I had the great pleasure last night of meeting Melissa of Shakespeare’s Sister fame and some of the great cast of characters she has assembled over at her blog, including Mr. Shakes, Litbrit, Paul the Spud, and others. The occasion was a visit to our northern suburb of Evanston to catch Al Gore’s movie An Inconvenient Truth. In fact I had already seen the movie, but was more than willing to see it twice. I am quick to admit that I am not a Gore fan, and the thought of paying hard cash to see a movie that consists mostly of him giving a Keynote presentation (there was plenty of Apple product placement) falls somewhat below “drinks at Clooney’s villa in Tuscany with the gang” on my list of exciting ways to spend an evening.
But it turns out to be a great film, oddly compelling, with at least one priceless joke about gold bars. It’s not a science documentary — many graphs have no labels on their axes (much less error bars), and much of the evidence adduced is anecdotal and aimed at the gut rather than the brain. But what anecdotes they are. It’s hard to see pictures of Russian fishing boats stranded in a barren sandy landscape that once was a major lake bed without thinking that something needs to be done.
There isn’t any scientific controversy over whether or not climate change is happening, or whether or not human beings are a major cause of it. That argument is over; the only ones left on the other side are hired guns and crackpots. But the guns are hired by people with an awful lot of money, and they’re extremely successful at sowing doubt where there shouldn’t be any.
Their task is made easier by the fact that the atmosphere is a complicated place, and the inherent difficulties in modeling something as messy as our climate. But climate models are not the point. The point is not even the dramatic upward trend in atmospheric temperature in recent years. The actual point is made clear by the plot of atmospheric CO2 concentration as a function of time, which I just posted a couple of days ago but will happily keep posting until I save the planet.
Here is the point: We are taking an enormously complex, highly nonlinear, intricately interconnected system that we don’t fully understand and on which everything about our lives depends — the environment — and repeatedly whacking it with sledgehammers, in the form of atmospheric gasses of various sorts. Statements of the form “well, we don’t really know what that particular piece of the system does, so we can’t be rigorously certain that smashing it with a sledgehammer would necessarily be a bad thing” are, in some limited sense, perfectly true. They are also reckless and stupid. The fact that there are things we don’t understand about the environment isn’t a license to do whatever we like to it, it’s the best possible reason why we should be careful. And being careful won’t spell the doom of our economic system, bringing global capitalism crashing to the floor and returning us all to hunter-gatherer societies. We just have to take some straightforward steps to mimimize the damage we are doing, just as we very successfully did with atmospheric chloro-fluorocarbons to save the ozone layer. And the best way to ensure that those steps are taken is to elect leaders who are smart and determined enough to take them.
CapitalistImperialistPig,
Wolfgang may not agree with any of us on global warming, but I assure you that he knows more statistics than any of us.
Do enlighten us about Wolfgang’s ‘expertise’ in statistics(not how to lie with statistics)…
era,
what I know or do not know about statistics is none of your business.
If you suggest that I lie then I suggest that you i) indicate who you are instead of hiding behind a generic address and ii) perhaps point out what exactly I am lying about.
Perhaps I should help you make your point. My main statements are/were:
i) the climate forcing due to CO2 behaves like ln(C)
ii) there is no good evidence that CO2 increases exponentially.
The CO2 produced by human activity should give the strongest signal for recent
years, i.e. after WW2, but the most reliable data (Mauna Loa) does not show this.
By the way a proponent of global warming seems to agree with me
http://mustelid.blogspot.com/2006/01/co2-and-sres.html
Again, I am eager to hear where exactly I am wrong.
And neither the words “troll”, “crank”, “liar” count as strong argument in my book.
This picture shows the recent trend in 5 major greenhouse gases
http://en.wikipedia.org/wiki/Image:Major_greenhouse_gas_trends.png
CO2 can at least keep the linear trend up …
Wolfgang, it should be clear to you if you read the responses you got that we disagree on (ii). There is a lot of evidence that humans are responsible for the CO2 increase in the atmosphere from *1800-present* – and you can find links and detailed references starting here. I didn’t call you a liar, but to cherry pick data to fit your conclusion (as you did when you threw out everything before 1955) is dishonest and misleading. We only have your *assertion* that human impact on atmospheric CO2 is only visible in the Mauna Loa data and invisible in any previous data. This does not count as an argument in *anyone’s* book, and if you post again without references to back up this claim, you will confirm yourself as a troll.
