Guest commentary from Richard Millar (U. Oxford)
The recent Lewis and Curry study of climate sensitivity estimated from the transient surface temperature record is being lauded as something of a game-changer – but how much of a game-changer is it really?
The method at the heart of the new study is essentially identical to that used in the much discussed Otto et al. (2013) study. This method uses a simple equation of the energy balance of the climate and observations of global temperature change and estimated ocean heat uptake anomalies along with a time series of historical radiative forcing (code), in order to make inferences about the equilibrium climate sensitivity (ECS – the ultimate equilibrium warming resulting from doubling carbon dioxide concentrations) and its shorter-term counterpart the transient climate response (TCR – the warming at point of doubling after carbon dioxide concentrations are increased at 1% per year). [Ed. An overview of different methods to calculate sensitivity is available here. The L&C results are also discussed here].
Lewis and Curry use an updated radiative forcing estimate over that used in Otto et al along with slightly different assumptions over the periods used to define the observational anomalies. They use the latest IPCC numbers for radiative forcing and global temperature changes, but not the latest IPCC ocean heat content data. Their result is a 5 – 95% confidence interval on ECS of 1.1–4.1K and for TCR is 0.9-2.5K. These confidence intervals are very consistent with other constraints, from paleo or emergent observations and with the range of GCM estimates. For the TCR, arguably the more important measure of the climate response for policy makers as it is a better predictor of cumulative carbon budgets, the 5-95% confidence intervals are in fact almost identical to the AR5 likely range and similar to the CMIP5 general circulation model (GCM) estimated 5–95% range (shown below).
Figure 1: The 5-95% confidence ranges for transient climate response (TCR) taken from various studies as in Fig. TS.TFE6.2 of IPCC AR5 WG1. The green bordered bar at the top of figure is the estimated 5-95% range from the CMIP5 GCMs. blue bordered bar at the top of the figure is the 5-95% range from the Lewis and Curry (2014) study. The grey shading represents the AR5 consensus likely range for TCR.
There is a difference between the Lewis and Curry 17-83% confidence intervals and the IPCC likely ranges for TCR and ECS. However, for all quantities that are not directly observable, the IPCC typically interprets the 5-95% confidence intervals as likely ranges to account for the possibility that the model used to derive the confidence intervals could be missing something important (i.e. non-linearity that would not be captured by the simple models used in Otto et al and Lewis and Curry, which can particularly be a problem for ECS estimates using this method as the climate feedback parameter is assumed to be constant in time) [IPCC AR5 WG1 Ch10.8.2]. In this case, accounting for more complete surface temperature changes (Cowtan and Way, 2013), or the hemispheric imbalance associated with aerosol forcing (Shindell, 2014), or updates in the OHC changes, may all shift the Lewis and Curry distribution. [Ed. This expert judgement related to structural uncertainty was also applied to the attribution statements discussed here before].
The median estimate of the TCR from Lewis and Curry (1.3K) is towards the lower end of the IPCC likely range and lower than the CMIP5 median value of around 1.8K. A simple way to understand the importance of the exact TCR value for mitigation policy is via its impact on the cumulative carbon budget to avoid crossing a 2K threshold of global surface temperature warming. Using the Allen and Stocker relationship between TCR and TCRE (the transient climate response to cumulative emissions) we can scale the remaining carbon budget to reflect different values for the TCR. Taking the IPCC CO2-only carbon budget of 1000 GtC (based on the CMIP5 median TCR of 1.8K) to have a better than 2 in 3 chance of restricting CO2-induced warming to beneath 2K, means that emissions would have to fall on average at 2.4%/year from today onwards. If instead, we take the Lewis and Curry median estimate (1.3K), emissions would have to fall at 1.2%/year. If TCR is at the 5th percentile or 95th percentiles of the Lewis and Curry range, then emissions would need to fall at 0.6%/year and 7.1%/year respectively.
Non-CO2 emissions also contribute to peak warming. The RCP scenarios have a non-CO2 contribution to the 2K peak warming threshold of around 0.5K [IPCC AR5 WG1 – Summary for Policymakers]. Therefore, to limit total warming to 2K, the CO2-induced contribution to peak warming is restricted to around 1.5K. This restricts the remaining carbon budget further, meaning that emissions would have to fall at 4.5%/year assuming a TCR of 1.8K or 1.9%/year taking TCR to be equal to the Lewis & Curry median estimate of 1.3K (assuming no mitigation of non-CO2 emissions).
While of some scientific interest, the impact for real-world mitigation policy of the range of conceivable values for the TCR is small (see also this discussion in Sci. Am.). For targets like the 2 K guide-rail, a TCR on the lower end of the Lewis and Curry and IPCC ranges might just be the difference between a achievable rate of emissions reduction and an impossible one…
References
- N. Lewis, and J.A. Curry, "The implications for climate sensitivity of AR5 forcing and heat uptake estimates", Climate Dynamics, vol. 45, pp. 1009-1023, 2014. http://dx.doi.org/10.1007/s00382-014-2342-y
- A. Otto, F.E.L. Otto, O. Boucher, J. Church, G. Hegerl, P.M. Forster, N.P. Gillett, J. Gregory, G.C. Johnson, R. Knutti, N. Lewis, U. Lohmann, J. Marotzke, G. Myhre, D. Shindell, B. Stevens, and M.R. Allen, "Energy budget constraints on climate response", Nature Geoscience, vol. 6, pp. 415-416, 2013. http://dx.doi.org/10.1038/ngeo1836
- K. Cowtan, and R.G. Way, "Coverage bias in the HadCRUT4 temperature series and its impact on recent temperature trends", Quarterly Journal of the Royal Meteorological Society, vol. 140, pp. 1935-1944, 2014. http://dx.doi.org/10.1002/qj.2297
- D.T. Shindell, "Inhomogeneous forcing and transient climate sensitivity", Nature Climate Change, vol. 4, pp. 274-277, 2014. http://dx.doi.org/10.1038/nclimate2136
- P.J. Durack, P.J. Gleckler, F.W. Landerer, and K.E. Taylor, "Quantifying underestimates of long-term upper-ocean warming", Nature Climate Change, vol. 4, pp. 999-1005, 2014. http://dx.doi.org/10.1038/nclimate2389
- M.R. Allen, and T.F. Stocker, "Impact of delay in reducing carbon dioxide emissions", Nature Climate Change, vol. 4, pp. 23-26, 2013. http://dx.doi.org/10.1038/nclimate2077
2C in 100 years. Yawn. No one is going to vote themselves into a recession for something that may happen in 100 years. Especially when it is known the remedy we would take on, will not even solve the problem. How much are you willing to write a check for, right now, to save mansions in Miami Beach in 2100?
