Polynomial Cointegration Tests of the Anthropogenic Theory of Global Warming
Michael Beenstock and Yaniv Reingewertz – Department of Economics, The Hebrew University, Mount Scopus, Israel.
Abstract:
We use statistical methods designed for nonstationary time series to test the anthropogenic theory of global warming (AGW). This theory predicts that an increase in atmospheric greenhouse gas concentrations increases global temperature permanently. Specifically, the methodology of polynomial cointegration is used to test AGW when global temperature and solar irradiance are stationary in 1st differences, whereas greenhouse gas forcings (CO2, CH4 and N2O) are stationary in 2nd differences.
We show that although greenhouse gas forcings share a common stochastic trend, this trend is empirically independent of the stochastic trend in temperature and solar irradiance. Therefore, greenhouse gas forcings, global temperature and solar irradiance are not polynomially cointegrated, and AGW is refuted. Although we reject AGW, we find that greenhouse gas forcings have a temporary effect on global temperature. Because the greenhouse effect is temporary rather than permanent, predictions of significant global warming in the 21st century by IPCC are not supported by the data.
Paper here (PDF)
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
I’m really amused by some of the comments regarding the data. The time series of cointegrated comments over the last 15 years sounds like this:
Warmists – this data shows that CO2 is heating up the planet
Skeptics – that data is all messed up
Warmists – is not!
Skeptics – is too!
New Paper – the data shows that CO2 is NOT heating up the planet
Skeptics – that data is all messed up
Warmists – that data is all messed up
Skeptics – yeah, that… huh? what?
I also think it would be very interesting to see this analysis done again, but with a better data set because I think the methodology itself makes sense.
JonesII (16:22:37) :
It´s the SUN….! PERIOD.
If it was truly the sun then it would only effect heat and cold. Not evaporation or precipiation that is increasing.
Joe
David L. Hagen (15:45:26)
_______________________________________
As you noted, the Beenstock and Reingewertz paper did not include water vapor in their calculations and you refer to the recent work by Miskolczi in which water vapor is a central issue. I would like to suggest that the Miskolczi model and the rationale underlying Willis Eisenbach’s Thermostat hypothesis may in fact have a lot in common. Miskolczi sees the water effect as stabilizing the IR layer to a constant value while Eisenbach thinks of cloud formation, which depends on water vapor, as the fundamental “governor” of global temperature. My own experience as a sailor in the Caribbean resonated very strongly with the effects described by Eisenbach. My belief is that the Earth is a huge heat engine about which know little. I feel that both Miskolczi and Eisenbach are on the right track to helping our understanding.
If I understand it corectly, it should no matter if aerosals, cloud behavior, land use, etc are factored into this analysis – the analysis looks for relationships between specific variables and the behavior of the system as a whole does not need to be understood to do this. What gets me is that I really thought someone would have done this already(?)
If not that means the GCMs were built off of theory-only, most likely with a heavy dose of tweaking to get their behavior to match the historical data. If this really is the case, I have given the RC crowd WAY too much credit…
My Econometrics is very rusty, so feel free to shoot me down if I’m mis-stating.
A couple of random thoughts in parting: 1.) lack of correlation in violation of established physical relationships (assuming, incorrectly of course, all other things being equal) could also point to the data being crap, 2.) since it hasn’t been mentioned, the quick feedback to neutralize CO2’s temperature could also be the cloud behavior proposed by Lindzen
Do you think perhaps your thesis should be totally reversed, for both ’cause and effect’ and the direction of effect, vis-a-vis:
“9/11 caused man-made global cooling ”
.
.
“Beenstock and Reingewertz paper did not include water vapor in their calculations”
As I understand the paper this is not very relevant/accurate. They matched the observed temperature to CO2 concentration and found a mis-match in trends from what should be expected if the AGW models were realistic.
It doesn’t matter what effect water vapour is having if the paper shows that climate doesn’t respond to CO2 in the way previously advertised. The paper is looking at the effect of CO2 in combination with everything else that is happening including water vapour effects.
George Turner
My point is not to defend any GCM models. I would be amazed if they could predict the future.
And who knows, I might find out that buried away in that paper they actually reviewed the results of an actual climate model. Or the derivative of a climate model.
It’s just seems like the Coleman report all over again – disproving the IPCC by proving something the IPCC agrees with.
