Climate Science Glossary

Term Lookup

Enter a term in the search box to find its definition.

Settings

Use the controls in the far right panel to increase or decrease the number of terms automatically displayed (or to completely turn that feature off).

Term Lookup

Settings


All IPCC definitions taken from Climate Change 2007: The Physical Science Basis. Working Group I Contribution to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Annex I, Glossary, pp. 941-954. Cambridge University Press.

Home Arguments Software Resources Comments The Consensus Project Translations About Support

Twitter Facebook YouTube Mastodon MeWe

RSS Posts RSS Comments Email Subscribe


Climate's changed before
It's the sun
It's not bad
There is no consensus
It's cooling
Models are unreliable
Temp record is unreliable
Animals and plants can adapt
It hasn't warmed since 1998
Antarctica is gaining ice
View All Arguments...



Username
Password
New? Register here
Forgot your password?

Latest Posts

Archives

16  ^  more years of global warming

Posted on 10 January 2013 by Kevin C

Update 21/02/2013: Troy Masters is doing some interesting analysis on the methods employed here and by Foster and Rahmstorf. On the basis of his results and my latest analysis I now think that the uncertainties presented here are significantly underestimated, and that the attribution of short term temperature trends is far from settled. There remains a lot of interesting work to be done on this subject.

Human greenhouse gas emissions have continued to warm the planet over the past 16 years. However, a persistent myth has emerged in the mainstream media challenging this.  Denial of this fact may have been the favorite climate contrarian myth of 2012, first invented by David Rose at The Mail on Sunday with an assist from Georgia Tech's Judith Curry, both of whom later doubled-down on the myth after we debunked it.  Despite these repeated debunkings, the myth spread throughout the media in various opinion editorials and stunts throughout 2012. The latest incarnations include this article at the Daily Mail, and a misleadingly headlined piece at the Telegraph.

As a simple illustration of where the myth goes wrong, the following video clarifies how the interplay of natural and human factors have affected the short-term temperature trends, and demonstrates that underneath the short-term noise, the long-term human-caused global warming trend remains as strong as ever.

In particular, once the short-term warming and cooling influences of volcanic eruptions, solar activity, and El Niño and La Niña events are statistically removed from the temperature record, there is no evidence of a change in the rate of greenhouse warming. This replicates the result of a study by Foster and Rahmstorf (2011) under slightly different assumptions.

The human contribution to global warming over the last 16 years is essentially the same as during the prior 16 years¹. Human-caused greenhouse warming, while partially hidden by natural variations, has continued in line with model projections². Unless greenhouse gas emissions are brought under control, we will see faster warming in the future³.

Implications:

  • The 16-year temperature trend provides no evidence to suggest that the consensus understanding of human-caused climate change is incorrect.

Further Reading:

For details of the method, see the Advanced rebuttal to the myth 'no warming in 16 years'.

The results of this analysis are consistent with a statement by WMO Secretary-General Michel Jarraud:

"Naturally occurring climate variability due to phenomena such as El Niño and La Niña impact on temperatures and precipitation on a seasonal to annual scale. But they do not alter the underlying long-term trend of rising temperatures due to climate change as a result of human activities"

Credits: Video: Kevin C. Voiceover: Daniel Bailey. Advice: The SkS team.
Teaser graphics: What happened next? Does this look like global warming?

Footnotes:
We have attempted to keep the language in this video at the same non-technical level as the media stories it refutes. As a result, it has been necessary to simplify much of the terminology. The following notes are for technically literate readers.
¹ i.e. If a change in gradient is allowed at 1997 then the change in gradient is not statistically significant (even at the 1σ level).
² i.e. Within the envelope of AR4 trend projections.
³ On the basis of both AR4 projections and that global GHG emissions are increasing.

Update 21/02/2013: Troy Masters is doing some interesting analysis on the methods employed here and by Foster and Rahmstorf. On the basis of his results and my latest analysis I now think that the uncertainties presented here are significantly underestimated, and that the attribution of short term temperature trends is far from settled. There remains a lot of interesting work to be done on this subject.

