In a previous post, I stated that the current trend in global temperature increase is sufficient that by 2015, data will probably establish that the planet’s temperature has definitely not stabilized or begun to decline. Specifically:
By 2015, the expected temperature from the regression-line fit and that expected from the “no change” hypothesis will be far enough apart that we’ll probably be able to distinguish between them with statistical significance. In other words, by 2015 either we’ll know that global warming has changed (possibly stopping, possibly reversing), or there’ll be no more of this “global warming stopped in 1998” malarkey.It’s entirely possible that the numbers may give us statistically significant evidence even before 2015. If so, I’ll report the result. If it turns out that global warming is not continuing (which I seriously doubt), then I’ll readily admit that I was wrong. In fact, I’ll be keeping a close eye on the future evolution of global temperature and actively looking for such results, so if we do get valid evidence that global warming has stopped, I just might be the *first* one to say so.
If 2015 rolls around, and temperature have [sic] risen above present-day levels by enough to be demonstrably significant, I’ll announce that too. Will those who have so often chanted the “no more global warming” mantra admit that they were wrong? Somehow, I doubt it. I suspect that instead, they’ll be flooding blogs, newspapers, magazines, and Faux News reports with claims that “global warming stopped in 2013.”
I didn’t intend this as a “challenge,” but the idea was loosely based on various proposals I’ve seen for “bets” about future global average temperature. The “challenge” aspect has taken on a life of its own among readership; therefore I’m willing to make it official. I will also reiterate that the divergence between warming and no-more-warming isn’t married to the year 2015! That was a choice of a future year in which it’s likely that the issue will be statistically distinguishable, but a significant result might be available before then, or not until after. Also, the choice was based on intuition, not on any quantitative analysis. Eventually a significant difference will emerge; if not by 2015, then not long after.
I’ll also emphasize that I’m not interested in betting money on it. I have a family to provide for, and I can’t afford to have my money tied up in escrow while waiting years for a bet to be settled. Besides, I don’t gamble. Although if I did …
So I’ll outline more precisely what terms I would suggest for a wager, challenge, or whatever you like to call it, on the question “Is global warming continuing or has it ceased?” I’ll try my best to be fair to both sides, so that if you firmly believe that global temperature will continue to rise and you’re eager for a wager, I suggest this is the one to make, and likewise, if you’re firmly convinced that global temperature has peaked and is not going to continue and you’re eager for a wager, this is the one to take. The winner will definitely be decided by the reality or not of continued warming, not by any clever design of the terms and conditions of the wager. In my opinion, settling such a challenge should be based on statistical significance, not on choosing a specific year, so this proposal is based on statistical significance rising above the noise level, not on the temperature at a fixed future time (but as we’ll see, there is a time limit to how long the bet can last).
First let’s review the data leading up to the statement. Here are global average temperature estimates, all set to the same zero point (using the reference period 1950.0 to 1980.0), from NASA GISS, NCDC, and HadCRU:
The trend lines are determined from the data covering the time span 1975-2000. The graph is intended to show that the data after 2000 are not inconsistent with the claim that the trend is continuing, in fact they’re following the line with “wiggles” (i.e., noise) that make trends impossible to identify over short time periods but clear over longer time periods. (and indeed that is so). For the terms of this wager, it is not necessary to recompute the data using the 1950.0-1980.0 reference period I’ve used in this graph. This graph just gives us the essential idea behind it.
And the idea is this: if global warming is continuing, global temperature will continue to follow a rising trend plus noise. If global warming has ceased, it will stay at its present level (or decline) plus noise. So we should outline what global temperature will be in those two cases.
First let’s look at annual average temperature. I used the trend from 1975 to the present to estimate the trend, and used the standard deviation from the residuals (after subtracting the trend from the data) to estimate the noise level. The trend is upward at 0.018173 deg.C/yr, and the standard deviation of the residuals is 0.0959 deg.C. Here in fact are the annual averages (black dots), together with the trend (solid red line), and (dashed) lines two standard deviations above and below that trend line:
This gives the expected range of annual averages — between the dashed red lines — and 95% of all years should fall within those lines. If one wishes to be precise, these limits should be modified to account for the red-noise character of the data, but in this case it’s a small correction and I’m going to ignore it. Note that all the annual averages from 1975 to the present fall within the dashed red lines. As an aside, the above graph is about as clear a graph as I’ve seen showing that there’s really no evidence — none whatsoever — that global warming has stopped.
