Arguments
Recent Comments
Prev 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 Next
Comments 46201 to 46250:
-
CBDunkerson at 07:56 AM on 20 April 2013Climate's changed before
In conversation I actually enjoy this kind of misleading statement;
Denier: 'You know, climate has changed before.......'
Me: 'Good point! The link between past CO2 changes and past climate changes gives us a clear idea of what to expect with our current CO2 emissions.'
-
Tom Dayton at 06:19 AM on 20 April 2013Global Warming: Not Reversible, But Stoppable
Well, heck, you're right, Dana! Maybe Andy could add a link to that post to his post?
-
Robert Marston at 05:31 AM on 20 April 2013Global Warming: Not Reversible, But Stoppable
Hat tip?
Excellent article clairifying the science. I think the key issue as in the linked article and your above post is to make certain that people understand warming is mostly stoppable, but not reversible.
Moderator Response:[AS] Thanks for the link to that article, which I hadn't previously seen. I made it clickable.
-
dana1981 at 05:25 AM on 20 April 2013Global Warming: Not Reversible, But Stoppable
Tom @4 - sounds like you're asking for this post, no?
-
rockytom at 03:40 AM on 20 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
KR #8 To give credit where credit is due, the two different clocks analogy is based on Einstein's Special Theory of Relativity from the following paper "ON THE ELECTRODYNAMICS OF MOVING BODIES" By A. Einstein, June 30, 1905.
Moderator Response:[DB] Hyperlinked paper.
-
Bob Loblaw at 03:32 AM on 20 April 2013Climate's changed before
Phil:
Yet another "advantage" of not stating the implied argument is that it leaves the imagination of the listener free to come up with any number of arguments. I was going to phrase that as "possible arguments", but then I realized that they don't even need to be possible - at least in a real sense. As long as the listener thinks they are "possible", then the speaker has sown the seeds of doubt.
Twenty listeners can come up with 20 mutually-exclusive implied arguments, and as long as nobody goes into detail everyone is happy (except for the scientists and policy people that actually want to examine the logic, validity, and strength of the argument).
As you say, getting the person to explicitly state the argument, rather than leaving it implied, allows you to examine it properly. Unfortunately, you often end up playing whack-a-mole as the fake skeptic refuses to provide a proper argument, and keeps saying "that's not what I meant" as you cycle through the possibilities and explain why they are wrong.
-
Tom Dayton at 02:55 AM on 20 April 2013Global Warming: Not Reversible, But Stoppable
Thank you for this post! I've been needing to point to something like this.
Next we need a post that shows why the uncertainty about sensitivity is irrelevant to our need to take significant action immediately. In a business as usual scenario:
1. Here are five versions of graph C--one for each of the sensitivities 1.5, 2, 3, 4, 4.5.
2. The consequent temperatures in each of those five graphs are sufficiently bad that it is far more cost effective to prevent than to purely adapt.
I guess we first need a peer reviewed paper for item 1.
-
villabolo at 02:42 AM on 20 April 2013Global Warming: Not Reversible, But Stoppable
Is it not possible to use biochar on an industrial scale to reduce C02 levels? If it were to take a hundred years to reabsorb the CO2 to preindustrial levels it would be worth the effort.
-
shoyemore at 02:03 AM on 20 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Dave123,
I tried to address this in #21. Lewis claim is that his transformations of the data and the parameter space renders his prior distributions objective, and makes his results tightly constrained. I lack the experience to pass judgment, but I am sure somebody can. From earlier discussions, James Annan (who also uses Bayesian methods, as do many climate scientists) has previously said he sees nothing wrong with Lewis' methods. Perhaps the best person to look at it might be Andrew Gelman, a well-known Bayesian statistician at Columbia U, author of many books, who blogs here:
http://andrewgelman.com/
-
CBDunkerson at 01:57 AM on 20 April 2013Global Warming: Not Reversible, But Stoppable
Excellent article. Even people who have paid attention to the science often have trouble with these details, so it is good to have a clear summation. The only point I might have stressed more is that the bit about warming stopping when human CO2 emissions do is conditional on how high the temperature goes. Currently it is true. At some future higher temperature permafrost and/or methane hydrates may be releasing enough greenhouse gases that atmospheric levels and temperatures continue rising even without further human emissions. This is alluded to in the 'Wrinkles' section, but is one of the major reasons action is needed now. Otherwise, foolish people might reason that we can continue letting the temperature rise and 'just stop emitting when things get bad'. Even in the best case that would mean we'd be stuck with the 'bad' conditions for hundreds of years, but it could also put us past the point where carbon feedbacks kick in and things do keep getting worse no matter what we do.