0 0

Printable Version  |  Link to this page

Comments

Prev  1  2  3  

Comments 101 to 139 out of 139:

  1. 0 0
    Moderator Response: [RH] Fixed image.
  2. 0 0
    Moderator Response: [DB] Fixed image.
  3. O.K. see above graphs. Here is objection number one to Zhou and Tung. Foster/Rahmsdorf get 0.17 degrees C per decade. Lets check this with expectations. 1. C2/C1 = 2^ (temp/Tc.s.) 2. log base 2 (C2/C1) = temp/T c.s. 3. temp = T c.s. log base 2 (C2/C1)= T c.s. Ln (C2/C1)/ .693 [ .693 is natual log of 2] 4. Expand Ln (C2 / C1) around unity. Leading linear term is [(C2/C1)-1]. What is ( d/d time ) of [ (C2/C1) - 1 ] these days? maybe 2 ppm /year. Say {(384 - 280)/280} minus {(382 - 280)/280} ~ .3742 - .3643 = .007 . multiply by C.S. of 2, divide by .693 = .02 per year or 0.2 per decade. 5. Foster Rahmsdorf get about .17 degrees per decade, which is pretty close to what I get (0.20) for transient climate sensitivity of 2 degrees per doubling CO2 in the limit where we only need the first term in the expansion of the log around unity. By Zhou and Tung their AMO removal makes the increase per decade due to CO2 go down to less than half that which is the C.S. of CO2 alone no water vapor feedback!!!!! Spencer, et all are rejoicing!!! But if you do this right, as in my top graph, clearly everything fine with transient C.S. of about 2 degrees per doubling of CO2. Now go to second graph.....I think I recall there was recent work attributing the big 1940 peak partly to aerosols, somehow, but I also think I recall a phrase to the effect that they could not completely get all that peak this way. Then it is IMO most interesting that if you subtract expected CO2 , ENSO, solar, and volcanoes at the end of the day we are left with three peaks with about 70 year spacing? So maybe the "What's left over" does contain some AMO. And it looks like we are presently kind of near the top of a smaller such peak,and global temp increases are slowing, though clearly if analyzed correctly, the CO2 eventually takes over and we do have about the expected transient C.S.
    0 0
  4. Curiousd @102, once again, total CO2 forcing is not the same as total anthropogenic forcing. Taking one recent assessment of the anthropogenic forcings since 1850, Skeie et al (2011), we see that total anthropogenic forcings were negative till about 1880, and zero from then till about 1900. After that they rose till about 1940, before falling back to zero around 1970, rapidly rising there after (red dashed line): That is quite different from the pattern exhibited by CO2 forcing over the same period (dark blue line): Note that Skeie et al do not show the close approximation between CO2 forcing and total anthropogenic forcing in the 21st century that is a feature of the IPCC AR4 and GISS forcing data. The aerosol forcing is uncertain, so it is unsurprising that this disagreement should exist. More importantly, the close approximation of the two values in the early twenty first century in those data sets that show it is a coincidence only, and does not apply throughout the nineteenth or twentieth century. It follows that what you show as a residual of the temperature record minus ENSO, solar, volcanic, and anthropogenic forcings actually contains a large component consisting of the time varying difference between CO2 and total anthropogenic forcing. If you use all forcings (in this case from GISS, and with GISTEMP as the temperature record), the residual will be much smaller, and without the apparent pattern shown in your residual: (Kevin C shows a picture of anthropogenic forcings used here. I would appreciate it if he were to show his actual residuals as well.) Obviously exact results will depend on which set of forcing data you consider more accurate and use.
    0 0
  5. Yes, but I am doing the CO2 forcing correctly, no, and the transient C.S. due to CO2 I get is reasonable. Somehow Zhou and Tung are getting a wildly wrong answer, with a CO2 C.S. about half what it really is?
    0 0
  6. Tom: The natural-subtracted versions of the plot give a pretty good idea of the residuals... see here. The pure residuals will need me to implement csv output on that web app, it's on the to-do list, but very busy atm.
    0 0
  7. Practical point here: Forgetting about whether my graphs in 101 and 102 are "publishable", clearly not...there is the practical question of how to best present the case for significant AGW to a hostile (mostly deniers) audience. I still like these long term plots of temperature versus log CO2 ratio. I show in 103 above that the 2 degrees C.S. is consistent with Foster and Rahmsdorf and others in the recent limit in which the first term in the expansion of the log is all you need. If I do the same log plot with Berkeley Earth I get 3 degrees C.S., not 2 which makes sense as Berkeley Earth is land based. (I recognize that these C.S. values are a form of transient response, and that once Arctic sea ice melting really strikes the C.S. will probably increase?) IMO, if you need to make the case that AGW is real and dangerous to an audience, the argument that "The climate simulations say it is happening and dangerous, therefore it is" is an argument that does not hack it, even to scientifically trained, but non climate trained, scientists and educators. This AMO thing......I am now bound and determined to go through the publications with a fine toothed comb...is there real evidence for a historical 70 year cycle or not, I mean going back before there was any AGW at all? If there is such evidence we can't just say there is no AMO and will have to deal.