We can of course extend those lines into the future. We can also quantify the hypothesis that global temperature hasn’t changed since 2001; the average from 2001 to the present is 0.5432 deg.C, so we can simply draw a line at that value and dashed lines two standard deviations above and below it. Putting the “no-more-warming” range in blue, we get this:
If the “continued warming” hypothesis is correct, future values should fall between the dashed red lines. If the “no more warming” hypothesis is correct, future values should fall between the dashed blue lines. If the earth has actually started cooling, future values will eventually dip below the blue lines.
So here’s the bet based on annual averages: the still-warming side wins if temperature goes above the top dashed blue line; the not-warming side wins if temperature goes below the bottom dashed red line.
If temperature rises above the upper dashed red line, we have evidence that the planet is warming even faster than the present trend. In that case the still-warming side also wins. Alternatively, if temperature falls below the lower dashed blue line, we have evidence that the planet is actually cooling, and the not-warming side wins.
Finally, I’ll add one last condition. It’s unlikely but possible that a value can fall outside either range just because of noise. So, my “bet” is that as soon as there are two years (not necessarily consecutive) which are in either decisive region, the side with two decisive years is declared the winner. Therefore:
If annual average global temperature anomaly (land+ocean) from GISS exceeds 0.735 deg.C for two (not necessarily consecutive) years before it falls below the value(where t is the year) for two (not necessarily consecutive) years, then the still-warming side wins; if it falls below the above equation for two years before it rises above 0.735 for two years, then the not-warming side wins.
By the end of 2015, it is in fact likely but by no means certain that one or the other side will have won. Eventually, the two regions get far enough apart that it’s certain to happen. In fact, by 2028 we’re sure to have two years outside the limits of one or the other side, so the bet can’t take longer than 2028 to be decided. But this test isn’t based on a particular future year; it’s possible (but highly unlikely) that either side could win if 2008 and 2009 both fall into its winning region.
Although it’s unlikely, it’s possible that this bet could be undecided until the end of 2028. This is because the noise level is very high compared to the signal level; the noise level is about 5 times as large as the present annual trend! We can reduce the noise level without affecting the trend rate by using, not annual averages, but 5-year averages. That gives us a graph like this:
It’s straightforward to modify the terms of the test in order to base it on 5-year averages rather than annual averages. It’s also straightforward to adopt the test method to the use of HadCRU data, or NCDC, rather that NASA GISS.
I’ve seen other proposals for wagers, some of which strike me as perhaps unfair, having what seem like overly complicated conditions which may be designed to take advantage of statistical naivete as much as depend on the future progress of global temperature. On the other hand, some seem like fair but poorly chosen (too much chance of a false result due to random noise). If any part of this proposal favors one side over the other for purely statistical rather than climate reasons, I swear that it’s an oversight, not intentional. This proposed test is designed to be a fair test of competing ideas, and to be settled by a genuinely significant result, not by accidental changes due to “noise” in the climate system.
A final note: in further reader comments it was pointed out, quite correctly, that even if AGW is completely correct it’s still possible for temperature to show no increase long enough for the “no-further-warming” side to win this wager, IF unexpected events happen to alter the behavior of the climate. For example, large volcanic eruptions will cover the world with aerosols which will lead to significant cooling (such as seen after the Mt. Pinatubo and el Chicon eruptions) even if AGW is completely correct and uninterrupted; a series of large eruptions in succession may cause enough cooling to put future temperatures into the “no-warming” region. Likewise if sulfate aerosols from the booming economies of India and China get so great as to overwhelm the warming influence of greenhouse gases. I leave it to those who have money to bet, to estimate the probability of such things happening, and what additional conditions to impose to account for such a possibility. As for me, I suspect (even though I haven’t estimated the probabilities!) that it’s unlikely enough, that I’d still take this bet (for continued warming) without additional caveats. Of course that’s easy for me to say, since I’m not a gambling man.
If, however many years from now, the no-more-warming side wins the bet, and no unequivocal caveats are identified, then I’ll admit that our understanding of global climate is insufficient and that we can’t rely on the prognostications of the climate science community. I doubt it’ll happen. If, on the other hand, the still-warming side wins the bet … what will be the response from the skeptic side?
39 responses so far ↓
chriscolose // January 31, 2008 at 3:28 am
skeptics have it made…no warming = AGW has stopped (or never was), warming = “something else is causing it.” You’ll never convince them
CO2 should be up around 400 ppmv then. I haven’t looked into this much, but I would guess that with thermal lag (and taking into account just CO2), we’d be at the temperature response of around a 360-370 ppmv like atmosphere, which would be ΔT of ~ 1 K. Barring any unpredictable cooling events like volcanoes, the trend will clearly continue as it has.