Call me old-fashioned but I find it more enjoyable to see a discussion where the presenter actually knows what his opponents believe and demonstrates the flaw in those arguments.
I’m clearly in a minority though, but that’s ok.
“Pa Annoyed (16:46:44) :It is a capital error to theorize without data.
As someone working in an industrial laboratory, it is silly to collect data without some sort of theory. You must at least believe the data has some sort of meaning. I have learned a number of times that the data will often refute the theory. Actually this almost always turns into a great learning experience. It has happened so many times we often try not only what we think will improve the product but also the exact opposite.
I’ve got a new model. It is now up to me to see if there is some data on the height of the tropopause over the last 50 years. I would not even know to look for this without some type of theory.
The only mistake i could make would be to assume my model is true and then make all kinds of announcements about the consequences if it were true. This is what the climate modelers have done. Now we know that every single climate model that assumes constant relative humidity is wrong. Yet they have been scaring people with thermageddon while ignoring that their models have been proven wrong. Its time to go and tell all of the biologists that all of the papers that assumed dramatically increasing temperatures are pretty much useless. Oops.
I wouldn’t put any more faith in this paper than I would have if it pointed in some other direction. Sounds like mumbo jumbo to me. “If all the economists in the world were laid end to end they would still point in different directions.”
I think if the study is looked at from the context of emergence in complex systems – coherence of multiple cooperant forces – Beenstock and Reingewertz’s approach makes sense. Here’s my take on it…
Characteristics presented (or being observed) tend to be observable only while the cooperant forces remain in a stationary relationship or a time series relationship where the size scale and phase of the relationship remains reasonably the same. Because a stable system is dominated by negative feedbacks or regulators, it is highly unlikely that the relationship of the cooperant forces can remain the same for long periods of time if one or more cooperant forces are the product of disequilibrium.
According to the AGW theory, GHG change forces the system out of equilibrium, resulting in long term temperature rise. If AGW theory is true then GHG change must be cointegrated with temperature trend and solar irradiance (among other cooperant forces).
Beenstock and Reingewertz look primarily at greenhouse gas forcings, global temperature, and solar irradiance, in their tests for cointegration. Cointegration testing doesn’t require that all involved cooperant forces be identified, rather only several that can be quantified are needed.
The argument that some have made about water vapor not being included isn’t important to the outcome of the analysis. If water vapor is one of the cooperant forces (and it most surely is), then its affect doesn’t change the test for polynomial cointegration of GHG’s with temperature trend and solar irradiance. It changes only the size scale of the relationship.
If this study is correct then temperature rise is emergent and will occur only as long as it takes for the system to return to the old or a new equilibrium. That time period is dependent on feedback response times and the casual relation across the different scales, phase, and time series of the cooperant forces.
I wouldn’t put any more faith in this paper than I would have if it pointed in some other direction. Sounds like mumbo jumbo to me>
I think some people are nervous about these results or rejecting them outright because the analysis technique is so foreign to how we look at data, but this technique is not new. I first came across it 20+ years ago when I was introducing this revolutionary computer technology called a graphics terminal (yes! you can draw lines on the screen!) to researchers. One of my customers was a geophysicist who was trying to locate micro-seismic events in three dimensions from an array of sensors. I remember him showing me some of his data capture from the sensors on the screen. There was a squiggle and he’d show me the math to determine how far away and how deep (back then I could still do the math…. now not so much). Then another squiggle which he identified as a heavy truck going by on the nearby highway. Another squiggle was from heavy artillary at a military base a few miles away. Then a BIG squiggle. What was that? Using similar techniques he showed that the big squiggle was a seismic event, a truck going by, and an artillary strike all at the same time. He could figure out how much of the squiggle was seismic event and how much was “other stuff”.
One of the first big events he captured was too far away for him to calculate location, just distance, but we knew it was big. The next morning the quake that clobbered Mexico City was on the news, so about 1985 I think?
I’m sorry but statistics don’t replace logic, they substantiate or refute it. This just doesn’t sound logical as long as you accept the GHG theory. Or perhaps I just don’t have a clue what they are talking about.
It may seem strange having a couple economists using one of their statistic techniques on climate, but both sciences are about finding cause and effect in a massively chaotic system. Data-wise, I see more similarities than differences and I think economics is a little more mature on the statistics side…..