    0 0
  8. Tom Curtis and Kevin, Does the the lowest graph in Tom's post 104 still contain the CO2 forcing? In my case I have subtracted the CO2 forcing for my graph in 102 above. It is not obvious to me that you would not get something similar to my graph in 102 if the lowest graph in 104 still contains the CO2 forcing, but that was removed using my method.
    0 0
  9. I found "Atlantic Forcing of Persistent Drought in West Africa" by Shanahan, et al. in Science, 324, 2009. They did x-ray fluorescence of elements in lake sediments that they tie to temperature and drought, connecting the monsoon to the AMO. Anyway, going back 1000 years plus, it sure looks like they see something similar to our modern record in terms of these oscillations, but with a power spectrum that peaks at about 40 years, not 70. So, why should not I say....experimentally so far in my literature search the AMO is real? They also say there is a good correlation between their results and tree ring studies. I was expanding some of these plots on 500 year time scale with photoshop. I am going to systematically dig more of this stuff up. If there is an interst I will post some it.
    0 0
  10. I'm certainly interested. Detecting a 70 year cycle in a 130 year record is pretty much a fools errand, but for a 40 year cycle you might begin to have a chance (although you'd have to be supremely confident that you had correctly separated out the volcanic signal first). When Tamino argued against the AMO he got some serious scientists arguing for it. I'm afraid I don't know anything about it beyond that discussion though.
    0 0
  11. The right way to do the graph would be (temperature-enso) vs (forcing). The graph Tom showed was temperature response vs temperature. The temperature-enso term is the second graph in the post linked from my #106. I'll try and produce the numbers, but either I have to revive some R code or do a lot more work on the browser version.
    0 0
  12. curiousd @109, the wavelet analysis shown in Figure 2 b of Shanahan09 shows a 40 year period from 500 to 300 BC, then again from 100 to 200 AD, and again from 1000 to 1200 AD, and finally, just briefly around 1950 AD. I will leave aside my doubts that any mathematical analysis can show a genuine 40 year periodicity in a 10 year period. Instead I will focus on the very transient nature of this periodicity. Tamino has analysed similar attempts to find AMO periodicity in paleodata. His main point is that in picking out statistically significant periodicities, you need to allow for the fact that you are examining so many periods. Given essentially random fluctuations, if you examine enough periods you will find some during which, just by chance, the fluctuations will appear to be periodic with a given frequency. That is not significant unless the apparent periodicity is persistent. Ignoring the large number of periods examined is like making a hullabaloo over discovering a sequence of three dice rolls in a row where you roll a six and ignoring the fact that it occurred as part of a sequence of 100 throws over which the mean result was 3.5 and the rolls follow a poison distribution. It seems probable to me (although I have not, and cannot do the maths) that this very episodic 40 year periodicity will fall to the same criticism.
    0 0
  13. More proxy data. Available with no paywall is tree ring reconstruction and prediction results of Gray, et al in Geophysical Res. Letters 31, 1994, L12205; URL LINK 1. Calibrated tree rings versus known records during instrumental period, extended results into earlier times. Wind up with essentially a three dimensional graph (third dimension by color) between period, power, and year. 2. Most "power" appears between 40 and 70 years. Clearly, if real, this thing is no clean "70 yer cycle." One take home message, this. 3. The hair raising statistical unapproachability of such studies for the non expert is emphasized by the statement that they use a "multi-tapered method coherency spectra --- based on MTM analysis using red noise assumptions."
    0 0
    Moderator Response: [RH] Hot-linked URL
  14. Question for anyone. Can you recommend a land based only set of historical temperatures other than Berkeley Earth? Not that I think there is anything wrong with Berkely Earth. Its just that a second temperature set would be good as a check. Historical would mean back far enough that the trend is clearly not linear,because the next term in the log expansion shows up.
    0 0
  15. Do you have some reason to discount GISTEMP (GHSN stations) Land only index ?
    0 0
  16. PDO is not noise. El Nino is not noise. They are fundamental processes that define the system. They work in conjunction and they work in antagonism both to each other and other processes and oscillations. The Australian hotspell is noise. A random resonance that defines no fundamental recurring process or oscillation. For a hypothesis regarding thermohaline influence on PDO and ENSO please see: http://geosciencebigpicture.com/2013/01/26/loose-fire-hose-and-the-aborted-nino/