I think that you’ll start being able to take into account a decline in aerosol effects rather soon, and if so that would be the equivalent of letting the other GHG’s (e.g. methane) “show up” since the net RF and CO2 RF are about the same. Without a substantial cooling mechanism, and I do not see a plausible way to not get an increased trend, unless negative feedbacks are much bigger than we think.
– C
Greg Simpson // January 31, 2008 at 3:51 am
To reduce the chance for a false positive, perhaps due to volcanic aerosols or a temporary increase in solar output, the requirement could be changed to two non-consecutive years. Note that a period of three consecutive years always includes two non-consecutive years, so a three years outside the bounds would always win the bet.
Aaron Lewis // January 31, 2008 at 4:32 am
Global warming is about heat. Most of the heat ends up in the water. Why are so many of these discussions focused on air temperature ?
(PS very good work)
Hank Roberts // January 31, 2008 at 4:53 am
Now, about ocean pH, that should be done about the same way, right?
Raven // January 31, 2008 at 6:03 am
Interesting.
It has been my position that the temps in 2015 would have to be at least 0.2 degC higher in 2015 to validate the AGW position. This was a crude calculation based on the CO2 sensitivity ranges presented in the IPCC report.
Your graphs come up with that same spread for different reasons.
That said, I don’t trust the GISS and HadCRUT datasets for the same reason I would not trust the unaudited financial statements produced by the Enron. I realize the satellite measurements have their own issues but there are two competing groups using the same dataset which helps ensure that self-serving data manipulation is kept to a minimum.
Would you take the bet on the average of UAH and RSS or is it limited to the GISS/HadCRUT?
Instead of betting money would you be willing to publicly acknowledge that AGW alarmists got the science wrong if you lost? Would you be willing to publicly apologize to skeptics who you have denigrated?
If not what would it take for you to do that?
The way I see it bets for money are a red herring since most people are prudent and would never bet more than they could afford to lose even if they were 95% sure about the outcome. Instead of fooling around with bets for money you should state what it will take to change your mind.
[Response: Your implication that GISS or HadCRU is guilty of “self-serving data manipulation” is mean-spirited, offensive, and unsupported by any evidence. Unless you can offer EVIDENCE with documentation to back it up, don’t repeat it here.
Your questions make me wonder whether you actually read the post. What part of “I’ll also emphasize that I’m not interested in betting money on it” is unclear? What part of “If, however many years from now, the no-more-warming side wins the bet, and no unequivocal caveats are identified, then I’ll admit that our understanding of global climate is insufficient and that we can’t rely on the prognostications of the climate science community” is unclear?
Shame on you.]
cce // January 31, 2008 at 6:53 am
GISS, Hadley/CRU, and NCDC are three competing groups using largely the same dataset.
And to repeat tamino’s last paragraph, “If, however many years from now, the no-more-warming side wins the bet, and no unequivocal caveats are identified, then I’ll admit that our understanding of global climate is insufficient and that we can’t rely on the prognostications of the climate science community. I doubt it’ll happen. If, on the other hand, the still-warming side wins the bet … what will be the response from the skeptic side?”
Raven // January 31, 2008 at 7:04 am
Yes I did misread your post about not wanting to make bets for money. I apologize for that. I had mixed you up with James Annan who has frequently talked about betting money.
For my I part I would also concede that the AGW point view is most likely correct if the warming trend continued into the ranges you identify. I feel your targets are a fair representation of the two possibilities.
My suspicions of the GISS and HadCRUT datasets comes from a general suspicion of any situation where there is a conflict on interest. One could argue that most executives would not deliberately manipulate their books even if they did not get audited. However, I would never invest money in company that did not allow its books to be audited by third parties.
A lot of money is riding on the temperature data so feel there is no excuse for allowing the ‘perception of possible bias’ to go on. The fact that many resist acknowledging the potential for bias simply re-enforces my view that the data should not be trusted unless it is audited by third parties.
[Response: All the data used by GISS can be downloaded from the web; the procedures they use are documented in the peer-reviewd literature; even the code for their computer programs is freely available. Their books are “open” and have been subjected to intense scrutiny.]
Raven // January 31, 2008 at 7:28 am
US taxpayers a fair amount of money for the GISS data to be produced. Expecting volunteers to replicate this work for free is not enough. Also volunteers have attempted to use the computer programs that were made available but were forced to give up because of poor documentation and OS/compiler problems.
You cannot say it has been audited unless proper funding has been provided to people who sole objective is to identify problems and ensure they get corrected.