Ever wonder how the scientific literature would read if Rupert Murdoch took over Nature ?
The answer may lie in the WSJ’s editing of Judith Curry’s op-ed version of the article under discussion
#26 I hope you do not work for my government. What I do not understand is how so much of the climate discussion, after all these years, can be dominated by the idea that establishing warming and a role for man in causing it is sufficient to justify “urgent action”. For anyone interested in the lot of humanity, that argument simply does not follow on its own. The question is whether the human costs of the deep interventions necessary to reduce emissions are worth the human benefits, including environmental costs and benefits which humans of course care about. Where this rather hard question has been looked at, eg Lord Stern, the evidence supporting action is not compelling in my view.
@40 I did manage to miss out one of my intended links. As this would have been to the work of one of our hosts (and so not because Victor@36 deserves a more complete rebuttal), I here make good.
Climatology has no problem linking OHC to AGW, or indeed SLR to AGW which can “explain(ing) 98% of the variance.”
#47–(Victor) {https://www.realclimate.org/index.php/archives/2014/10/climate-response-estimates-from-lewis-curry/comment-page-1/#comment-609524}
Sure, it’s both a reality and a ‘mystery’–at least in the sense that it has specific causes which would be scientifically interesting, and which are in fact being investigated.
However, that does not make it a ‘problem for AGW.’ As the Santer paper showed, as people have been saying for a very long time, aand folks more interested in polemics than science have been ignoring, such periods are neither unexpected nor problematic. Saying that there ‘should be’ a more obviously correlation between the surface temperature record and GHG concentration–“But if one were the cause of the other there would be a clear correlation between them at all, or almost all, points.”–signifies nothing, regardless of how many times it is repeated. (Note in this regard, that climate science, far from asserting that CO2 is ‘the cause’ of the surface temperature, has taken great pains to quantify the influence of *all* relevant causes.)
<blockquote)But you can’t simply dismiss a graph as cherry picked simply because it doesn’t suit you …
Well, I could. Deniers do it all the time, and I could certainly follow their poor example.
However, I didn’t. I dismissed it as cherry-picked because it displays a strategically-picked date and data set, while ignoring other possible periods and data sets, apparently in order to make the strongest rhetorical case. Now, let’s see what a random online check finds as definition of “cherry-pick”:
Had you looked at other time frames, or other data sets, it would be another story. But you didn’t.
The part you really need to know about this question, though, is what cherry-picking means for statistics: standard statistical tests work pretty well most of the time. However, if you do an ex post facto cherry pick, all that goes out the window; what would otherwise be reasonably indicative shows little or nothing. And that’s what you have here. Do a linear fit starting in the second quarter of 1999, and here’s what you get:
http://www.woodfortrees.org/plot/rss/from:1999.3/plot/rss/from:1999.3/trend
Hmm, looks like the warming is back–even though this is still RSS, which shows a very different result than GISTEMP for this period, as I showed in my previous comment. The lack of consilience is a sign here, that the result upon which you are pinning your blog post is, as they say, “not robust.”
And nice demonstration is here:
http://www.woodfortrees.org/plot/rss/from:1999.3/plot/rss/from:1999.3/trend
#51–(Eric Zahn.)
Yawn. Just yawn. Rhetoric, falsehood, more rhetoric.
Whatever, dude.
#53–(ben.)
I think you may have missed some analysis, here and there. For example:
http://report.mitigation2014.org/spm/ipcc_wg3_ar5_summary-for-policymakers_approved.pdf
There’s more, of course–much, much more. But please don’t try to tell me that benefits and costs haven’t been examined.
Re- Comment by Victor — 11 Oct 2014 @ 10:53 AM, ~#47
Victor, if you could explain to me what you mean by “the hiatus” I would be very interested. Please check out- https://www.realclimate.org/index.php/archives/2013/09/the-new-ipcc-climate-report/#more-15766 and on the first graph point out what, as you say, “most climate scientists agree on that score.” I am having a real problem with this, especially on the lower half of the graph where the stair steps don’t seem to be decreasing much within the 15 to 18 year period you have identified. Because you understand what the scientists agree on I would like to know what you think they are talking about.
Steve
#55 Kevin McKinney
It was never my intention to present a balanced view of the problem but simply to provide a clear illustration of it. I presented three different graphs that illustrate what has already been acknowledged by a great many recognized climate scientists, including the authors of the NASA paper I cited. There was no need for me to cherry pick since I had no expectation that anyone would be convinced solely by the evidence I presented. If you are unwilling to accept the reality of the hiatus, you are entitled to your belief, but please don’t assume this is the consensus view, because it is not. The hiatus is widely accepted as real, and as a problem for climate science. And as I’ve already made clear, it really doesn’t matter whether 21st century warming has been flat or somewhat elevated. What matters is that the assumed correlation has broken down.
My argument is based on fundamental principles of science, not my or anyone else’s understanding of climate science per se. Judging by the defensive tone of those responding here, this is an emotional issue for most of you — that’s your problem, not mine. I have no skin in the game and have no interest in attacking anyone simply because they disagree with me.
If you want to see an excellent example of truly outrageous cherry picking, I invite you read the most recent comments on my blog: http://amoleintheground.blogspot.com/2014/10/common-sense-on-climate-change.html#comment-form
Kevin@41&42,
and
Santer et al base their analysis on TLT data – UAH and RSS.
I’m not smart enough to sort out what they are saying when their introductory remarks seem inconsistent with their conclusions.
I don’t see how the the requirement multi-decadal records squares with ‘at least 17 years,’ or with that interval compared to the comment that AGW can be continuing even if a decade is no warmer than the previous one.