    0 0
  17. Yes, absolutely they are not noise. Weather is more akin to 'noise' in the statistical sense, although the statistical meaning of the term noise is somewhat context dependent - it can include unmodeled signals as well as measurement errors, however if those unmodeled signals are not noise-like they can invalidate your conclusions. So when people describe the F&R calculation as removing noise from the trend, that is certainly an oversimplification. The adjusted time series is the temperature series with the best approximation to the influence of the known natural oscillations subtracted out, which provides therefore provides a better estimate of the residual processes in the system.
    0 0
  18. Concerning YouTube postings, why not do guest postings over at WattsUpWithThat.com? According to Alexa they have significantly more traffic and are just slightly better educated than the audience here. Sure there are some hardcore nuts there but you don't just want to preach to the choir. I read all the climate sites and find there to be a very high level of discussion there compared to many sites.
    0 0
  19. trunkmonkey: As Kevin C wrote, the term "'noise'...can include unmodeled signals as well as measurement errors." But Kevin overgeneralized, because in many scientific and mathematical usages, the term "noise" is used to refer to any information that is not the particular information you are interested in, regardless of how systematic that undesired information is, and regardless of whether you know its source. So it is perfectly legitimate to refer to F&R's treatment as removing the noise since we are interested in the trend that is not due to those removed influences.
    0 0
  20. freddyv33, Skeptical Science conserns itself with actual discussions of the science of climate change...and the debunkings of the nonsense spewed about it. While WUWT has the latter (nonsense) in abundance, the utter lack of consistency in moderation practices there makes it impossible to have civil discussions about the science there. Unlike SKS, where the even enforcement of a strong comments policy serves to ensure that actual discussions of the science can happen, free from invective, slander, innuendo, unsupported assertion and character assassination in favor of promulgating false equivalence to support the ephemeral facade of "debate" and "sides". By your "preach to the choir" quip you nakedly reveal your ideological base.
    0 0
  21. Reply to scaddenp in 115: No I have no reason to discount GISTEMP (GHSN) land only index. And I thank you. As a retired professional physicist new to climate science, one of the many aspects I find bewildering is the "alphabet soup" of various temperature records. For anything new I delve into in the climate change area, I find it useful to check out my assumptions first at SKS. There is no one at my University who knows more than thing one about this topic, and therefore this site is for me an information godsend.
    0 0
  22. Enthalpy is transferred almost exclusively from the surface to the atmosphere.Oceans constitute 70% of the surface and oceans have continued to warm (albeit very unevenly)over the past 16 years.I suspect the residual shown in the graph above represents the baseline inevitable transfer of energy from a warming ocean to the atmosphere.If I get a minute I'll plot your residual against Reynolds "all ocean" SST's.
    0 0
  23. Trunkmonkey: Just to be sure I'm understanding you correctly - are you saying that the deviations of the 'human contribution' (or more correctly the ENSO/solar/volcanic-removed temperature) from linear (or more correctly from anthropogenic forced response) are due to variations in the rate of heat transfer to the deeper ocean? If so, then I think there is another mechanism to consider. A while back Kevin Trenberth was answering questions on a post here, and I asked about the nature of the unaccounted-for radiation imbalance since 2003. If I remember correctly he attributed part of the answer to anomalous cloud response during an ENSO event. In which case it could be that clouds can play a significant role on short timescales. I suspect these two mechanisms account for much of the deviation, but I wouldn't like to guess in what proportion.
    0 0
  24. By burning much midnight oil I got that Zhong - Tung devolcanized, de ensoed, data into excel and got so I could do a kind of manual regression, where I use the entire range of data and use the proper log function for the CO2 contribution from 1863 to present, and manually tweak weighting factors by hand for long periods of time. I used the Zhong - Tung version of the AMO index. I can do well enough by such fiddling to convince myself that the C.S. is no longer 1.9 something (as I get above on post 101 here by a simple log straight line fit), but by regressing this AMO is indeed a lower C.S., maybe 1.3. There is another definition of the AMO index I found on Wiki - "Oldenwhatever's methodology???" - that gives a much different AMO index and it looks to me like it would give a larger C.S. So would like to compare these results. But now I need something which has the computer, not me, systematically doing the prowling around in parameter space to give the smallest residual, then give a goodness of fit parameter. I want to buy such a program and learn to use it. Suggestions on what to purchase?