More importantly, there have been a number of people producing analyses that suggest that the data is quite biased yet these criticisms are ignored (Anthony Watts, Ross McKitrick, Roger Peike Sr). You cannot claim that the GISS data has been subject to intense scrutiny if legitimate criticisms are regularly dismissed by the gatekeepers.
[Response: What a load of crap. Watts, McKitrick, and Pielke have generated ZERO real evidence — just as you have zero evidence of any misconduct — but they’ve slung a lot of unfounded insults — just like you have.
It seems you’re one of those who, no matter how closely the data and results are examined, will just invent yet another reason to claim it’s not enough. Keep moving the goalposts.]
Raven // January 31, 2008 at 7:52 am
It does not make a difference if you think they have zero evidence. What matters is whether their criticisms have been dealt with reasonably. Your response is typical and demonstrates that they are not being dealt with reasonably.
Right now the keepers of the data are free to arbitrarily dismiss any criticism. This state of affairs is unacceptable. Governments should take control of the data aways from the agencies developing the models. The conflict of interest is huge and would not be tolerated in any other field.
[Response: It doesn’t matter that they have zero evidence of any wrongdoing or mistake? It doesn’t matter that after years of trying to discredit the surface record, they’ve managed zip? What dream-world do you inhabit?
It would appear that from your point of view, it’s the *truth* that doesn’t matter.]
cce // January 31, 2008 at 8:01 am
It’s also worth pointing out, that the Surfacestations.org project documenting weather stations in the US, to date, has produced data and trends from the “highest quality” stations that almost exactly track the GISS analysis for the lower 48 states.
This is likely (>66% probability) the reason why discussion of the USHCN and the UHI effect has virtually disappeared in recent months. They were not “forced to give up.” They didn’t like the answer so they moved on to the next trumped up controvesry.
You can download JohnV’s “opentemp” program and run it yourself.
http://www.opentemp.org/main
dhogaza // January 31, 2008 at 9:21 am
Uhhh Uhhh grunt grunt oh my aching back ugghhh ugghhh …
That’s the sound of Raven moving goalposts.
First it’s “if the data’s not available for auditing, it can’t be trusted, as the three groups working with it all have a conflict of interest”.
Then it’s “oh, shit, the data and code’s all publicly available, so, ummm, now what, oh yeah, expecting volunteers …”
There’s plenty of money available on the denialist side from the fossil fuel industry to finance a legitimate alternative to the mainstream surface temp computations.
Why don’t they spend money doing that, rather than, say, have the heartland institute finance a shill conference (with paid speakers and free transportation and hotels for politicians), as they’ve recently announced they’re doing?
Could it be that they know that the surface temp record is sound? Could it be that they know that Christy and Spencer took their best shot with the first UAH satellite temp reconstructions and rather than knock down the surface temp record, only exposed themselves as being shoddy scientists who couldn’t even get their algebra right?
dhogaza // January 31, 2008 at 9:23 am
And before Raven suggests that the fossil fuel industry doesn’t finance legit debunking of the surface temp record because “AGW believers” have a lock on the literature …
They could easily get such a paper published in their own shill journal, Energy and the Environment.
Bruno De Wolf // January 31, 2008 at 10:12 am
As an (amateur) climate skeptic: I applaud the idea. I think both sides lack producing claims which are falsifiable in a decent timeframe. So, yes, I’m almost in.
One thing I don’t agree with in your method: GISS as the only measurement. As you know, GISS, and measurements by surface stations in general, is highly contested (correctly or not, I leave this out of this debate) among skeptics. Ideally, you should combine 1 metric based on surfuce station and 1 on based on satellite measurements (UAH or RSS).
For example: if a year falls outside the band you’ve drawn for both metrics, it’s a point. In this case: I think 2 years out-of-band should be enough to name a winner. There are other possibilities too, but I think you get the idea of involving satellite measurements as well.
[Response: All the brouhaha to contest GISS, and surface measurements in general, has generated no reason to have less confidence in its correctness. Quite the opposite, it has made confidence in the surface record far greater. But it has made effective propaganda points for those who simply refuse to accept the truth.]
JCH // January 31, 2008 at 1:20 pm
“would not trust the unaudited financial statements produced by the Enron. …”
The financial statements produced by Enron were audited. Every one of them. The auditing firm was one of the most respected in the industry.
Every single financial statement produced by the savings and loans industry were audited. Just from memory, but I do not think a single going concern opinion was issued in an industry that sustained pervasive collapse.
