, and back. There is no real connection there. (This is arguably, the most common mistake of all liberal thinking.)

But Terrence Mann wasn’t an inexperienced data scientist. He was a serious (cough) academic thinker, with many letters after his name. So how was it possible that he made such catastrophically simple errors in his analysis of the data? It wasn’t logical to assume that he was actually a ‘Chauncey Garner’ like imbecile finding his way through academia spouting wise sounding allegories he never really understood. The only logical conclusion was that he was not an idiot.

And since that was so, he must be lying.

Why would he lie? The answer was elegantly reduced in the classic movie ‘The right stuff’, when the astronauts were arguing with the engineers for a more passenger friendly design of the Mercury capsule – “Do you know what makes this rocket take off? Money! No Bucks, no Buck Rogers!”

Whether Terrence Mann knew he was lying about the hockey stick is unknowable, at least by me. But I know he was lying, because headlines drive funding, and politically motivated funding is what drives the debate around ‘climate science’. His laughable hockey stick is only the worst example. There have been numerous other ‘errors’ which didn’t get the same headlines, but had the same effect.

Then there is the issue around the ‘solutions’ to it, like the Paris Climate accords.

Is the world warming? I don’t know. Probably. It certainly makes sense to me. Is man in part responsible? I don’t know. Probably. It certainly makes sense to me. (no one else does either… all they have is conjecture and notoriously unreliable predictions.)

Should does it then follow that we should do all we can, including shutting down a significant portion of our global economy and putting another portion of it into the hands of a group of people who have demonstrated their willingness to lie in order to gain political influence? No. Absolutely not. We would be stupid to do that. Thankfully, we aren’t as collectively stupid as Terrence Mann and the ‘climate scientist’ seem to think we are.

1 comment:

MikeCLT said...

The science/physics of global warming may be pretty good but the prediction models are ludicrous. Any policy proposals based on those models would be wrong. Scott Adams, who admits he has no math background, had some good posts about the problems with the prediction models from a persuasion standpoint. Basically that there are so many models that even a broken clock is right twice a day. This guy at The Reference Frame had a really good post about the problems with the models.

http://motls.blogspot.com/2017/03/selection-of-climate-model-survivors.html

Even my English major/law school background can understand this. I do not practice any more but work in a law related business that depends on technology. But I am shocked at the innumeracy and technological illiteracy among lawyers. I deal with lawyers who went to Ivy Leagues schools who are highly intelligent. Yet they cannot use Excel and its basic functions to manipulate reports we provide them. These are the people we have in Washington who are determining policy. Asking them to spot the errors you did is a bridge too far.