-
From Peru at 01:38 AM on 20 April 2013Global Warming: Not Reversible, But Stoppable
Two things:
1)The link to the PDF is broken. Please help!
2) What about the "warming in the pipeline" caused not by the current global radiative imbalance, but by the masking of part of the greenhouse forcing by short-lived, man-made cooling aerosols(the so-called "global dimming")? They have masked nearly half of the CO2 forcing, and have very brief lifetimes, unlike CO2 and the other greehouse gases.
Moderator Response:[AS] 1) Could you let me know which link is broken? 2) The aerosol issue was discussed briefly in the "Wrinkles" sub-section. I understand that Matthews has a paper "in the pipeline" addressing aerosols and non-CO2 GHGs, but I haven't seen it.
[DB] The source paper PDF is here.
-
citizenschallenge at 00:41 AM on 20 April 2013Global Warming is Accelerating, but it's Still Groundhog Day at the Daily Mail
Thank you Rob for that excellent article.
Good enough to Repost:
http://whatsupwiththatwatts.blogspot.com/2013/04/examining-david-roses-no-warming-daily.html
Thank you SkepticalScience.com for your generous Reposting policy !
Peter
-
Dave123 at 00:02 AM on 20 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
I'm not following this closely, but the implication is that Nic Lewis is a known climate contrarian. This would make his exercise, laudable admittedly for peer review publication, an excercise in motivated reasoning. Now admittedly there are good reasons to try to find out how you can create low ECS estimates. Determining what ECS is most sensitive to would seem a sensible procedure. But the question of biases built into Bayesian priors, the ability to skew the results as a result of those choices coming from a prior advocacy position sets a Bayesian prior of its own: Will a contrarian cherry pick to get a desireable result? Wouldn't past history with Watts and crew suggest that the Bayesian prior for this would be a strong "Yes"? Will Bayesian priors selected by a contrarian be likely to face questioning by Climate Hawks? Undoubtedly the prior here is a strong yes also.
So, as part of any exposition on this, should we be looking to experimental determinations rather simply allowing people to make prior estimates? (When I retire I'm really going to have to take some refresher stat classes)
-
Kevin C at 21:42 PM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
OK, this question has finally puched me into what I've been meaning to do for a long time, and get my simple response function model online.
With this you can vary the strength of the different forcings and see how that affects the fit to 20thC climate, the TCR and the RCP projections, as well as a few other things. You can also see the effect of switching between GISS and Potsdam forcings.
While on defaults the model gives a very similar result to CMIP-5 projections, that proves nothing. Try switching to the Potsdam forcings and you get a smaller response. Try doubling the Black Carbon term (in line with the recent study) and you get a yet smaller term. The main take-home is that while current forcing can explain 20thC climate, the forcing are sufficiently uncertain that this doesn't allow us to very significantly constrain future climate change.
A few notes:
1. I don't think the Potsdam forcings include the 2nd aerosol indirect effect. You can try multiplying it by 1.5 to compensate.
2. Hansen's aerosol forcing is larger than most.
3. I've rushed this out, there may well be mistakes.
4. Start date is fixed at 1880, so not directly comparable.
See what you can do with it.
-
Phil at 19:51 PM on 19 April 2013Climate's changed before
I think one of the "strengths" of this argument (from the Denialist point of view) is that there are actually two possible implied arguments. One is (as others have pointed out above) is
Climate has changed before [naturally and so the current instance of climate change must be natural too]
The second is this
Climate has changed before [and "the planet" survived that and so will survive the current climate change, irrespective of how its caused]
I've put the incorrect implication in []. The first case is a failure of logic, the second case mis-identifies both the risk of climate change, and what "taking action" on GHG emissions seeks to accomplish.
So when responding to the simple statement "Climate's changed before", you can't be sure which one of these two arguments is being implied, if you rebute one the other can still be thought to be valid, if you try and rebute both it can get very involved.
Perhaps this helps explains why, as an argument, it is so popular with contrarians.
A final thought: It seems to me that as soon as the implied arguments are stated, the flaws in them are very obvious, and this, together with the fact that the prima-face statement is true and uncontestible, explains why the implied arguments are always just that - implied - which is why SKS haven't found a quote to adorn this acticle that exposes the real myth(s)
-
Dikran Marsupial at 17:29 PM on 19 April 2013Models are unreliable
@bouke The current policy decision to do nothing [and it is a policy decision] is not based on the projections of sea ice extent, so you did dodge the question.