Auditing has a purpose, and it also has severe limits. I’ve been around a lot of auditors. My wife controls auditors. They cannot do what you seem to think they can do, and I can recite tons of proof. Auditors f up about as often as Barney Fife; as in, every show.
Don Fontaine // January 31, 2008 at 2:22 pm
Great post.
A minor observation. In the 5 yr average graph the last data point appears to be plotted at 2007, but all prior points appear to be plotted at the midpoint of the five year period eg 2002.5, 1997.5, etc. so the spacing for the last point isn’t the same as that for the earlier ones. Is this as you intended?
[Response: The final 5-year period (2005-2010) is incomplete, so the average temperature has been plotted at the average *time* of the data so far.]
Phil. // January 31, 2008 at 2:32 pm
Re GISS reliability I’d like to hear from Raven how GISS able to ‘fix’ their data while maintaining an approximately constant offset from the satellite data?
http://bp0.blogger.com/_0HiXKAFhRJ4/R5gl9hMkqcI/AAAAAAAAAgU/8Z6sJBs_XnU/s1600-h/Variation.JPG
Bruno De Wolf // January 31, 2008 at 3:36 pm
@tamino
Is it about ‘trying to do your best to be fair to both sides’ or about your opinion? I’m willing to take it as official as you do (I’ll do a posting and a follow up on my blog as well, no money involved, I’ll change my opinion if I lose), but I do insist in using more than 1 metric to make the conclusion. NONE of the popular metrics are absolutely 100% correct, RSS did some corrections to their historical data this month, GISS found out their Y2K error in the summer 2007 …
My comment was not about bashing GISS, it’s about having faith in 1 metric, so I reiterate my request to base your results on the combination of 1 surface station metric and 1 satellite metric. Is it that unreasonable?
[Response: None of the metrics — popular or not — is 100% correct. And correcting the GISS Y2K error led to a net change in global average temperature anomaly of 0.003 deg.C.
As I said, I’m not betting money I’m trying to establish conditions under which we can confirm or deny various hypotheses. It was framed as a bet because that seems to be popular for discussion, and it does force one to be explicit about exactly what conditions will lead to a declaration for one or another hypothesis. For a bet, I think it’s better to keep it simple and agree on a single source of data for decision.
But for determining the outcome with highest reliability it’s better to use multiple data sets. I intend to keep track of GISS, HadCRU, and NCDC, and I’ll probably keep my eye on satellite data from RSS, UAH, UMd, and UW as well. I’ll report any significant results, regardless of the nature of the result or the source of the data. I expect they’ll end up telling the same story.]
Hank Roberts // January 31, 2008 at 3:51 pm
Tamino, have you looked at the Hadley Centre’s ten year paper? It’s based, as I understand it, on the new flood of data coming from the Argo system and indicates a ten-year stretch with a lot more rearrangement than trend, followed by a return to the upward trend.
I hope you’re not setting up a betting range that would give a false negative if Hadley is correct about the coming decade — hoping you took that into account. The big decadal fluctuations do need to be modeled, and Hadley as far as I know is the first group to predict a ’step’ in the trend line.
http://inel.wordpress.com/2007/08/09/hadley-centre-decadal-climate-prediction-system/
[Response: Yes I’ve seen the paper, in fact I posted about it.]
Raven // January 31, 2008 at 4:09 pm
If the GISS has been fixed the Micheals and McKitrick would not have been able to find correlations between temps and social economic data. Nor would Peike be able to demonstrate biases in the measurement techniques. BTW - IPCC IR4 acknowledges the correlations found by McKitrick but dismisses them as mere “coincidence”.
Auditing is not perfect but it is a lot better than doing nothing and expecting to people to blindly trust the data. Especially when the gatekeepers like Hansen have long since dispensed with any notion of scientific objectivity and become political activists.
The limited disclosure of GISS methods was only done after the government forced NASA to do so. More critically: the information that was disclosed did not allow others to replicate the work which means the disclosure was meaningless.
I consider the fact that it was necessary to fight to get any disclosure from NASA is more evidence of bad faith on the part of the gatekeepers and yet one more reason why the data should be treated as suspect until proven otherwise.
The idea that the fossil fuel industry should finance the effort is absurd - you know damn well that you would reject any work funded in that manner. The money should come the governments that fund the people making the alarmist claims.
[Response: You’re entitled to your own opinion. But you’re not entitled to your own facts.
GISS procedures have been part of the peer-reviewed literature for nearly a decade, and have always been an open book.
