Thursday, May 23, 2013

Does global warming cause tornadoes?

It was sort of inevitable that the deadly tornadoes in Oklahoma would ultimately be blamed on global warming and CO2 by someone. While most people – including those alarmed by "climate change" – reject this attribution, you can find pretty powerful people who promote this incredible link.

Senator Barbara Boxer was perhaps the most powerful person who enthusiastically supported the idea that the tornado outbreak was a message from Nature telling us to introduce new carbon taxes. She really sounds religious.

You may find lots of stories in the media that discuss a possible connection between tornadoes and the enhanced greenhouse effect. Thankfully, almost all of them (e.g. NY Daily News, Washington Examiner) say that there's no connection. But Barbara Boxer knows that such a connection would strengthen the case for the new taxes – so it must be a part of the consensus, right?

Without actually thinking about the science or asking researchers, leftwingers generally assume that whatever is convenient for their "cause" must be a part of the "scientific consensus".

But let us look at another magazine, National Geographic. It concludes by saying that a hypothetical influence of "global warming" on the frequency and strength of tornadoes could go in both ways and there is no evidence of a trend in either direction. They quote Roger Pielke Jr, among others. Still, the title asks whether there is a connection and it offers some ideas that could support such a link.

Which ideas?

Before they admit that it's always tricky to link a particular weather event to climate change, they offer these two paragraphs:
It sounds intuitive: Of course global warming should lead to more—and more powerful—tornadoes.

We're adding energy to the atmosphere by trapping heat with greenhouse gases, and tornadoes are the very picture of terrifying atmospheric energy.
What I find problematic is that it is not explained why these would-be arguments in favor of the connection are wrong. So many readers may just think that they're actual valid arguments in favor of the connection.

The quote above says that tornadoes are "the very picture of terrifying atmospheric energy" and we are adding it, so we are probably strengthening tornadoes. Is that a valid reasoning?

All forms of energy are convertible to each other – only the total energy is conserved. This is the statement known as the first law of thermodynamics. So adding energy in one form may increase the energy in other forms, too. However, there is also the second law of thermodynamics that says that you can't construct a device that does mechanical work by extracting the thermal energy (heat) from a colder object (of from an object indefinitely). In other words, the perpetual motion machine of the second kind is impossible.

(The perpetual motion of the first kind is a gadget that would create energy out of nothing and it is also impossible – because of the conservation of energy or, equivalently, the first law of thermodynamics.)

But the creation of a tornado out of the "global warming" would be exactly such a perpetual motion machine of the second kind! The reason is that whether we like it or not, a tornado is a place with concentrated "useful" mechanical energy – stored in the macroscopic ensembles of molecules – while a warmer atmosphere is the "useless" thermal energy i.e. energy stored in the chaotic motion of individual atoms. And one simply can't create the former (useful, concentrated energy) out of the latter (useless, chaotic energy).

Once the energy is converted to heat, i.e. universal chaos, it can't be converted to the "useful" mechanical forms anymore! Only if the temperature is non-uniform, the temperature differences may be exploited to do some mechanical work, like in "heat engines". But global warming creates no new temperature differences so it creates no new opportunity for such "natural heat engines" to produce new mechanical work!

Why is it so? Because the global warming is global, stupid. The widely discussed effect is caused by the increased CO2 concentration which adds something to the expected temperature at any place. But because the carbon dioxide molecules are almost instantly (within weeks at most) spread uniformly over the whole Earth's atmosphere, the greenhouse effect is equally strong everywhere. Well, there is some dependence on the latitude etc., and especially the feedbacks (e.g. the ice-albedo feedback) depend on the latitude, but this dependence is extremely slow. To extract a useful mechanical energy e.g. to create a tornado, the greenhouse effect would have to seriously change from one mile to another and it is clearly something that the uniform, "global" rise of the CO2 concentration isn't capable of doing.

In other words, the enhanced greenhouse effect doesn't change the magnitude of local non-uniformities, so it can't contribute to the tornadoes, either. The same comment is true not only for tornadoes but for any "special", localized weather event. These events don't really care about the overall temperature shifts and the overall temperature shifts (probably very modest ones) are the only thing that extra greenhouse gases may cause.

So at least in the leading approximation, CO2 isn't capable of changing the frequency of tornadoes – in either way. But you could hypothesize that there is some subtler, higher-order effect that causes such an influence, anyway. For example, the thickness of the troposphere (the circulating part of the atmosphere between the surface and the tropopause, about 10-20 km above the surface, where all the "complicated weather" takes place) depends on the overall strength of the greenhouse effect on the Earth.

It's plausible that as the troposphere gets thicker, there is more room for circulation and various manifestations of the circulation may strengthen. That's great but we should be interested not only about the Yes/No answer to the question "whether such an influence may exist" – in principle, almost everything influences everything else – but also about the estimated strength of such an influence. These estimates will be order-of-magnitude estimates of a sort; we won't discuss whether the effects boil down to wind shear, jet streams, funnel clouds, supercells, or other fancy concepts in atmospheric physics.

We should ask: by how many percent the frequency of tornadoes could rise or decrease (we really don't know the sign) if CO2 raised the global temperature by 1 °C?

That's a meaningful question, suggesting that we want to understand these things beyond demagogic slogans. So let's try to think a little bit. I proposed a mechanism suggesting that the strength or frequency of tornadoes could scale with (a power of) the thickness of the troposphere or, almost equivalently, with the total temperature increase caused by the greenhouse effect on Earth.