This paper is cited by skeptics announcing that 17 years are up with little to no warming in the TLT records. I always point out the last paragraph, and suggest that 17 years is a bare, but not necessarily determinate mnimum.
If anyone can explain the results, I would greatly appreciate it.
Victor #47. Victor says “I also pointed out the lack of correlation prior to 1975. Sure, we see a situation where both warming and CO2 emissions have clearly risen over the past 100 years or so. But if one were the cause of the other there would be a clear correlation between them at all, or almost all, points.”
Not really. We have evidence of a longer term warming trend due to CO2, and an additional interplay with this trend from shorter term natural cycles like el nino or periods of particulate emissions. So there will be no “clear correlation” but rather a rough correlation.
Victor @58
You are an exceedingly foolsh fellow to persist with this denialist nonsense here. As you do persist, let me explain your failings to you.
The blog post you are so proud of attempts to show that levels of atmospheric CO2 are not convincingly correlated with global surface temperature and and with sea level rise. The silly error you make concerning sea level rise is described fully @40/54.
This leaves the CO2/temperature correlation which is banal nonsense found on most denialist websties. So what is new? What do you bring to the party that is any different from all the other nonsense? What is your evidence? You talk of “hiatus” And then what? I so nothing beyond “hiatus”. Not a squeak. To you, the “hiatus” is proof enough but it doesn’t matter to you because you are in denial.
Look at you evidence, man!! You present a graph from Wikipedia which plots GISStemp monthly data and a 60-month running average. This shows quite clearly a “hiatus.” So as to be sure there is no doubt, you cut&paste the second graph from Wiki showing it in close up. So how long is this “hiatus” you are waving at us?
If you zoom in on the detail of Wiki gaphs or even resort to the GISS data, that “hiatus” begins in June 2003. The running average then continues to March 2012. So how long is this “hiatus”? That wound be 8 years and 9 months, and counting. But that isn’t long enough, is it? Not for you!!
So do you consider the references at the Wiki site? AR5 Technical Summary Box TS.3 talks of a ” a much smaller increasing linear trend over the past 15 years.” The NOAA Q&A referenced talks of “the ‘pause’ in global warming observed since 2000.” These are not as exact as the 8 years and 9 months shown by GIStemp but you could convert “over the past 15 years” into ‘a 15-year-long hiatus.’ Who would notice the slight of hand? Of course somebody might and also those references don’t say zip abut this “hiatus” being incompitable with CO2-the-climate-driving-GHG. Indeed those references make pains to point out that the “hiatus” is not a problem with that fundamental scientific finding.
So you don’t reference those folk. Instead you ignore them, which isn’t very scientific.
So where do you look?
You seek out the wisdom of the mad Viscount Monckton of Brechley. Of course his point is not that there has been a 17 year 8 month “hiatus” but rather that there has been“no global warming statistically distinguishable from zero over the 18 full years from March 1996 to February 2014” but even that was not good enough for your purpose.
And quite right too. A lack of statistical significance does not mean something doesn’t exist.
Instead you present Monckton’s graph of RSS data where he shows in this noisiest non-surface temperature record, there is a period with a linear trend that is actually equal to zero warming. As has be pointed out to you by Kevin McKinney @55 such a use of the data is cherry-picking. But it is worse than that. By dodging His Lordships main finding, Victor, you even cherry-pick from the cherry-picker. That is unbelievable.
And for you to then persist in presenting such silly nonsense here where the name of the game is science, words fail me. Frankly, Victor, if you weren’t so incompetant, I would be accusing you of trolling.
#59–(Victor)
“It was never my intention to present a balanced view of the problem but simply to provide a clear illustration of it.”
Well, you’ve failed. Your ‘illustration’ is not only unclear, it’s misleading.
“There was no need for me to cherry pick since I had no expectation that anyone would be convinced solely by the evidence I presented.”
Cherry-picking is a matter of using a valid sample, not a matter of intent–it’s a technical question, not a legalistic one.
“If you are unwilling to accept the reality of the hiatus, you are entitled to your belief, but please don’t assume this is the consensus view, because it is not.”
Perhaps you should re-read my comments for comprehension. Obviously, the ‘hiatus’ is real. The question is, does it mean much? Santer et al, and numerous other scientists, having been saying “probably not” AND providing reasons why.
“The hiatus is widely accepted as real…”
Yes…
“…and as a problem for climate science.”
No.
“And as I’ve already made clear, it really doesn’t matter whether 21st century warming has been flat or somewhat elevated. What matters is that the assumed correlation has broken down.”
No. As Santer et al showed, and Tamino less formally showed, there is no evidence of that. The hiatus has not achieved statistical significance. And, while I’m responding to this sentence, the correlation (over longer time frames) isn’t ‘assumed.’
“My argument is based on fundamental principles of science, not my or anyone else’s understanding of climate science per se.”
Really? You seem to be a bit selective about which principles are ‘fundamental,’ and it’s a irony: your central point is that ‘the correlation has broken down’, but you seem quite ignorant of basic stats.
“Judging by the defensive tone of those responding here, this is an emotional issue for most of you — that’s your problem, not mine. I have no skin in the game and have no interest in attacking anyone simply because they disagree with me.”
That’s your perception. My perception is that you seem pretty attached to your ideas, and refuse to pay attention to people who point out problems with them. Indeed, the two last sentences could be considered an ‘attack’, basically boiling down to “Oooh, you alarmists are too emotional!”
#59–Good points, Barry, but let me emphasize the “at least”–it’s a *minimum*. Similarly, 17 exceeds one decade, so loosely…
You said it well yourself: “I always point out the last paragraph, and suggest that 17 years is a bare, but not necessarily determinate minimum.”
Victor writes:
“The hiatus is widely accepted as real, and as a problem for climate science.”
I don’t think climate scientists see it as a problem for climate science, if by that you mean questioning anything fundamental about climate forcings or sensitivity or anything like that. The known forcings plus unforced variability account for the recent variations in global temperature rather well – e.g. Kosaka & Xie:
http://www.nature.com/nature/journal/vaop/ncurrent/full/nature12534.html
@Victor #58
You might take a look at this reconstruction, one of very many examples
You can look up on the web.
http://www.psmsl.org/products/reconstructions/jevrejevaetal2008.php
You rely heavily on the conclusion that sea level rise predates global warming, based on the wiki source, that is not very convincing.