Does anyone know any other studies that can account for CO2 only having a temporary effect? Could that be attributable to the black carbon issue or is a 1 year half-life too long for that?
Who found this paper? It seems odd, if it really is to be published in Nature, that it could be released NOW, before publication. Most journals I know, especially Nature, place an embargo on when a paper can be publicized. This right now would seem to be a big no-no. Secondly, it looks like it is a pdf of a word file from 12/21/09. Awfully fast turnaround time to get a paper accepted, especially of this magnitude. If, in fact, it was submitted in December that makes it all the more incredible since trying to find anyone, much less willing reviewers, in December is like herding cats. I haven’t read the paper, and most likely won’t since math ain’t my bag, baby. Still, the provenance of this seems odd.
rbateman – I have some serious concerns with the spatial nature of both the surface temperature record and now the CO2 record. These cyclical anomalies do not seem to distribute randomly – should the surface temperature record be “gridded” as if natural temperature variation was a random function or a local characteristic (eg UHI).
I reckon that our Gav thinks that this is a load of crap.
Which, to my way of thinking, will be the proof of it.
ed432 (18:52:40) :
Sounds like mumbo jumbo to me.
To me too ! And the underlying assumption is that the data is good. We know that the sunspot numbers, TSI, and likely the temperature as well are not well constrained and that there is significant doubt on the long-term variation of all of them. In view of that, it is doubtful that statistical test have much meaning [either way].
Smokey (18:05:28) : Ice/snow cover, Northern Hemisphere, 2009 v 2010: click
Scary. Extrapolate that out 5 years.
jorgekafkazar (15:05:34) :
JDN (14:13:26) : “Where the hell do these guys get off using “nonstationary time series” and “methodology of polynomial cointegration”? … [snip]
Part of the risk of visiting science blogs is that you may run into terminology peculiar to areas of science of which you are partly or totally ignorant. Most people either look it up or let it go, instead of flaunting their ignorance in public.
Actually, I’m flaunting my disgust. I’m a scientist of various training with a heavy emphasis on mathematics and biology at this point, so, I’m accustomed to being ignorant of one subject or another. This paper is a case of people deliberately trying to obfuscate their work. You see it all the time in mathematics. At this point, if people refuse to write clearly and try to disguise their methods, I’m not interested in looking it up. They need to summarize their independent variables, justify their choices of variables (e.x. they introduce rfCO2 without explanation… it could be anything), explain why they are using linearized equations for those variables instead of raw data, provide some explanation on how their statistical method (something not useful for establishing causality) overturns predictions that have explicit causality (AGW calculations for all their flaws still pretend to be causal), etc., etc., & etc. As far as I’m concerned, this paper reads like the Sokal hoax.
rbateman (14:52:50) :
Yes. BRAVO.
I hadn’t known, before reading this paper, that it is already known that CO2 is I(2) and temperature is I(1). In economics, a finding like that that is normally taken to mean “Game Over” for any theory that says a permanent increase in the level of CO2 will cause a permanent increase in temperature. The theory has failed at the very first empirical hurdle. I am surprised that this result has not been more widely publicised. It’s a big problem for the CO2 theory.
(What Beenstock and Reingewertz do is try to help the CO2 theory jump over that first hurdle, by getting some help from solar irradiance. Even with help, the CO2 theory fails to jump the first hurdle. But solar irradiance can jump the first hurdle.)
Summary of the paper for non-economists:
Temperature has increased linearly. CO2 has increased exponentially. If CO2 would have driven temperature, it would have increased exponentially too. It has not. Therefore, CO2 does not drive temperature.
That’s what Beenstock and Reingewertz say.
I disagree, because they do not account for the fact that the ocean warms only very slowly.
mikep (15:08:23) :
http://www.ulrich-fritsche.net/Material/murray1994.pdf
Thank you for this link!!
This should be read (short, clear, sweet) before tackling the subject of the post.
JDN (14:13:26)
Ah, yes, the famous “I can’t understand it so it must be wrong” argument …
Richard Tol: “Temperature has increased linearly. CO2 has increased exponentially. If CO2 would have driven temperature, it would have increased exponentially too.”
That sounds a dubious statement. CO2 is subject to saturation effects. Therefore temperature should increase at logarithmic function of CO2. If CO2 is increasing exponentially that would make temperature linear as found?