But do you not find it a little bit odd that the data fits your assumption of exponential growth only when you include data before 1955; i.e. when the impact of human activity was the smallest?
The mixing of CO2 in the atmosphere occurs on a timescale of a few years only.
No you did not call me a “liar”, just a “troll” and it was ‘era’ who suggested that I was lying. Now you suggest that I am dishonest.
We cannot agree on the statistics of CO2, but perhaps we can at least agree on the statistics of personal attacks.
> We only have your *assertion* that human impact on atmospheric CO2 is only visible in the Mauna Loa data and invisible in any previous data.
Re-reading your text it is clear to me that you do not understand my argument;
It is exactly the other way around. If you want to see exponential increase (due to human activity, there is no exponential increase due to natural cause) you have to look at old data.
I really think I am wasting my time here; Thanks anyway.
Wolfgang,
The undisputed contribution from humans so far (50, according to you in #40), out of ‘natural maximum’ of 300 is HUGE considering the time scale involved (i.e., relative to geological time scales). The Taylor series explanation is apt: the above data looks even more benign if plotted against seconds instead of years, decades or centuries… Which is the proper time scale for this problem?
If it continues ‘linearly'(as you put it) so it is 600 in a couple of centuries, is it still no big deal? Well that is your own biased opinion (to say the least).
It is not just the Taylor expansion that hides the exponential growth of the CO2 concentration on short time scales, but mainly the error in the data, and consequently the fitted curve. That is why you need to look at this over a reasonably long time scale.
The CO2 concentration growth can suddenly have gone linear with time, but that would require a mechanism for why this sudden change has come about.
I actually just tried to fit just the Mauna Loa data that Wolfgang linked to, just for a check. You can already see the non-linearity in this small timescale. A linear fit is a poor fit.
Hiranya,
I really appreciate that we are back to science.
I did the exercise some time ago. A direct fit (linear vs. exponential) does not show anything significant, both fits give you about the same confidence.
If you want to test the hypothesis that the ML data shows exponential growth you should examine the first differences and check for exponential growth of those differences.
A simple stationarity test rejects this hypothesis as you can see in
this chart.
The black line includes a 2nd order term in t and curves the wrong way; the growth rate flattens instead of accellerating as it should.
You can *assume* exponential growth and fit the growth rate from the data, but the data itself does not indicate exponential growth.
Wolfgang, I didn’t make any exponential assumptions. I tested the hypothesis whether the data was described a *linear* function. Since we are being scientific, kindly provide the scientific reasons why you abandon data before 1955, and reasoned refutations for the human impact on the atmosphere pre-1955 that I cited before.
PS: And since you are ignoring the basic point which nano expressed very well (that at best you are postponing the inevitable a couple of generations) I won’t waste my time on this debate anymore.
> kindly provide the scientific reasons why you abandon data before 1955
because economic activity was lower by a factor of about 10 and thus the signal is less by the same factor and in addition you have to merge data of different quality and the merging itself introduces all kinds of artifacts.
If I want to estimate the growth rate of the internet I would mainly use data from the 90s until now and leave out the data from before 1950 as well 😎
> I won’t waste my time on this debate anymore.
This was my conlclusion yesterday and I think I will stick to it from now on.
Unfortunately I had no time to make a timely response.
Hiranya says:
“I was referring to the last 650,000 years in my comment (i.e. the era of the modern humans).”
And I was discussing natural occurences. Otherwise we are cherry-picking data.
“The conditions during the PETM does not instill me with a great deal of confidence that the capitalist system will take it in its stride :(”
You don’t know that, since a) the natural system recovered from much larger disturbances b) there was no capitalist system during PETM. 🙂 But see my comment about the “Venus effect”. 🙁
Torbjorn: the whole “scary” part of climate change is the rapidity of the “change”. Its not that the capitalist system has to adjust to the PETM conditions while the human race evolves over a stable period of order a million years. Its the chaos caused by a sudden transition from one stable climate configuration to another that I don’t think the capitalist system can handle very well. Note that sudden is in geological terms. I don’t think a 650,000-yr baseline is cherry picking data since (a) these are the conditions during which the human race evolved, and (b) going from memory it includes 5-6 periods of glaciation to get an idea of the natural cycle of climate change.