Furthermore it is not a good idea to expect really a 1:1 relationship to global warming, as (for instance) also cahnges in patterns of fresh water use /damming plays a major role
barry @59,
Does this address the question you have? The term that describes why the minimum period is >17 years or so for statistical significance to appear in a trend is “signal-to-noise.” The bigger the wobbles & the greater the scatter, the longer the data run you need. For constant noise, uncertainty levels are roughly proportional to Amount-of-Data^1.5, so with monthly data with a similar scatter and inter-annual wobbles and a similar slope, the trend will emerge from the noise after about the same time period. And for monthly global temperatures rising at 0.1-0.2ºC/decade, that comes in at 17 years or so. The TLT data is more wobbly that the surface temps so will require longer if the slope of trend is the same.
Victor wrote: “My argument is based on fundamental principles of science, not my or anyone else’s understanding of climate science per se”
Your “argument” is based on repetition of falsehoods, distortions, misrepresentations, irrelevancies, fallacies and nonsense — and after multiple commenters have thoroughly demolished it, all you can do is accuse them of being “emotional”, and repeat the claims they have already demonstrated to be false.
It’s the most boring and stupid sort of denialism, and why the moderators are not dumping your comments into the Bore Hole is beyond me.
#64. Thanks, Icarus, for the link to the Kosaka and Xie abstract. Here is a direct quote: “Despite the continued increase in atmospheric greenhouse gas concentrations, the annual-mean global temperature has not risen in the twenty-first century, challenging the prevailing view that anthropogenic forcing causes climate warming.” Looks like the “banal nonsense” I’ve been accused of disseminating is accepted by recognized climate scientists as well. Two of them, at least. From my readings on this topic there are a great many more who’ve expressed essentially the same opinion. If you think the “hiatus” can easily be dismissed you are wrong.
The fix offered by these authors is ingenious. The apparent breakdown in the correlation between emissions and warming can be explained by “recent cooling in the eastern equatorial Pacific.” According to a summary of this paper in Nature, the cooling found by Kosaka and Xie is generated by a La-Niña effect. “The equatorial Pacific cooling turns out to be strong enough to offset the general rise in temperature induced by anthropogenic greenhouse gases.”
This paper should be compared with a more recent one, from NASA, by Paul J. Durack et al., alleging that surface ocean temps have been grossly underestimated: “Relying on the close correspondence between hemispheric-scale ocean heat content and steric changes, we adjust the poorly constrained Southern Hemisphere observed warming estimates so that hemispheric ratios are consistent with the broad range of modelled results. These adjustments yield large increases (2.2–7.1 1022 J 35 yr-1) to current global upper-ocean heat content change estimates, and have important implications for sea level, the planetary energy budget and climate sensitivity assessments.”
Thus we have two fairly recent studies offering two radically different assessments of temperatures in the upper oceans. If the NASA report of radically higher ocean temps is correct, then it’s hard to see how the adjustment offered by Kosaka and Xie, based on a cooling of the ocean, could work.
“Despite the continued increase in atmospheric greenhouse gas concentrations, the annual-mean global temperature has not risen in the twenty-first century, challenging the prevailing view that anthropogenic forcing causes climate warming.”
At best that is poorly worded, at worst it’s utter nonsense. The surface temperature record shows a warming trend over the 21st century. In no way can a warming trend be turned into a claim that temperature hasn’t risen. And the part about human forcing causing warming being in doubt is ludicrous.
“Despite the continued increase in atmospheric greenhouse gas concentrations, the annual-mean global temperature has not risen in the twenty-first century, challenging the prevailing view that anthropogenic forcing causes climate warming.”
It’s badly worded nonsense I am afraid. The surface temperature record shows a positive trend over the 21st century. In no way can that be turned into a claim that temperature hasn’t risen over the 21st century.
#68–Well, thanks for turning to some actual science, Victor. (Though I must say I find it a bit odd to cite a paper proffering an explanation for the hiatus within the framework of conventional thought on climate as an example of the hiatus-as-fatal-inconsistency thesis, just because they used the word ‘challenge’ in the abstract.) They essentially do the same thing others have done in investigating short-term weather/climate variability–that is, they look at the influence of other factors they think may be relevant.
Just go to Google Scholar and search “climate hiatus” to see similar work.
As to comparing the two you cite, I’d note that there’s a big difference between cooling sea surface temps in a region of the equatorial Pacific, and hemispheric scale increases in ocean heat content down to several hundred meters of depth. It’s quite possible that both trends could coexist.
MARoger@66,
Thanks for the reply. No, it doesn’t clear it up.
Firstly, Santer et al is based on MSU TLT records. There is no need to suggest that longer records are required when considering TLT records, as they are the records applied in the paper. (If I understand your comments correctly)
Secondly, I’m not clear on the apparent ambiguity between their abstract and concluding remarks.
“..at least 17 years” – vs- “multidecadal periods are required”
“..at least 17 years” – vs – “we do not expect each year to be inexorably warmer than the preceding year, or each decade to be warmer than the last decade, even in the presence of strong anthropogenic forcing of the climate system…”
Any commentary I’ve seen on that does not specifically clear up this ambiguity, but usually emphasises whichever period is preferred to make a point. Is 17 years rolled gold, sufficient, a bare but not necessarily determinate minimum? Should we ignore that figure in the abstract and go with the concluding remarks? And why?
(I am quite aware of the signal to noise issues, and how they are statistically different for surface and satellite temp records)
Markus @69/70.
If you examine the source of Victor the troll’s quote, Kosaka & Xie (2013), you will perhaps see he has edited the quote in that he deleted sign of two references from it.
The two references put the issue addressed by Kosaka & Xie (2013 in context. Thus from Easterling & Wehner (2009) we can read that the “challenge” is less scientific in origin but more to do with blather in the media etc:-
The issue of “hiatus” as a marker of AGW is also seen as mainly non-scientific by the second reference Foster & Rahmsstorf (2011):-
I would thus suggest that the opening sentence of Kosaka & Xie (2013) is indeed “badly worded” as you say, but only in the context of an extra-scientific audience. However, the authors are well aware of that audience and should have therefore done better at avoiding the creation of such obvious ammunition for such a rowdy audience.