I'm sorry, I have better things to do with my time than engage in evasive rhetorical discussions. If you want to demonstrate that is not what you are trying to do, then give a cogent explanation of how sea ice extent projections support the policy decision to do nothing. The models clearly indicate that sea ice extent is diminishing and that we will see effectively ice free summers on a timescale relevant to human lives, so even though they underestimate the rate of ice loss, the clearly argue that we should do something, not do nothing.
-
DSL at 13:47 PM on 19 April 2013Climate's changed before
So the answer to your first question is "What's wrong is that Lindzen actually wrote it."
In other words, Lindzen thinks you're stupid or at the very least simply ignorant, and willingly so. If you said to him "Duh! What's your point?" he'd probably bluster a little and then show you some really cool graphs.
-
DSL at 13:43 PM on 19 April 2013Climate's changed before
No. There's nothing wrong with the information itself. The myth occurs when the information is voiced--made public. I'd bet you that in 99.99% of the cases in which those facts are strung together and published, the person doing the stringing is attempting to make an implied argument that since climate has changed before without rapid accumulation of atmospheric CO2, and therefore the current climate change is not the result of human CO2 emissions. I've engaged with at least 400-500 people on comment streams using this very implied argument.
-
scaddenp at 13:41 PM on 19 April 2013Climate's changed before
Mark, "the climate has changed" before is mostly used to imply that humans arent responsible for current warming and/or that past changes are unforced. As a skeptic argument, it is framed as an excuse to do nothing. In the source where the quote came from, Lindzen tries to touch both bases, along with an a priori assumption that climate sensitivity is low. He has been trying to find some plausible reason for low sensitivity without much success so far. If you want to get down to real myths, then I guess you have to look at what logical step follows from "climate has changed before" but for many, this somehow seems to be enough.
-
Tom Curtis at 13:32 PM on 19 April 2013Global Warming is Accelerating, but it's Still Groundhog Day at the Daily Mail
Karellen @21,
the key point is that a confidence bound is a line that the observed value is expected to fall below a certain percentage of the time. The confidence interval is then the interval between the lower and upper confidence bounds.
Thus, the way to read the uncertainty interval is that, for the lower bound (5%), the observed value is expected to fall below that value 5% of the time. The upper (95%) bound is the limit such that the observed value is expected to fall below that value 95% of the time. Similarly you can have 25% and 75% bounds, which the observed value would be expected to fall below 25% and 75% of the time respectively. The predicted value (the central line) is in fact just the 50% confidence bound, ie, the line that the observed value is expected to fall below 50% of the time. You will also sometimes come across 33% and 66% confidence intervals, which the observed value is expected to fall below respectively 33% and 66% of the time.
When the observed values minus the predicted value, which is called the residual, is normally distributed the confidence interval is often expressed in terms of Standard Deviations (SD), or Standard Error (SE). With a normally distribued residual, the two SD confidence interval is approximately a 95% confidence interval, ie, the lower bound is such that the observed value is expected to fall below it 2.5% of the time and the upper value is expected to fall below it 97.5% of the time. For the one Standard Deviation (or Standard Error) confidence intervals, the lower bound is such that the observed value is expected to fall below it approximately 16% of the time, while the observed value is expected to fall below the upper bound approximetely 84% of the time. These bounds are often designated by the mathematical symbol for a Standard Deviation, ie, a lower case greek letter sigma (σ) which is eqivalent to "s" in the English alphabet. Likewise it will sometime be reffered to simply as one or two "sigma".
-
Mark Bahner at 13:02 PM on 19 April 2013Climate's changed before
Mark @345 - there's nothing directly wrong with 'the skeptic argument' as articulated by Lindzen here.
So the "#1 global warming myth" is not really a myth at all...?
-
engineer8516 at 11:08 AM on 19 April 2013Global Warming is Accelerating, but it's Still Groundhog Day at the Daily Mail
I should add that the 5-95% uncertainty range is the dark red region and lighter red region combined.
-
engineer8516 at 11:06 AM on 19 April 2013Global Warming is Accelerating, but it's Still Groundhog Day at the Daily Mail
@ karellen
The 25-75% uncertainty range, means that the actual temperature is expected to be within that range 50% of the time (75-25=50). The 5-95% uncertainty range means that the actual temperature is expected to be within that range 90% of the time (95-5=90).
Another way to think about it is that the actual temperature has a 50% probability of landing in the dark red region and a 90% probability of landing in the red or lighter red region.