BTW since this is a science blog I thought I would say what I did to check the hypothesis that the ML data was fit well by a linear curve. Note that I am a theoretical cosmologist, so I don’t make any claims about how good the data are – the climate scientists who accept it are far more qualified to make that judgement than I am.
The data appear to show a pretty monotonic upward trend modulated by yearly oscillations. There don’t appear to be statistical errors given for the data points (at least I couldn’t find any). One can assume that the statistical errors are constant from measurement to measurement, and do a least-squares fit to the data of a linear function modulated by a sinusoid. Once you get the best fit parameters, you subtract the “best fit” model from the data. If your hypothesis is right, the residuals should look like white noise. In practice, this is not quite the case, since the oscillations are not completely described by the simple sinusoid so they leave a residual too. However this effect is very easy to identify and filter, as it has a particular frequency. Now you check the power spectrum of the residuals (e.g. by taking a fourier transform of it). You can see that the power spectrum does not resemble white noise. Therefore either the data are not described by a linear function, or else the statistical errors were not random. Since the data are accepted in the field, I conclude that the former is the case.
Of course if I could find the statistical errors there are other, more sophisticated tests I can perform, but this one works well to show the non-linearity of the data.
Tobjorn: Let me also add that I have very little concern about the natural system recovering from what we do to it. It has taken far worse blows in the past and recovered with new types of biodiversity after massive extinction events. Its just that I have some small concern about our own survival from the consequences of what we do to the natural system 😉
> the whole “scary” part of climate change is the rapidity of the “change”.
Torbjorn,
as I wrote above, the data suggests that the increase in CO2 is less than exponential (and flat to down for other greenhouse gases). The average increase from one year to the next grows less than linear (when it should grow exponentially for the really scary scenarios). This makes a huge difference, because it is important whether we have 10 years (as Al Gore suggests) or 100 years to change technology.
And there are indications that we will have to move away from oil over the next 100 years anyway.
#7 s.y.: I have been reading extensively on the relevant issues since watching this movie, and I found these criticisms addressed in detail at a non-technical level here and many other places on the same site, including the links to the refereed publications criticising the claims of Steve McIntyre, which discredit his methodology. Hope this is helpful!
Wolfgang, there is no evidence in the only dataset you find admissible that there is a “less than linear” trend. There is a “more than linear” trend. The beauty of science and having public data is that such claims are very easily checked.
I think you misinterpret what I wrote:
“The average increase from one year to the next grows less than linear.”
Just to make myself clear. In an exercise like this you do not fit C(t) but
d(t) = C(t) – C(t-1) as a function of t.
The first derivative of an exponential is still an exponential, but d(t) grows slower than a + bt.
Wolfgang, your other statement about the trends of other greenhouse gases does not seem to be correct either. I apologise for calling you a troll, I do now think you sincerely believe what you are saying. But anytime I think you post something misleading here and it’s within my ability to verify it, I am going to correct you, because I believe it is my duty as a scientist to the people who stop by to read this and have not made up their minds.
I agree with you that the precise rate of climate change has some uncertainty (as far as I can determine the behavior is within a power-law to an exponential). However this gives us only a time-frame between 10-100 years (as you seem to agree), and given the non-linear nature of the problem and its complexity, it is not at all clear when the “point of no return” is passed and one cannot prevent drastic climate change. I am not at all worried about the future of the planet, but I *am* worried about the future of the human race. For my part, I prefer to do something about it now, rather than have the blood and suffering of future generations on my hands.
PS: I understand perfectly well how to fit a time series 🙂
> the trends of other greenhouse gases does not seem to be correct
I refer to the picture in #54. The one who pointed the flattening out was James Annan who is a proponent of global warming.
I would like to clarify a few things by the way:
i) I am certainly not an expert on climate change, all I did was to take the available data, load it into R and play around with it. I think my exercise was correct, but the beauty is that everybody can do it for herself and does not need my judgement or yours.
I would really appreciate to see how a professional approaches this and in particular what models are used (e.g. about the feedback of the biosphere which seems to take advantage of the increased CO2 supply and thus slow its growth).
So far I have mostly seen papers which simply *assume* that the growth is exponential.
ii) I think it would be a big step forward if this debate would not be about global warming: yes vs. no or global warming: Gore vs. Bush
but if it would be possible to discuss more subtle issues, like how much and how fast, without running the risk of being insulted.