Barry @72.
Sorry, but I do not see where within Santer et al. (2011) that they contrast “…at least 17 years…” (as mentioned in their abstract) with the “multidecadal periods” that they mention in their conclusions. Are we looking at the same paper?
Victor @68.
More foolishness from you, I see. Is it truly your understanding that Kosaka & Xie 2013 (which considers the effects of La Nina, a cooling of SST in the “equitorial Pacific”) is incompatable with the work of Durack et al. (2014) (who address the assessment OHC of the Southern Hemisphere)? What is the problem? It is not as though one of them is arguing for something to be bigger while the other argues for something to be smaller. And even if they were, they are discussing different things (SST & OHC) in different places (tropics/SH).
This is, as you tell us @68 that you are capable of, not simply “banal nonsense.” This is now childish gibberish. Well done you!!
Barry–#72, currently—
I can’t speak for MARoger, but to expand on my response to your original comment, let me say a couple of things.
1) Good catch that Santer et al was based on TLT records. I honestly wasn’t remembering that aspect of the paper when I originally commented, and so bear primary responsibility for leading the discussion off into the weeds on that issue. It rather begs the question as to what a parallel analysis of surface temperature records would show–I’d think that the minimum would be shorter than 17 years, since that’s a less variable data set, but as far as I know–usually not all that far!–that’s not an analysis that has been done.
2) I think that 17 years *is* a ‘multidecadal period by one reasonable interpretation: it’s a span exceeding one decade. Therefore it is not clear to me that there is a contradiction in:
I don’t know if that is what Santer et al intended, but it seems to me not unreasonable.
3) I really don’t get your confusion in other respects. I see no contradiction whatever in:
Rather, these seem complementary. As illustration of what I mean, take the case of Victor, which has been mooting, as you will have observed. He asserts that the lack of warming (and I’ll accept that there has been a lack of *statistically significant* warming over the ‘hiatus period) is a “problem” for climate science, by which he seems to mean a fundamental structural blow, calling the whole edifice into question. (As opposed, say, to a “problem” in the sense of a good research question for those who want to investigate short-term forcings or natural variability, and to relate that to existing theory, which is what Kosaka & Xie appear to me to be trying to do.)
If I’m interpreting things correctly, he sees the lack of marked warming as surprising, or exceptional, in light of the consensus model. But if he were to see a period during which “each year [were] inexorably warmer than the preceding year”, then he presumably would feel different. It’s too early to tell whether the present decade will turn out to be warmer than the last, if we go by the conventional calendar decades. But if we compare the averages for 1984-93, 1994-2003, and 2004-2013, we find that they are, in order, .24 C, .46 C, and .58 C (rounding to the nearest hundredth, and based on January-December averages.)
So, despite the hiatus, we do in fact have a situation in which each of the last three decades has ‘inexorably’ been warmer than the previous. (That has not, however, been the case throughout the 20th century.)
By the way, that is for GISTEMP data, which is available here:
http://data.giss.nasa.gov/gistemp/tabledata_v3/GLB.Ts+dSST.txt
Second ‘by the way’: the year to date average is .65 C, with a good probability of rising by year’s end. The warmest year to date in GISTEMP was 2010, coming in at .66. So we will be close to, or perhaps even a tick above, the GISTEMP annual record–not proof of anything, but perhaps underlining the fact that excessive confidence as to a real shift in the long term warming trend would be misplaced.
Returning to the point: Barry, I think that the 17 years is exactly what you already called it–a minimum period for which we might expect to see a trend in TLT records. That would imply that we are just now at the point at which the hiatus is starting to look unusual. As to what timeframes are better, well perhaps WHO guidelines, which use 30-year norms, are indicative. But it’s really going to depend upon the data, and upon how it changes over time. Should warming accelerate–which there is some reason to expect–then the times would shorten.
Barry@72
At a typical warming rate of 0.1C/decade and confidence limits for the temperature record of +/- 0.1C it takes 20 years for a statistically significant trend to become apparent. Reality is a bit more complicated, but the bottom line is that your confidence in the validity of a trend improves as the record gets longer.
@71-Kevin McKinney
“As to comparing the two you cite, I’d note that there’s a big difference between cooling sea surface temps in a region of the equatorial Pacific, and hemispheric scale increases in ocean heat content down to several hundred meters of depth. It’s quite possible that both trends could coexist.”
They could co-exist, certainly. However, as far as their effects on the worldwide warming data are concerned, wouldn’t they cancel one another out?
#61 MARodger
“The silly error you make concerning sea level rise is described fully @40/54.”
I found that graph (from https://www.realclimate.org/index.php/archives/2013/09/the-new-ipcc-climate-report/#more-15766), but I’m still puzzled, as I see no rising trend until ca. 1910 in either the yearly or decadal graph. I also see a cooling trend (or stasis) in both graphs from roughly 1940 through 1975. Yet during this entire period, as I understand it, greenhouse gas emissions were steadily rising.
Regarding sea level, if you take a look at the 5th graph down on my blog (http://amoleintheground.blogspot.com/), representing sea level from 1870 to 2008, it looks as though the rise in sea level dates from roughly 1885, several years prior to the beginning of the temperature rise. So why is it silly to wonder about g.w. causing the sea level to rise if that rise began prior to the rise in g.w.? Just asking.
> Barry … “I am quite aware of the signal to noise issues …”
Is the specific data set you’re using becoming noisier in recent decades?
> Barry … I am quite aware of the signal to noise issues
Has the TLT data set you’re referring become more or less noisy, over the time it’s been captured?
If so, does that change the length of record required to detect a trend in shorter segments of it?
Victor @79.
I can but assume you found my comments @40/54 too challenging for you to comprehend. Also, thank you for demonstrating that you are prepared to amend your blog to remove the signs of egregious error.
So I’ve done you a simple graphic (two clicks down here) which you are at liberty to use to correct your blog. The data sources are Jevrejeva et al. (2008) and BEST land temperatures. Land temperatures are used to provide 100 years more data than with global temperatures relying on SST. There are, of course, large uncertainties with instrument data stretching so far back in time, but the point is made. There is no demonstrable disconnect between global temperature and SLR.