So how many percent may this unknown influence add? The key realization is that the greenhouse effect caused by the man-made CO2 is a small fraction of the overall greenhouse effect on Earth. The greenhouse effect on Earth is dominated by water vapor which adds over 30 °C to the temperature of pretty much every place on our blue, not green planet. Even if the whole 20th century warming was due to CO2 (which seems unlikely to me), we have only added about 2 percent to the overall warming caused by the greenhouse gases. We may add another percent, perhaps, but we're still talking about a few percent.

The thickness of the troposphere may have risen as a power law which also means by "several percent" and the same thing holds for all quantities describing the wind speeds, the total number of vortices, and so on, and so on. I actually think that this is an overestimate and because the troposphere is getting thicker, the atmospheric phenomena and energy are actually diluted into a larger volume so less is left for the near-surface weather. But let's be agnostic about the sign, exponents, and coefficients.

It's still true that these unknown influences only add or subtract several percent to or from the strength and frequency of the tornadoes we ultimately observe. And this is such a tiny change that it's ultimately undetectable. Why is it undetectable?

Because the number of tornadoes is notoriously variable. The interannual noise is so large that it totally prevents us from seeing a hypothetical signal that would only scale as a few percent. For example, look at this list of tornadoes in the U.S. spawn by tropical cyclones. The records (with lots of holes) begin around 1811. You see a few tornadoes a year, sometimes a dozen or two. There are exceptions like 115+ tornadoes in 1967 and 103+117 in 2004. It's pretty clear that in the past, the actual numbers were larger but people couldn't see everything due to the limitations of their observational technology.

At any rate, if the frequency of tornadoes grew by 2%, it would mean that there should have been just 113+ tornadoes and not 115+ tornadoes in 1967. Similarly for other numbers. The years for which the number of tornadoes are of order one are shifted by a tiny fraction of a single tornado. Try to statistically evaluate these chaotic tables in any way. It is absolutely clear that you couldn't possibly see a trend, whether it is a positive one or a negative one.

I believe it is a sufficient reason for us not to talk about such influences. Given the fact that we can't observe such an influence empirically, every proposed claim about such an influence has to suspected to be an artifact of errors, neglected terms and effects, and other things. We just don't know how large the influence is and what its sign is. Within the error margins, we observe the influence to be zero. So we should always act as if we were assuming that the influence is zero. Anything else amounts to bias – and violations of the presumption of innocence and other things.

Incidentally, I wrote about "noise" in the number of tornadoes on a given year. But the word "noise" is actually too disrespectful because what we really meant is "everything that has nothing to do with CO2". However, there could be – and there almost certainly are – many much stronger and signal-like influences on the frequency of tornadoes than the CO2 concentration. For this reason, it's already tendentious to pretend that the data are composed of a "signal" and of "noise", especially if we want to implicitly claim that CO2 is the only "signal" although it's almost certainly not the case.

Please, Ms Barbara Boxer and others, stop talking about these medieval hypothetical links that sound almost identical as the accusations against the witches in Salem, Massachusetts. Science supports none of your fantasies and every attempt by a person to rationalize such fantasies shows that the person lacks scientific integrity – and sometimes plain human honesty, too.

And that's the memo.


  1. I would think that increased co2 would increase convection by some tiny amount that is probably beyond the power of human calculation, but which may be significant enough to spawn a few extra tornadoes. I wouldn't be that surprised if windiness in general were to increase due to CO2, but yeesh, these kinds of deadly storms are not a new thing. What is infuriating is pretending that they are, and the denigration of those of us old enough to remember similar incidents, such as Xenia, OH, as out of touch old farts.

    This goes back to the fact that our culture is basically Christian and the myth of being expelled from the Garden of Eden is still very strong, even if the belief in the literal reality of the characters involved has faded.

  2. Barbara Boxer is a lunatic whose only saving grace is that she is less adept at raiding the US Treasury for her own personal gain than her sister California Senator Diane Feistein.

  3. I guess that global warming can't increase convection if it warms the air equally in the whole column of air. So, I guess your argument about the warming being "global" applies to this case as well.

    It would have to create some kind of mechanism that collects heat in one area, and not another.

  4. What? Carbon tax? Sorry, too late for that. And no, global warming is not *causing* tornadoes, nor any other extreme weather event, for that matter.

  5. Boxer is a ridiculous woman. Everything will be a sign of AGW for her -- snow in summer, hot weather in summer and extreme weather (such a vague emotional term). How did this idiot get elected to US Senate?

  6. Barbara Boxer is proof that stupid people should not breed.

  7. I'm not at all confident that the Communists weren't smarter than people like Boxer.

    Marxism was held to be scientific in the sense that, unlike Utopian Socialism, it was a materialist theory, not a set of moral ideals. This has little to do with whether the Communists were smarter or dumber than people like Boxer.

  8. The same way other idiots get elected to the U.S. Senate (not saying all U.S. Senators are idiots).

    Not only government officials, but the whole world, on average, is stupider than people imagine (although government officials are probably extra stupid). I'm not confident that the world could function if it faced up to how stupid it is.

  9. The convergence of cold air from Canada caused the tornado. No more. No less.

  10. NPR reporters are sounding more and more like Sheep Dogs whose job is to get anyone reporting weather to convert to the Global Warming Religion. They are shameless - they show up whenever there is a public appearance, and pressure the guy. But these 'reporters' actually sound like Valley Girls.