-
Tristan at 10:57 AM on 19 April 2013The anthropogenic global warming rate: Is it steady for the last 100 years?
Nice work DS
I couldn't even show her the difference between a negative temperature trend and 'statistically significant cooling'.
-
Don9000 at 10:38 AM on 19 April 2013Global Warming is Accelerating, but it's Still Groundhog Day at the Daily Mail
Karellen@21
My layman's understanding is that the uncertainty ranges are also called "confidence intervals." The 5-95% range is saying that the value being tracked will fall within the wider band 95% of the time. Similarly, the 25-75% uncertainty range is saying that there is a 75% confidence that the value will fall within the narrower band. I'm sure someone with more technical expertise will chime in with a better explanation. That said, between you and me and everyone else, I would like to see the Climate Science Glossary under the Resources tab get a bit better about explaining this kind of thing. -
Karellen at 09:36 AM on 19 April 2013Global Warming is Accelerating, but it's Still Groundhog Day at the Daily Mail
Sorry if I'm being stupid here, but I don't understand how the uncertainty ranges on figure 3 are supposed to work.
At the blue line, the average value looks to be (approx) +0.6°, with the "25-75% uncertainty range" going from +0.5 to +0.7, and "5-95% uncertainty range" going from +0.3 to +0.9.
So, the outer band is both more uncertain (95% > 75%) *and* less uncertain (5% < 25%) than the inner band? That doesn't make any sense to me.
Surely if your average value is +0.6, you might be 50% certain that the value will be between +0.5 and +0.7, and 90% certain that the value will be between +0.3 and +0.9. But that would make your inner band "50-100% uncertain" (0-50% certain), not 25-75%, and your outer band "10-50% uncertain" (50-90% certain), not 5-95%.
I mean, I'm pretty sure I understand what the graph is telling me. I think the graph, and the bands, make sense. It's just the labelling that I can't figure out. The Y axis is measuring temperature, not certainty. Certainty is an expected deviation from a value, with zero uncertainty being zero deviation. Zero uncertainty is not as far as you can go below the expected value, it *is* the expected value.
Isn't it?
Clearly, everyone else has no problem with this graph, and is reading it differently from me, but presumably you're all reading it the same way as each other. Therefore, the most obvious conclusion is that I'm the one reading it wrong. But how?
Can someone please tell me which angle I need to tilt my head at, in order to look at this in the right way so it actually makes sense for me?
-
K.a.r.S.t.e.N at 07:52 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Thanks Tom!
In order to be more transparent, I uploaded the principal figure from Lewis 2013 (Fig 3). It shows the posterior PDFs in comparison with Forest et al. 2006 (which I erroneously referred to as Forster et al. 2006 in #14):
Figure caption: "Marginal posterior PDFs for the three climate system parameters by diagnostics employed, method used and either upper-air diagnostic EOF truncation parameter (panels a, b and c) or surface diagnostic EOF truncation parameter (panels d, e and f). Panels a, b and c use the F06 diagnostics; panels d, e, and f use revised surface and deep-ocean diagnostics, and no upper-air diagnostic. Panels show marginal posterior PDFs for, from the top, Seq (ECS), Kv (effective vertical deep-ocean diffusivity) and Faer (total aerosol forcing). In panels a, b and c, the solid red lines show marginal PDFs using the new objective Bayesian method and kua=12 (number of EOFs for upper air (ua) and surface (sfc) diagnostics), while marginal PDFs using the F06 method are shown at kua=12 (solid green lines) and at kua=14, as used in F06, (dashed blue lines) , and the dotted black lines show the published F06 PDFs. In panels d, e and f, marginal PDFs using the new objective Bayesian method are shown with ksfc=16 (solid red lines) and ksfc=14 (dotted blue lines), while corresponding marginal PDFs using the F06 method are shown with respectively solid green and dotted black lines. The box plots indicate boundaries, to the nearest fine grid value, for the percentiles 5–95 (vertical bar at ends), 25–75 (box-ends), and 50 (vertical bar in box). Parameter ranges are given by plot boundaries."
His prior assumptions for Seq, Kv and Faer are: 0.5 to 15 K, 0 to 64 cm2/s, -1.5 to 0.5 W/m2.
He explicitly states that: "Faer represents net forcing (direct and indirect) during the 1980s relative to pre-1860 levels, and implicitly includes omitted forcings with patterns similar to those of sulfate aerosols." As aerosol forcing seems the most constrained parameter, I wonder what a small shift in the prior to a more plausible range of -1.5 to 0.0 W/m2 would do to the final results. Given that the F06 aerosol forcing mode isn't too far off the one in Lewis 2013, it seems that the method is very sensitive to the prior choice. I might very well be wrong in this regard.