I will leave unmentioned here your puzzlement based on what is actually AR5 Figure SPM.1a. If you try very hard, you may be able to appreciate that your puzzlement here has no relevance to any SLR-temperature correlation.
> Victor
Those are the kinds of questions that researchers ask — and answer.
You can look this stuff up. Here’s one recent summary for example.
You should look for yourself, though, not trust some guy on the Internet.
I’m just another reader like you. Possibly one of the researchers reading here will comment, but if you do some of the work for yourself, your questions will be more interesting.
Note, at the original page, the footnote numbers are links to their sources:
http://www.nature.com/nclimate/journal/v4/n5/full/nclimate2159.html
Victor—
No, not necessarily. There’s an oceanic boundary layer in the top tens of meters (the depth varies a lot according to local conditions in time and space–there’s a strong annual cycle, and dependence on things including insolation, wind, salinity, currents and more.) Below it, there are ‘property gradients’–relatively sharp transitions in temperature (thermocline), salinity (halocline), and density (pycnocline). All of these vary dynamically.
What that means is that the properties of the boundary layer (which is where ENSO mostly happens) and the underlying water (which is where the OHC changes mostly happen, because there is a lot more water below the boundary layer than in it) show a considerable degree of independence from one another. (Of course, one influences the other, but nevertheless measured temperature, salinity and density can and do differ sharply.)
So it’s possible that OHC below the thermocline could increase with not much observable change in SSTs. I believe that is essentially what people such as Kevin Trenberth mean when they talk about warming being ‘concealed’ in the ocean:
http://www.reportingclimatescience.com/news-stories/article/global-warming-pause-due-to-pacific-says-trenberth.html
A couple of random things I looked at in connection with this comment:
http://onlinelibrary.wiley.com/doi/10.1029/94RG01872/full
http://en.wikipedia.org/wiki/Mixed_layer
The first is a 1994 review paper investigating parameterization of boundary layer fluxes and properties needed to improve coupled numerical models. It’s logged over 2000 citations, so it seems to have had pretty good impact over the years. A lot of it is, frankly, over my head–it’s a highly technical subject, after all. But I found the graphs of specific observational experiments quite instructive (and the discussion intermittently so!)
In reply to replies above,
My layman’s difficulty is that the abstract of the Santer paper vouches that “at least” 17 years of TLT temps is required to winnow out an antrhopogenic signal. Skeptics flourish this quote, point to RSS and UAH and opine that the recent 17-year period shows little to no warming. Thus, ‘according to Santer,’ anthro warming is shown to be next to nothing. Clear so far?
So I point out the ‘multidecadal’ quote in the conclusions.
However, I cannot reconcile that 17 years = multidecadal. 17 years is not even two decades. I am confused by an apparent discrepancy. I cannot ignore the 17-year period just to win an argument. I don’t want to cherry-pick bits of the paper to support a point. I want to understand what the authors mean. If 17 years is sufficient for an anthro signal to emerge, we’re not seeing it in the last 17 years (and that’s how the skeptics put it).
Hank, as far as I can tell, RSS (or UAH) is no noisier in the last 17 years than it has been for the periods discussed in Santer et al. If it is statistically noisier, then that would be good to know, although I doubt it would make a whole heap of difference to the point.
Kevin, the fact of the last decade being warmer than the previous does not really speak to the recent 17-year flatline/little warming. And if you think 17 years is ‘multidecadal’ (I don’t), then the skeptics now can point to the conclusion and the abstract both in Santer et al to make their point.
I think the only way to sort this out is to analyse the paper and perhaps get a better fix on the caveat “at least” (17 years). It’s way beyond me, unfortunately. For now, all I can do is point to the concluding paragraph and suggest that “at least,” is not ‘definitely’, but that’s pretty weak.
(Appreciate the repies)
Thanks to Kevin, Hank and MARodger for your very helpful efforts in educating me regarding the complexities of this science. Thanks especially to MARodger for the very interesting graphic comparison, which I’ll be adding to my blog post soon, along with some discussion, once I have a chance to examine it more carefully.
I admit that I’ve been very hard on some of those responding to my blog post, either here or on the blog itself, because of issues I haven’t been able to resolve in my own mind regarding not so much the data itself but the manner in which it is so often presented and interpreted. On both sides of the debate.
At this point I have to say I’m still not convinced. However, it seems to me as though the problems should clear themselves up in a few years, when it will be possible to test the various proposals offered to account for the hiatus anomaly. If the hiatus persists beyond the scope of ENSO and the other mitigating elements proposed by various researchers, then it looks like it will be a lot harder to deny AGW falsification. If, on the other hand, the hiatus abates in a manner consistent with these proposals, then I’ll be willing to reconsider my position. Again, thanks.
For Victor: always go past the original paper; read the papers citing that one:
http://scholar.google.com/scholar?cites=16172201651443602391&as_sdt=2005&sciodt=1,5&hl=en
For a laugh, compare the results on each search you do
— in Google Scholar (which, altho imperfectly, gets mostly good scientific sources)
— in Google (which gets blog posts, mostly from Watts)
— in Google’s “image search” (which gets illustrations, often misleadingly captioned or entirely faked)
#85, Barry–
Let me take one last crack at this.
Looking at Santer et al (2011), one might be struck by the number of times that they use the phrase “10-year” periods. For example: “The relatively small values of overlapping 10-year TLT trends during the period 1998 to 2010 are partly due to the fact that this period is bracketed (by chance) by a large El Ni˜no (warm) event in 1997/98, and by several smaller La Ni˜na (cool) events at the end of the MSU record (Fig. 7B).”
Or, “On 10-year timescales, distributions of unforced and forced TLT trends overlap markedly (Figure 4A). This overlap occurs because even under anthropogenic and natural external forcing, interannual climate noise has a large influence on short, decadal trends.”
This emphasis on 10-year slices–periods of 20 and 30 years are also mentioned and compared at various points–suggests to me that “multidecadal” might be intended loosely to include any span longer than one decade. (I suppose ‘paradecadal,’ or ‘metadecadal’ might have been more exact, but find I can’t regret that Santer et al didn’t opt to coin either one.)