As for correlations between temps and social economic data, quite a bit of fudging and cherry-picking was included to make the correlations appear stronger than they really are; essentially, they simply removed the data they didn’t like. And if you don’t believe in the existence of coincidence, you don’t know much about statistics.
Your statement that “gatekeepers like Hansen have long since dispensed with any notion of scientific objectivity” is nothing short of libelous. It’s the last time you’ll make such a statement here; reiterations will go straight to the trashcan.]
Bruno De Wolf // January 31, 2008 at 4:32 pm
Tamino,
Agree that in the long run, both satellite and surface measurements will tell the same thing, there can be however quite big differences in between metrics in a single year. E.g. compare 1998 (yes, I know but let me speak) with 2007
GISS: 2007 was as hot as 1998
UAH: 2007 was 0.2 °C colder than 1998
Or: year-to-year variability between measurement systems can go up to 0,2°C
Such an exceptional year as 1998 (or vice versa, a year with a exceptional ‘cold’ event) can easily create a false positive or negative if your confidence interval is 0,4°C. If you truly want to stick with one source (yes, I agree, it’s simpler), you will have to score 3 points before you convince me (or vice versa: I give you the right to doubt for up to 3 times outside the confidence interval).
Heretic // January 31, 2008 at 4:32 pm
Gee, of Mc Kitrick and Hansen, wich could be the scientist and which the political activist? Mmmm, dunno, perhaps a look at respective numbers of science papers in peer-reviewed publications could give us a clue…
Lee // January 31, 2008 at 4:46 pm
Oh, good christ, raven.
The raw data is all available to anyone.
The methods are published you can read exactly what they have done.
The computer code is released and available you can read the code that does the work. Whining that it doesn’t compile on a different computer does not change that - its sitting RIGHT THERE to be read.
HADCrut does an independent analysis of mostly the same data - different criteria for which data to include, but that is part of the independence of theri analyis - with very similar results.
The satellite analyses, with completely different data and analytic methods, analyzed by independent groups, track both GISS and Hadley very closely.
The Surface Stations project, which set out to discredit the surface record, instead confirmed that the “high-quality” station data nearly perfectly tracks the GISS analysis.
How much more fricking utter transparency and confirmation do you need?
george // January 31, 2008 at 4:47 pm
raven said: “Governments should take control of the data aways [sic] from the agencies developing the models. ”
I always thought that NASA was part of the US “government” (ie, a “US government agency”)
Perhaps the “N” stands for “NON-governmental”? (or perhaps “NON-objective”?)
You learn something new every day on the internet.
Hank Roberts // January 31, 2008 at 5:03 pm
> … Hadley … posted about it
Blush. Thanks for the reminder. I thought I’d looked before asking.
[Response: No problem. I appreciate pointers to interesting stuff. I just can’t keep up with it all, but occasionally a reader points me to an extremely valuable work … and there’s no harm in a pointer to something I’ve already seen.]
JCH // January 31, 2008 at 5:05 pm
Just to satisfy your doubts, how much are these audits going to cost? Wouldn’t a little dutch boy cost-benefit analysis demonstrate that money would be much better spent on treating aids victims in Africa, or spraying DDT in the tropics?
george // January 31, 2008 at 5:52 pm
Tamino said:
and
Isn’t that based on the assumption that warming continues at the same rate given by the trend over the past 30 years?
What if the warming trend continues, but at a lower rate (lower slope), from now through 2015?
Then we have an upward trend line with lower slope bracketed by 2-sigma error lines also with lower slope (defining yet a third region)
Wouldn’t that mean that it would be possible for the temp to fall below your bottom dashed red line for two (not necessarily consecutive) years even though global warming (ie, an upward trend in temperature) continued? (assuming the same standard deviation for the residuals)
[Response: You’re quite right. I considered whether or not to address this issue, which complicates things of course, and decided that for the purpose of this blog it was best to omit.
Both basic physics and computer models suggest that the warming rate won’t be decreasing, rather it’s likely to increase. So IF I were betting money, I’d still go with the bet suggested in the post.]
Bob // January 31, 2008 at 5:54 pm
So the bottom of the GISS rising temp range is currently at about .39 C and the bottom of the 2001 to 2007 average range is .35 C. FYI, the December 2007 monthly anomalies in C are as follows:
GISS 0.39
UAH 0.11
RSS 0.08
HadCRUT3 0.21
NCDC 0.40
[Response: FYI, the random variation is *monthly* averages is bigger than in *annual* averages (just as the variation in annual averages is bigger than in 5-year averages), so the limiting ranges will be even wider than for annual averages.