    0 0
  25. curiousd... You know that a large amount of climate work is done in the R programming language. Is that what you're looking for?
    0 0
  26. I was hoping for some kind of canned program such as Origin or Sigma Plot. I have Origin, and it does linear regressions, but I don't think it does regressions of non linear functions, though I am not sure. I have heard of something you could get that works with excel. I don't want to learn a new programming language, for sure....I mean I know Fortran, Basic, and assembler language for the old Digital Equipment 8 kilobit DEC, and the IBM 1600 that used card stacks! (That DEC pre dated the Wang Calculator, which was the size of a suitcase, had a NIXE tube output, and could..da te de da te dahh - compute sines and cosines).
    0 0
  27. gnuplot does nonlinear least squares fits sidd
    0 0
  28. curiousd, I'm not sure what you mean by "regressions of non linear functions". You can do linear regressions of non-linear functions with Excel; for polynomial fits there's even a shorthand notation (e.g. "=linest(yvalues,xvalues^{1,2,3},true,true)" entered as an array formula for functions of the form y = ax^3 + bx^2 + cx + d) and there is also one for logarithmic functions, but even without that you just need to create extra columns (e.g. a column for x, a column for x^2, a column for x^3, etc., or a column for ln(x)) and use those as parameters to linest. This would work with any package that allows you to specify the input data by hand. The important thing to note is that while the function itself may be nonlinear, the fitting of the coefficients (a,b,c,d in the polynominal example above) is by linear regression. Sorry if I misunderstood your point.
    0 0
  29. curiousd: Although it is neither elegant nor efficient, you can (at least sometimes) do arbitrary least-squares fits to somewhat complex equations and data sets with Excel (or any spreadsheet with an equation-solving routine). In the version of Excel that I have (rather old - knowing Microsoft, details probably change with each new version), it is call the Solver. I think it may be part of the "Analysis Toolpak Add-In". In it, you can get Excel to change one or more cells, under various constraints, to yield a particular result. In order to do the odd-ball regression you need to set up a) a column with the data to be fitted b) a column with the calculated values. This needs to use a formula that refers to cells for the value(s) you want to try to fit to. c) cells containing the "fitting" values (the ones I mentioned in b). d) a column that calculates the differences between the observations and calculated values e) a column that squares the differences (or you could do that directly without the column of unsquared differences) f) a cell that is the sum of the squares (from e)... You need to set up the solver to manipulate the cells in c), with the goal of minimizing the sum of the squares in f). Though not straight-forward, it is readily available, and great way to understand more about just what it means to do a "least-squares regression" with something other than linear fits.
    0 0
  30. Hi Rob in 129. That suggestion sounds right up my alley.I.E. I have a shot at being able to do this.
    0 0
  31. curiousd: A couple of other points I forgot to mention: - you'll find that Excel is more likely to get a decent answer if you have a reasonably good guess of the initial values of your fitting parameters. If it has problems, try a different initial value. You can also try different initial values to see if it still comes to the same solution (which increases confidence that it is a good solution). - the times when Excel will have problems finding a good solution are usually associated with fitting two (or more) parameters that are strongly correlated. In such a case, it doesn't make much difference which parameter is altered - they both have a very similar effect. The effects, therefore, can also be offsetting - e.g., make A a lot bigger, and B a lot smaller, but the overall result (on the sum of squares) is very small, so Excel starts to think that neither has much effect when in reality they both have a large effect.
    0 0
  32. To detect if fitting parameters are correlated, look at the off diag elements of the correlation matrix that gnuplot dumps after the fit converges. sidd
    0 0
  33. For a good explanation of nonlinear least squares, i like Bevington, "Data Reduction and Error Analysis for the Physical Sciences" It has a good FORTRAN implementation. As for others, I seem to recall that the Bevington version is superior to the one in Numerical Recipes, or used to be. I use the open source Gnu FORTRAN compiler sidd
    0 0
  34. Another thought about the effect of the AMO on C.S.....There is the article by Booth,et al in Nature Letter, vol 484, April 2012, pp 228 - 232. (Probably already appeared on SKS, though not sure..) They do a simulation which attributes about 70 % of that peak in the global temperature record in 1940s to the effect of Aerosols. For sure, if they are right, then the transient C.S. goes back up considerable.
    0 0
  35. Kevin C @ 123