Note that there the paper comes with Supporting Information, which isn't available yet.
-
BBD at 07:03 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Much to ponder.
Without going so far as to suggest that this study is intended as a distratction, there is no doubt that it will be used as such in the 'debate'.
Perhaps this bears repeating:
Albatross #9 gives the best synopsis in my view - the meta-estimate of ECS is still stubbornly in the 2.5-3C range, and I somehow doubt if we will even pin it down more accurately. Deniers are just arguing the decimal points, as Professor Scott Denning told the Heartland Institute in 2011, but of course they did not listen then either.
-
bouke at 07:02 AM on 19 April 2013Models are unreliable
@Dikran
- You asked "Exactly what are the policy decisions that are being made based on predictions of SSIE?".
- I answered "The current policy decision is to do nothing".
- You stated "it is a pity that you dodged my question".
- I stated "perhaps we disagree on what constitutes a policy decision"
- You stated "I agree that "do nothing" is a policy decision, nothing I wrote suggests otherwise."
The only way I could reconcile (3) with (2) was if you didn't consider doing nothing a policy decision. That's where (4) came from.
Essentially the point is that for Arctic sea ice extent, the observations tell you what is happening, the models predict that the Arctic sea ice will last longer than the observations suggest to the extent that we know that their projections are wrong, and hence nobody should be taking them sufficiently seriously as a basis for policy.
Now you think that they do have an effect on policy, but do you have any evidence to support that?
The entire purpose of the IPCC is to inform policymakers. The IPCC's AR5 draft mostly discusses models and their projections. If you think that the projections of those models are wrong, why are these projections in AR5?
Anyway, this is the last post by me on this matter. I will read answers but won't post more. Except for CBDunkersons link, the discussion here hasn't been very useful for me and I doubt that will change.
-
Tom Curtis at 06:40 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Dana @22 & 23, the mode (unlike other parameters) is easilly estimated from the PDF shown in Fig 6 a. Using a pixel count, I can confirm the estimate of 1.6 C for the mode of the main analysis.
Contrary to your main article, this is not "simply a misrepresentation" of Aldrin et al. It is misleading in that he does not point out the difference between the modal value and the central estimate commonly quoted for the IPCC, and that they are not directly comparable. However, it is not incorrect on any matter of fact. That is, the modal value of Aldrin et al is the value Lewis quotes, and he does specify that it is the modal value he is citing. Consequently, as it stands, your section on "Misrepresenting Aldrin" is more misleading than is Lewis' citation of Aldrin.
Further, Aldrin et al do not exclude cloud and aerosol effects in general. The nine forcings included in the main analysis are specified as:
"long-lived greenhouse gases (CO2, CH4, N2O and halocarbons), tropospheric ozone, stratospheric ozone, stratospheric H2O, the direct aerosol effect, the cloud albedo effect (indirect aerosol effect), surface albedo because of land use changes, solar radiation and volcanoes".
(My emphasis)
Aldrin et al do exclude cloud lifetime effects from the main analysis and other cloud effects (ie, other than the aerosol indirect effect and cloud life time effects) from all analyses. As these are not defined as forcings in the IPCC AR4, they may or may not be included in other estimates of climate sensitivity. De facto, they will be included in purely observational estimates, but their inclusion in model based estimates (or hybrid estimates such as Lewis 2013) may well be doubtfull.
For what it is worth, by eyeball, inclusion of a cloud lifetime effect of -0.25 W/m^2 lifts Aldrin's modal estimate of climate sensitivity to about 1.8 C, and including a feedback of -0.5 W/m^2 lifts it to about 2.2 C. Aldrin et al cite confidence interval for the effect of -0 to -1.5 C from Isaksen et al (2009), and from consulting the figures for Isaksen et al (it is pay walled), the mean estimate is about -0.3 W/m^2.
-
dana1981 at 05:39 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
I should also note again that he's only looking at the Aldrin sensitivity distribution that excludes cloud and aerosol effects when he uses that 1.6°C mode figure.
-
dana1981 at 05:36 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Tom @16 - to be precise, Lewis is quoting what he thinks the mode of the Aldrin PDF is. I'm fairly sure that Aldrin never reports that mode. That being said, the abstract does say he's comparing the modes of the two studies.