As to ‘at least,’ I think it means exactly what it says–the time for a signal to emerge from the noise depends upon the strength of the signal, in part. In this case, that signal is only known to an approximation, and may well not be constant over time. All of which means that we can’t expect a single number to act as absolute criterion. What the wording of Santer et al appears to me to do is to set a lower bound: one can’t expect that the signal will emerge unambiguously for periods shorter than 17 years. So, ‘skeptics’ who point to the 17 years as offering ‘proof’ are quite wrong: we are only now at a period where we would expect to see the signal emerge. They are like their cousins who claim that Dr. Maslowski’s projection of an ice-free Arctic has been falsified, since there was not an ice-free Arctic in 2013–in reality, the prediction was for 2016, plus or minus 3 years, so we are only one third of the way into the prediction ‘window’, and five years away from the closing of that ‘window.’
That’s my take. However, there is another statement in the body of the paper that is worth considering. It says:
> “multidecadal” might be intended loosely to include any span longer than one decade.
I’d say that’s the obvious meaning intended, from the context.
‘oogled: define:multi
brings as the first result:
multidecadal:: more than one decade
which fits everything they’ve written.
As centuries pass one decade at a time, there comes a time in the history of any underperforming 100 year climate projection when disinterested observers begin to notice that the rate of change needs to accelerate in order to make up for lost time.
So far, the media at large have allowed the pros on the US Climate Open
lecture circuit three decadal Mulligans in a row-
Never mind that Hansen started the clock in 1988– the year 2100 curves those focused on framing electively promote aim at quantifying the hazards of the next 86 years, rather than explaining the last 26.
Even at the low end of the IPCC 5 scenario range, it will take decadal changes averaging in excess of .17 C per decade to make the 1.5 degree grade by 2100.
If the beginning of the beginning fails to emerge from the decadal noise in the remainder of this decade- as the Bayesian priors suggest it well my, then it will take an even more unprecedented decadal delta T to get back on track- or acknowledge that global thermal inertia has not been parametrized well enough for modeling as usual to continue.
Well Victor – you are not alone. I have been looking at this ‘imbalance’ issue since at least 2009.
I don’t think the latest OHC data does more than confirm Hansen’s reduced imbalance of about 0.5-0.6 W/M2 (over full Earth surface), which is reduced from AR4 and his 2005 estimate. I don’t think Hansen and Trenberth have yet resolved Hansen’s contention that this reduction is due to increased Asian aerosols, a reason for which Trenberth profoundly disagreed.
17 years covers a full 11 year solar cycle and a couple of ENSO’s, so the reasoning that the ‘pause’ is not statistically significant is looking weak.
The upshot is that CO2 has been steadily increasing and the OHC imbalance is being revised down and we have a surface warming ‘pause’ or ‘flattening’ over a significant period. Clearly the CO2 contribution to imbalance is either being exaggerated in models or masked by other factors such as albedo (Hansen’s aerosols?), ENSO (maybe an external forcing after all?).
This pause discussion here is getting pretty silly. What has jiggled a little is the surface temperature, which is only around 2% to 3% of the global radiation energy balance. Do you remember the “global” word? The climate scientists are concerned because they cannot predict these minor deviations, as of yet. Their ethic is to report findings accurately even though they know that opportunistic science denier clowns will take advantage.
Here is a nice simple explanation- http://skepticalscience.com/ipcc-global-warming-pause.htm with links to sources.
Steve
I suppose we should be happy that Curry at last appears to recognise that the climate will respond to a doubling of CO2 (which appears inevitable) in a way that will give us a dangerous level of warming, if the concentration stops there. With politicians using 2C as a limit, Lewis and Curry are telling us that we’ll breach that limit, maybe even with a TCR. And that is even without acknowledging the usefulness of corrections such as those by Cowtan and Way. I hope that Curry now drops her distaste of any actions to reduce our emissions, as her own work shows that not doing that will lead to dangerous climate change.
> Ken Lambert …
> 17 years covers a full 11 year solar cycle and a couple of ENSO’s,
> so the reasoning that the ‘pause’ is not statistically significant
> is looking weak.
Uh, no. That doesn’t make sense.
Look again at the arithmetic used to test whether a significant trend is present.
You’re doing it wrong.
Look at http://www.woodfortrees.org/
You can figure this stuff out. It’s not hard.
It’s best to understand it before you tell other people what you think.
——-
People on all sides of the climate discussion use Woodfortrees, and mostly learn by doing so.
As he says there:
And when you use Woodfortrees, folks, I urge you to chip in.
His tip jar directs help to a woodland trust, not his own pockets.
Like the site, the contributions go to the benefit of everyone.
Like the arithmetic, that’s true regardless of what you believe.
Thanks again, MARodger, for pointing me to your more comprehensive graph, now sitting on my blog along with some discussion of what I think it might mean. You’ve demonstrated the limitations of the sea level graph I used, which I assumed to be more informative than it actually is. I was wrong about that. In my defense, this graph has been widely distributed, and the only one I was able to find that went back as far into the history as it did. Your composite should be published as a corrective, since it goes back a lot farther.
However. As I stated on my blog, there are aspects of this graph that puzzle me. And hopefully you can clarify certain things.
1. From your graph, it looks like the upward trend in temperature begins around 1810, which precedes the Industrial Revolution. There’s a bit of a dip between 1830-1840, and then a very long rise continuing almost without a break until 2000. It’s occurred to me that the original rise might just be part of normal variation — but that later rise, from 1840 on, looks to be the beginning of the warming trend now associated with the “global warming” paradigm. And my question is: isn’t that rather early to be associated with AGW?
2. Similarly we see a more or less continuous sea level rise beginning around 1865 that also continues until 2000. It’s occurred to me that the sea level rise could have been a delayed response to the temperature rise, but 1865 also seems really early in relation to the sort of CO2 emissions that could affect global sea levels.
I’m assuming climate scientists are well aware of these issues and I’m wondering how they would explain them.
Tony Weddle: “I hope that Curry now drops her distaste of any actions to reduce our emissions, as her own work shows that not doing that will lead to dangerous climate change.”