And FYI, the autocorrelation of monthly averages is considerable bigger than that of annual averages, so the red-noise character of the data cannot safely be ignored for monthly data.
FYI, all the data sets to which you refer are on *different scales* because they use a different reference (comparison) period for computing anomalies. Hence each requires different numerical values for its range definitions, just as the “safe operating temperature” of a device will have different numerical values if the temperature is expressed as degrees Kelvin rather than Celsius.]
Kevin // January 31, 2008 at 6:30 pm
This ‘challenge’ is very interesting, and as usual the graphs Tamino presents contribute a lot of clarity to the data. In a way, though, in seems like this challenge accepts a prior move of the goal posts. That is, if this same question had been posed in this same manner in, say, 1989–when eyeballing the data could have given the same misguided impression that some skeptics are promoting today, i.e. that the warming may have “taken a break” in the preceding years–why, then the challenge would already have been resolved. The yearly averages have continued to vary around the increasing trend line. Unless there’s some physical basis to suppose the case is different now, then why must we wait several more years to consider the issue resolved?
Am I missing something? Is there a reason that the skeptics who post here can give why they expect this trend to end or reverse? I mean, a reason in terms of forcings? I think it’s been made quite clear that eyeballing the past 10 years’ averages doesn’t give a supportable reason to think anything is different, and neither does a regression line on enough data to be statistically significant. So why, exactly, should we expect the trend to go away?
[Response: How true! But “global warming stopped in 1998″ is the very public mantra du jour. When it’s shown to be false, we can link to this post in response to the “global warming stopped in 2013″ mantra.]
Julian Williams // January 31, 2008 at 7:16 pm
There has been a lot of work done with nonparametric statistics for reliability analysis on testing hypotheses until they can be either rejected or accepted, without any preset endpoint. The amount of work published on it (and its everyday use in detection equipment software) suggests that it is not anything like as simple as presented here, and there is a real danger of accepting the “wrong” result.
Brian Schmidt // January 31, 2008 at 7:20 pm
I will take Tamino’s side of this bet. For money. Skeptics, what do you say?
I have one modification - to deal with the Hadley Center issue, which is unrelated to AGW, years 2008-2010 don’t count.
Nice job, Tamino.
Aaron Lewis // January 31, 2008 at 9:01 pm
The bottom line is that all the temperature data that Tamino choreographs so beautifully are only proxies for climate effects on agriculture and our ecosystems.
It is easy to check agricultural and ecological data. For example it is easy to track harvest dates. Harvest dates are a very precise indicator of climatic warmth and have huge economic importance. NH fruit harvests were 3 to 5 days earlier last year than they were in 2000. This is very good evidence that Global warming has not stopped, and it has significant current economic impact.
As I write this, my nectarines are breaking bud. In the period of 1997 to 2000, the same trees reached the same stage of bud break on Valentine’s day, and the fruit ripened the last week in June. Last summer, the fruit ripened the second week in June. Due to the early flowering, this year I expect the fruit to ripen the first week in June. (If we have any bees to pollinate the blossoms so early in the year!) The early bloom has totally disrupted my spray schedules. And, never before have I seen red spider mites so active in January. I expect that is because, recently, we have had fewer frosts.
Citrus is a less precise marker, but my tangerines trees that used to ripen fruit the last two weeks of February, this year ripened the last two weeks of January, and were the sweetest ever. I guess the tangerine trees did not really like those frosts in the old days!
Thus, my records suggest that the commercial data actually understates the recent effects of global warming.
Here in California, climate warming has not stopped. I see its effects in the garden and orchard every day.
Zeke // January 31, 2008 at 9:41 pm
Raven,
If you don’t trust GISS, go out and survey all of the temperature stations in the U.S.
Choose the best rural stations.
Create a temperature record based on those.
I’ll spoil the ending, since its already been done. GISS is remarkably close to an independent temperature reconstruction using only the best rural stations for the United States: http://www.yaleclimatemediaforum.org/pics/gistemp.jpg
If you want to work through the data yourself, go to http://www.opentemp.org/main and try it out.
Granted, you can still wax poetic about the quality of temperature measurement stations in China. But since the same arguments were made for the U.S., and it turned out to be reasonably good after all, I’m giving GISS the benefit of the doubt for the moment.
J // January 31, 2008 at 10:14 pm
Okay. How about this?
Pretend that it’s January 1997, and you’re looking at the GISTEMP data. It looks like temperatures rose from 1975-1990, and then leveled off. Maybe global warming stopped back in 1990? Temperatures have been flat from 1990-1996!