-
shoyemore at 05:24 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Tom Curtis, #17,
My Master's thesis supervisor always emphasised that in Bayesian Methods, the parameters took on probabilities, and the data was regarded as a static "given", somewhat the reverse of Frequentist Methods. Lewis' work does seem to be a departure from that, and possibly has philosphical implications?
The key moves seem to be "whitening" the observations by an "optimal fingerprint transformation", using Bayes to derive a PDF for the data. Another move is a "dimensionally reducing change of variables" in the parameter space. I am just summarising the abstract, but this is claimed to result in the tightly constrained estimates. The suspicion is that the "tight constraint" is the subjective motivation rather than the objective result.
Unfortunately, my own knowledge is insufficient to properly critique his methods, but I hope somebody does. After all, the statisticians flocked around to "help" Michael Mann! If they can do it for a physicist, they can do it for a retired businessman.
Albatross #9 gives the best synopsis in my view - the meta-estimate of ECS is still stubbornly in the 2.5-3C range, and I somehow doubt if we will even pin it down more accurately. Deniers are just arguing the decimal points, as Professor Scott Denning told the Heartland Institute in 2011, but of course they did not listen then either.
-
Tom Curtis at 04:49 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
BBD @19, I also haven't read the paper. However, based on the abstract, the main result, ie, that which is discussed first, is the change in methodology and its effect on the climate sensitivity estimate. Of course, Lewis may want to put his revised low CS estimate up front in discussion, but that should not be assumed to be the reason the paper was published.
-
BBD at 04:32 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Tom Curtis # 17
By happy coincidence I have just said the exact same thing as your first paragraph in a comment elsewhere ;-)
I haven't read the paper (only the blog posts NL has made elsewhere) but I suspect that what you go on to say is correct. Must reserve judgement for now, though.
-
BBD at 04:28 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Moderator [JH]
Unless you have just added the defintion of PDF to your glossary, it is already there. The term is underscored in text and the definition appears in a pop-up when the cursor is passed over it.
Moderator Response:[JH] I may have been viewing it on the rolling summary of all comments. It appears that our nifty glossary does not do its magic in that venue.
-
Tom Curtis at 04:25 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
BBD @15, the main result of Lewis 2013 is the different approach to bayesian estimates of climate sensitivity, ie, the application of uniform priors to the observations rather than to the parameters. Regardless of the merits of the final estimate of CS, if that method is not transparently flawed it should be placed before the scientific community so that its merits can be discussed and analyzed.
As to the estimate of climate sensitiivty itself, it uses the output of a single 2 dimensional model to quantify likilihood of the three paramaters. Consequently it is no better than the accuraccy of that model. Further, the final quoted value depends on the addition of a number of La Nina effected years to the data. Given the use of a 2-D climate model, there is no way the effects of El Nino on temperature are modelled, and in consequence the resultant estimate of climate sensitivity is certainly too low. The 2.4 K mode quoted for the change of method only is therefore more likely to be accurate than Lewis' final result. Ideally, given that the model does not account for ENSO, it should be used against ENSO adjusted temperature data, such as that from Foster and Rahmstorf. Doing so will likely give a mode between 1.6 and 2.4 K, but closer to the later than the former; and is likely to give a mean value (the value cited by the IPCC and therefore the correct comparison with the IPCC range) slightly higher than that.
-
Tom Curtis at 04:07 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Dana, Aldrin et al quote a mean CS of 2 C. Lewis quotes the mode of the Aldrin PDF on climate sensitivity as 1.6 C, which is correct. The confusion arises because he is citing a different value.
Lewis, in fact, has a penchant for quoting the mode of climate sensitivity analyses because it is, as he says, the maximum likilihood result. Personally I do not think it is the best value to quote. Rather the median (where it can be determined), ie, the 50/50 mark is far more informative if you must quote a single number. Of course, with a long tailed PDF, the mode will always be less than both the median and the mean of the PDF. Without going so far as to say that is Lewis' reason for preferring it, that does make it a tremendously convenient number for "skeptics" to quote.
-
BBD at 04:01 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
K.a.r.S.t.e.n
# 10; # 14
I remember your discussion with NL and PaulS at James Annan's very well. And most illuminating it was too. What puzzles me (troubles me?) about this is that this study has been accepted by J. Climate (not say, E & E) after an apparently long period under review and yet these questions apparently remain outstanding. -
Dumb Scientist at 03:50 AM on 19 April 2013The anthropogenic global warming rate: Is it steady for the last 100 years?