    Not really what I was saying. Ocean enthalpy is all over the map from Trenberth's dropoff and "missing heat" to Levitus's steady rise. Finally found the time to plot Foster and Ramshorf against Levitus as I mentioned earlier to curiousd. Not conclusive but shows similar structure.

    http://geosciencebigpicture.com/2013/02/02/structural-sim…ocean-enthalpy/

    0 0
  36. Saw a story about this paper over at ScienceBlogs and I suspect it's on topic for this post.

    From the abstract:

    [1] Observations suggest that the optical depth of the stratospheric aerosol layer between 20 and 30 km has increased 4–10% per year since 2000, which is significant for Earth's climate. Contributions to this increase both from moderate volcanic eruptions and from enhanced coal burning in Asia have been suggested. Current observations are insufficient to attribute the contribution of the different sources. Here we use a global climate model coupled to an aerosol microphysical model to partition the contribution of each. We employ model runs that include the increases in anthropogenic sulfur dioxide (SO2) over Asia and the moderate volcanic explosive injections of SO2 observed from 2000 to 2010. Comparison of the model results to observations reveals that moderate volcanic eruptions, rather than anthropogenic influences, are the primary source of the observed increases in stratospheric aerosol.

     

    0 0
  37. The video in the post is "private". A mistake?

    0 0
  38. I'm not sure whether this is good or bad news for the faithful, but there will always be "short-term warming and cooling influences of volcanic eruptions, solar activity, and El Niño and La Niña events" so removing them is pointless.

    0 0
  39. Earthling, by talking of the "faithfull" you give the impression of being a troll, here only to provoke an intemperate response, without being interested in the scientific response to the point you have raised.  If that is not your intention, I suggest that you avoid such inflamatory terms in future posts.  If it it your intention, I suggest you find another blog where such behaviour is appreciated.

    The point of controlling for the effects of volcanic eruptions, solar activity and ENSO is to undestand what has cause changes in GMST in the past, not to be able to predict it in the future.  The point is to discover what part of the observed changes cannot be explained by volcanic eruptions, solar activity and ENSO.

    We know that there will be short term warming and cooling influences in the future, but if they do not have a long term secular trend, they will not be the cause of a long term trend in GMSTs, which is what climatologists are primarily interested in.

    0 0

Prev  1  2  3  

You need to be logged in to post a comment. Login via the left margin or if you're new, register here.



The Consensus Project Website

THE ESCALATOR

(free to republish)


© Copyright 2024 John Cook
Home | Translations | About Us | Privacy | Contact Us