S0 I repeated Tamino’s analysis more or less exactly, from the perspective of someone back in early 1997.
From 1975-1990, temperature rose at a rate of 0.0214 deg.C/yr, and the residuals have a standard deviation of 0.09753 deg.C/yr.
I projected that trend out into the distant future of the 21st century (all the way to 2007!), along with its +- 2 SD envelope.
I also looked at the 1990-1996 average, extended that as a “no-trend” line, and gave it the same +- 2 SD envelope.
Just like Tamino.
Now, let’s take our time machine forward to 2008, and look at what really happened to the climate. Since 1996, there have been *no* points that fell outside the “warming-trend” envelope.
On the other hand, there have now been *8* years that fell more than 2 SD above the “no-trend” line. (1998, 2001, and every year since).
In other words, using Tamino’s methodology, if we had made this bet back in January 1997, it would have been resolved definitively by 2001 (well, actually in early 2002, when the 2001 annual mean became available…) and nothing since then would have contradicted this.
The graph looks really nifty. I think I’ll make a copy to show to the next person who tells me that global warming stopped in 2001.
Bob North // February 1, 2008 at 12:19 am
I think it is a given that each organization (GISS, hadley, UAH) uses slightly different methods and assumptions in processing the raw data but they all give roughly the same result when looking at long-term trends. Therefore, it seems that using an unweighted average of the GISS, Hadley, UAH, and RSS estimates of the global temperature anomaly would be the most reasonable approach for future evaluations of whether or not warming is continuing. I don’t think that either outcome will necessarily prove or disprove the AGW theory. What it will tells us is that our estimates of the effects of increasing CO2 (ie., the forcing) are either too high, too low, or just right.
Rather than just looking at whether the current trend is continuing or not, perhaps the better test is how close the average trend from let’s
says 2000-2015 matches the average temperature increase predicted based on the increase in CO2.
Bob North
steven mosher // February 1, 2008 at 12:26 am
Nicely put bet Tamino. I think it’s fair. I have some quibbles ( mainly about the AGW hypothesis implying INCREASED warming) but I think you did a fair job. And it’s well explained.
On Hadcru. I see a bunch of people arguing that GISS and hadcru use the same data.
That’s an open question. Until we forced Dr, Jones via FOI to release his list of stations, we didnt have any confirmation that HAdcru used the same stations as GISS. In fact, we had some evidence that they used diferent stations.
Now, that comparision can be done. I think those who claim the stations are the same should have a look.
The other issue with hadcru is the lack of transparency WRT to the actual data .
GISS, after much lobbying (psst I invented the drive to free the code) has seen fit to release everything: Stations, data, code. GOOD FOR THEM. The day nasa did this I posted a thank you on RC. It was filtered. maybe it was OT.
So I repeat the thank you here
In the future I think IPCC should use GISS rather than HADCRU. Transparency reduces doubt. That’s a good thing.
[Response: I prefer GISS (I think it’s better to estimate unobserved-but-near-to-observed regions rather than omit), but I doubt IPCC will change, mainly because they’ve used HadCRU so far, and there’s an argument to be made for consistency.]
Bob // February 1, 2008 at 1:25 am
GISS and HADCRU do not use the same data sets, Check the links conveniently provided by Tamino. GISS uses satellite data for Sea Surface Temps and HADCRU uses ship measurements. This would appear to be a major difference in two data sets given the size of the oceans.
EliRabett // February 1, 2008 at 4:09 am
Given that the records all track each other it pretty much doesn’t matter which ones you use. Assuming that the AMSU stays on line, all the satellite reconstructions should have essentially zero variance from each other. The rest is noise.
steven mosher // February 1, 2008 at 1:58 pm
tamino, I’m not so sure on the GISS inclusion of the “unobserved” I’m assuming your refering to how they treat the artic. I’d take the hadcru approach and live with the greater uncertainity.
As I understand it Hansen et al use stations within 1200Km of each other. The issue is in Hansens orignal study the coorelation study to determine this distance showed that at that distance the correlation was around .6 for the northern hemisphere and .5 for SH. The problem is what is the correlation across the polar region? In any case I think reasonable people can disagree about this. And it has no bearing on AGW as far as I can see.
Jim Arndt // February 1, 2008 at 11:50 pm
Hi,
I think I’ll put my bet on Anthony Watts. Here he shows the correlation between AMO, PDO and TSI better correlate than CO2
http://wattsupwiththat.wordpress.com/2008/01/25/warming-trend-pdo-and-solar-correlate-better-than-co2/#more-597
Leave a Comment