Due to my inexperience, I have found it difficult to answer individual questions, mostly of them are technical in nature. I have tried to explain the technical details, but that did not seem to work. [KK Tung]
It didn't work because your technical details didn't address my point, as many have noted above. That's why I distilled my point into two yes/no questions. I look forward to the educational answers you've undoubtedly provided in your second post.
There is no obviously right or wrong answers; this is always the case when the science is unsettled---when the science is settled I will have to move to another field.
Science can be (loosely) defined as the search for answers which are less wrong than previous answers. That slogan is meaningless because all science has uncertainties.
One argues for the reasonableness of the assumption using evidence and physical mechanisms, and then proceeds to deduce what that assumption will lead to as consequences. In scientist publications, one always lists the assumption clearly so that others could refute it. We should argue whether the assumption is supported by the available evidence or not. But claiming that the argument is circular simply based on the technical fact that the consequence arose from the assumption is missing the bulk of the arguments in our paper leading to that assumption.
It's true that all science is based on assumptions, such as conservation of energy. But no study based on the assumption of energy conservation would conclude that energy is conserved. That would be a circular argument.
You regressed global surface temperatures against the AMO in order to determine anthropogenic warming. Because the AMO is simply linearly detrended N. Atlantic SST, this procedure would only be correct if AGW is linear. Otherwise you'd be subtracting AGW signal, sweeping some AGW into a box you've labelled "natural" called the AMO. So you're assuming that AGW is linear, and you're also concluding that AGW is linear.
I suspect a lot of that may have to do with the fact that our paper is behind a "pay wall", so that many posting here may not have read more than the abstract. I have posted a free link to the entire paper in the first few lines of my post. A correction: The link to our PNAS paper was deleted in this first post. I hope it survives in my second post, where it is provided again.
Actually, the first link in my article was a free link to the entire paper Tung and Zhou 2013.
Dr. Tung, I'd also like to thank you for your participation here. Right now I'm also trying to explain Antarctic ice mass balance at Jo Nova's, and the disappointing responses made me appreciate this civilized discussion even more.
-
K.a.r.S.t.e.N at 03:14 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
It's not that Nic wouldn't have been aware of the aerosol forcing issue. PaulS and myself, we had an interesting discussion with him over at James Place. His reluctancy to accept the opinion of other (more informed) people on the subject struck me as quite bold. Realizing that he didn't even discuss the uncertainties regarding the aerosol prior in the paper, is something which I consider a huge disappointment. Had he applied a more objective areosol forcing estimate (one may think of an "expert prior"), he wouldn't have come up with such a low ECS (or effective CS for that matter) number. I think he missed the opportunity to demonstrate that his method is superior to Forster et al. 2006 (who were using uniform priors). Too bad ...
-
Albatross at 02:55 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Dana @12,
Agreed. The misrepresentation of Aldrin et al by Lewis is egregious and should have been caught. The journal can still correct the paper (the online version is an "early release" and not necessarily the final version), or failing that Lewis should publish a corrigendum.
-
dana1981 at 02:47 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Interesting point, Karsten. That's an issue with a Bayesian approach - it's important to use accurate informed priors.
It bothers me a bit that the misrepresentation of Aldrin's results and calling this an estimate of equilibrium sensitivity weren't caught during the peer-review process, or at least weren't fixed in the final version. It's possible that the issues were raised in the extensive review comments but Lewis disagreed and the editor allowed the errors to be published. Whatever the reason, those are significant issues with the paper, in my opinion. He should at least have discussed the relationship between effective and equilibrium sensitivity. As it stands, the paper simply says it's estimating equilibrium sensitivity, which isn't accurate. And the Aldrin error being in the abstract is unfortunate too.
-
KK Tung at 02:41 AM on 19 April 2013The anthropogenic global warming rate: Is it steady for the last 100 years?
A correction: The link to our PNAS paper was deleted in this first post. I hope it survives in my second post, where it is provided again.
-
KK Tung at 02:37 AM on 19 April 2013The anthropogenic global warming rate: Is it steady for the last 100 years?
Replying to KR: Thank you for the comments. I appreciated the opportunity to discuss the various issues involved.
Due to my inexperience, I have found it difficult to answer individual questions, mostly of them are technical in nature. I have tried to explain the technical details, but that did not seem to work. Now I appreciate why Gavin Schmidt in Realclimate.org won a prize for communication.