Or not…
http://online.wsj.com/articles/judith-curry-the-global-warming-statistical-meltdown-1412901060
When you’re the darling of the denialoshpere, change is hard.
#90–(Ken Lambert)
“17 years covers a full 11 year solar cycle and a couple of ENSO’s, so the reasoning that the ‘pause’ is not statistically significant is looking weak.”
Hand-wave much? No, Santer et al, cited and discussed above, looked at the statistics over basically all timescales in the TLT record, analyzing signal-to-noise ratios in some depth, to reach their conclusion that 17 years was the *minimum* period for which we might expect to see the signal emerge.
Bringing in solar and ENSO cycles is rather ironic, too, given that Foster and Rahmstorf showed that if you adjust forcing for known solar and ENSO influences over the span of the hiatus, most of it is explained:
http://tamino.wordpress.com/2011/12/06/the-real-global-warming-signal/
“None of the data sets shows any evidence that the global warming rate has changed recently. A truly fascinating result is that increased precision enables us to establish the statistical significance of a warming trend using a shorter time span than with unadjusted data. All five data sets show statistically significant warming since 2000.”
Ah, well, you say–that was back in 2011-2012?
Consider last year’s NCDC ‘top ten warmest list’:
1 2010 : 0.66
2 2005: 0.65
3 1998 : 0.63
4 (tie)* 2013: 0.62
4 (tie)* 2003: 0.62
6 2002: 0.61
7 2006 : 0.60
8 (tie)* 2009: 0.59
8 (tie)* 2007: 0.59
10 (tie) 2004 : 0.57
10 (tie) 2012 : 0.57
And 2011 finished just out at #11, with an anomaly of .51 C.
This year is in the running to become the warmest in the NCDC record, despite the continued dallying of the next El Nino:
http://www.ncdc.noaa.gov/sotc/global/2014/9/supplemental/page-1
So, since we are waving hands here, which way do you think the needle has moved, statistically speaking?
NB–NCDC has just updated for September this morning. The summary:
http://www.ncdc.noaa.gov/sotc/global/2014/9
for Victor: https://www.realclimate.org/index.php/archives/2005/12/early-anthropocene-hyppothesis/
I wish I saw a more hopeful discussion here, recognizing that most climate scientists are backing away from the IPCC AR4 2005 (and earlier) best climate sensitivity estimate of 3 degrees C with recent estimates converging in the 1.5 to 2.0 C range, about one-half to two-thirds of the earlier best estimate. This good news of lower climate sensitivity has the potential to bring together the more moderate and data oriented skeptics, the lukewarmers, and those more moderate and data oriented folks who believed (based on climate sensitivities of 3 and higher)that we need need immediate action to reduce CO2. If the good news were recognized as good news, then most of those posting here (and on most other climate blogs) could speak with one voice to address policy that we all agree is needed: pressure on China and India to transition to lower CO2 fuels at a rate that doesn’t destabilize their and the world economy, reduction of carbon black (soot) which affects both Arctic albedo and world health, curtailment of food for fuel programs such as corn ethanol and also the biofuel programs that result in tropical forest clear cutting, and continued support for many kinds of alternative energy research. We in the affluent west with our 2000 and 3000+ SF houses and two car families are in a real bind. There’s not much we can tell the developing world without being extreme hypocrites. We need to recognize that, and figure out how to create a lower carbon, sustainable future here! My science back ground is biology where, arguably, the world’s best inductive scientist, Charles Darwin, wrote, arguably, the most lucid, modest, and compelling science book ever written, “On the Origin of Species.” I plan to assign that book in my Global Warming/Climate Change class as an example of science excellence. My attempt at climate peace in the midst of climate war is here-
http://climatesensitivity.blogspot.com/
Victor @95.
You say you were only able to find a graph of sea level (sourced on a really obscure website called wikipedia.org and going back to 1870) and thank me for “pointing” you at my more comprehensive graph. This suggests again that your abilities are rather limited.
I actually did more than point you at that graph. I felt it would be useful to create such a graph for you because you were unable to see the evidence you were already being pointed at. For instance, this graph of seal level back to 1700 (which orignates as IPCC AR5 Figure 13.27) and which you were pointed at @40.
I note that, while you take on board this pre-1870 data, you remain dismissive of it. On you blog you now tell the world:-
And the basis for your understanding of “CO2 emission at this very early stage of the Industrial Revolution “ is a apparently a postage-sized graph of CO2 FF emissions 1751-2007 (based on Boden et al. (2011)). I am mystified why any sane individual would consider such a graphic satisfactory evidence for anthropogenic GHG forcing during the early part of that period. So let me continue your education.
The emissions of FF CO2 are not the sole source of anthropogenic CO2. And whatever the difficulty of establishing the record of such emissions, goodness, we actually do have evidence of the atmospheric levels of CO2 from ice cores, levels shown on this Scripps Institute graphic.
And there wasn’t just CO2 at work. Methane was also on the rise. IPCC AR5 Annex II, Table AII.2, gives the rise of all anthropogenic LL GHGs forcings 1751-1840 (Note start date.) totalling 0.173 WM^-2. The table also shows that such a rise in forcing occurred roughly during 1750-1810 and in the following three decades GHG forcing had remained reasonably flat before continuing the rise after 1840.
If we accept this data at face value, that 1840 forcing took 60 years to build with a further 30 years to act and establish its full effect with a resulting increase in global temperatures of perhaps ~0.14ºC+/-0.07ºC. In this context, examine the graph and the wobbles in global land temperature in the BEST record. (An bear in mind land temperatures are more responsive than full global temperatures.) That big wobble in the red 20-year-rolling-average 1810-1840 measures to 0.24ºC peak-to-peak. Such a wobble-size would suggest that within the wobbles 1750-1840, there would be a significant contribution towards an upward trend (or perhaps ‘cancelling a downward trend’ would be more correct) from anthropogenic GHGs during this period, something in the order of 40% to 100% the size of that 1810-40 wobble.
That, I would suggest, deals withn the substance of you Question 1 @ 95.
As for your Question 2 @95 and your understanding of the detail of how increased temperatures translate into SLR, your grasp of the subject is dire. The short answer would be “Stick to century-long trends and ignore the detail if you wish to avoid playing the fool some more.”