A general comment not related to any post in particular: There is no obviously right or wrong answers; this is always the case when the science is unsettled---when the science is settled I will have to move to another field. One argues for the reasonableness of the assumption using evidence and physical mechanisms, and then proceeds to deduce what that assumption will lead to as consequences. In scientist publications, one always lists the assumption clearly so that others could refute it. We should argue whether the assumption is supported by the available evidence or not. But claiming that the argument is circular simply based on the technical fact that the consequence arose from the assumption is missing the bulk of the arguments in our paper leading to that assumption. I suspect a lot of that may have to do with the fact that our paper is behind a "pay wall", so that many posting here may not have read more than the abstract. I have posted a free link to the entire paper in the first few lines of my post.
-
Albatross at 02:26 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Hi Karsten @ 10,
"Correspondingly, he got a median posterior of -0.3 W/m2 with his method. It is certainly a underestimate, no doubt about that. So I don't put much credibility into his most likely ECS values, as they are an underestimate too (perhaps strongly so)."
You have just raised a huge red flag. His value for aerosol forcing is undoubtably wrong and as such way too small.
Remember, peer review is not perfect. Also, this paper was published in a reputable journal, but even they err from time-to-time. It is my understanding from what Lewis has been saying that two of the reviewers were very tough. They allegedly even brought in a forth reviewer to help settle things.
Regardless, the true test of Lewis starts when it appears in press. Somehow I very much doubt that it will stand the test of time. With that all said, kudos to Lewis for submitting his work for peer review; it also shows that there is no censoring going on in journals as "skeptics" claim.
-
K.a.r.S.t.e.N at 02:16 AM on 19 April 2013The anthropogenic global warming rate: Is it steady for the last 100 years?
Dr Tung, although one might receive my above comment (#30) as destructive (if not offensive), I'd like to stress that I merely expressed my personal opinion ... unfortunately in a rather dismissive and brisk tone for which I humbly apologize. I am just too tired of all these unconvincing attempts to blame the AMO for everything. As I said, I am looking forward to the second post (which is certainly gonna be up soon).
-
K.a.r.S.t.e.N at 01:58 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
Thanks for putting this together, Dana! Nic Lewis seems unwilling to accept the current (AR5 SOD) aerosol forcing estimate. The total anthropogenic ARF (all direct + indirect effects in terms of adjusted TOA forcing) is -0.9W/m2 since 1750. The most likely range is –1.5 to –0.3 W/m2. Scaling it to 1860 (Lewis' start year) might reduce the aerosol forcing slightly to -0.7 or -0.8 W/m2. However, his prior for the total ARF distribution lies between +0.5 and -1.5 W/m2. Correspondingly, he got a median posterior of -0.3 W/m2 with his method. It is certainly a underestimate, no doubt about that. So I don't put much credibility into his most likely ECS values, as they are an underestimate too (perhaps strongly so). He put an aerosol forcing which he believed to be true and he got a low ECS which he (apparantly) wishes to be true. What a surprise! I am underwhelmed ...
I really don't understand why no one has bothered to check the assumptions which went into his approach. Why hasn't the unwarrantedly low aerosol forcing prior not been picked up during peer review? I am a bit surprised to see what makes it through peer review as of late. Perhaps just stronger awareness or pure coincidence.
-
Albatross at 01:47 AM on 19 April 2013Climate Sensitivity Single Study Syndrome, Nic Lewis Edition
While recent research has suggested that an equilibrium climate sensitivity (ECS) above +4.5 C is highly unlikely, a meta analysis of all the evidence continues to stubbornly show a best estimate of ECS to be near +3 C.
This is pretty worrisome given that on our current trajectory we will probably quadruple CO2 levels before the end of the century (remember that estimates of ECS are traditionally for "only" a doubling of CO2).
Yet, "skeptics" are trying to reframe the argument by using the Lewis paper (which IMHO looks to have yielded an estimate of transient sensitivity rather than for ECS). "Skeptics" seem to think that a few studies out of literally dozens suggesting a relatively low ECS means that humanity should bargain/bet on ECS being < 2 C so that we can continue on with business as usual. This is both risky and irresponsible. Engineers design components to fail at the upper end of the range, not the most likely value, nevermind the lower range.
I doubt many of us would willingly board an aircraft or spaceship that we have been told meets only the minimum safety criteria. Well, planet earth is our one and only spaceship and we would be prudent to not bet on ECS being < 2 C, especially given the preponderence of evidence to the contrary.
-
Jim Hunt at 00:59 AM on 19 April 2013Global Warming is Accelerating, but it's Still Groundhog Day at the Daily Mail
For further information on how The Mail "con" their readers on climate change please see:
http://econnexus.org/the-strange-tale-of-the-mail-and-the-snow-dragon/
Amongst other things, they "conveniently omit" relevant comments
Prev 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 Next
