Arguments
Recent Comments
Prev 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 Next
Comments 25301 to 25350:
-
mancan18 at 08:11 AM on 25 February 2016Fossil fuel funded report denies the expert global warming consensus
One Planet Only Forever @7 and Jeff T @6
To evaluate a source of information you not only have to understand who the author is, but you need to pay careful attention to the language used, the style of writing and the material used to support the argument. Now, I suspect that some of the Heartland material is not written to seriously participate in the scientific debate. It is written to give a scientific veneer to the reference material used in non-scientific articles written by those who oppose the climate change argument in the popular media. These articles make it easier for denier motivated journalists who have a non-scientific background to make simple dismissive climate change denier arguments in the popular media. I have seen articles, particularly in the Murdoch press and some conservative journals like Quadrant, that use the word "discredited" when referring to John Cooke's Consensus Project. When reading opinion pieces in the Australian Press by the likes of Andrew Bolt, Miranda Devine, and Piers Ackerman, they often use a combination of words that are dismissive of the whole scientific basis for Climate Change without providing any real evidence. When they do provide "evidence", which is not very often, it is from institutions like Heartland and a few scientists attached to Australia's Institute of Public Affairs. I am sure that it is the same with the popular media in Canada, the US and the UK. It is very easy to make dismissive statements in this debate in the popular media. It is very difficult and not so easy for scientists and like minded journalists to refute those dismissive statements in the popular media. It requires the reader to have some understanding of scientific reasoning. Of course in scientific circles, where scientists argue in a scientific manner, dismissive arguments are not so easily made. They require evidence and logical argument. In science, it is the strength of the scientific argument and supporting evidence that wins the day, not necessarily who the author is.
-
sailingfree at 07:56 AM on 25 February 2016A Response to the “Data or Dogma?” hearing
6. Sailingfree I get why the troposhere for a denier:
The Earth's surface is warming. The stratosphere is cooling.
So somewhere in the troposphere, by interpolation, is a "Goldilocks Layer" with a level temperature graph. So "No warming since forever!" can be truthfully proclaimed.
-
Earth is warming 50x faster than when it comes out of an ice age
wili - In that Hansen and Sato 2011 they point out that we shouldn't ignore outlier estimates, noting that "...a 10-year doubling time was plausible, pointing out that such a doubling time from a 1mm per year ice sheet contribution to sea level in the decade 2005-2015 would lead to a cumulative 5 meter rise by 2095."
Note that they don't claim that is likely, but that it is plausible, and in the next paragraph note that "...more plausible but still accelerated conditions could lead to sea level rise of 80 cm by 2100." Note "more plausible" in that sentence. They then go on to discuss strong negative feedbacks as well.
The point they were making in that paper was that we can't at this time exclude extremely high sea level rise, and that more study is needed. They weren't claiming 5m by 2100 was the likely outcome, just that it wasn't impossible - despite multiple out of context claims by climate denialists, which seem IMO to be poor attempts to discredit Hansen.
-
wili at 07:02 AM on 25 February 2016Earth is warming 50x faster than when it comes out of an ice age
Thanks, scaddenp. I had seen the RC piece, but not the other one you linked to.
What do you think of Hansen and Sato's claims that we could see a doubling rate of ten years over this century with a resultant ~5 meter rise by 2095?
www.columbia.edu/~jeh1/mailings/2011/20110118_MilankovicPaper.pdf
pages 15 -16
-
scaddenp at 06:40 AM on 25 February 2016Earth is warming 50x faster than when it comes out of an ice age
While warming is faster, there is a lot less ice to melt. It is certainly an interesting question. You might like to look at new paper on it here. Some discussion about it over at Realclimate.
-
wili at 06:34 AM on 25 February 2016Earth is warming 50x faster than when it comes out of an ice age
So, with:
1) the meltwater pulse that happened about 8000 ybp raising sea levels 1-2 meters per century, people.rses.anu.edu.au/lambeck_k/pdf/262.pdf
2) and the Meltwater Pulse 1A at the end of the last ice age 13,800 years ago, raising sea levels over 5 meters per century,
3) and with current warming happening 50 times faster than during those events...
What rates of sea level rise should we be expecting, exactly. And how soon?
-
scaddenp at 06:30 AM on 25 February 2016Models are unreliable
The whole point of current focus on climate models is to inform policy with regard to emission strategies. That is under human control, not something you "forecast". We are only peripherally interested in effects on climate from variations in natural forcing because a/ they are either short term or small, and b/ there is nothing we can do about them. We can however control emissions and we really do want to know whether models will accurate predict future under various emission scenarios. For evaluating an older models, you can simply rerun with actual forcings (without changing the model code) and compare to what we actually got.
The hard bit in communication is having public understand the difference between a physical model and statistical model in terms of value in hindcasting.
-
Nick Palmer at 02:52 AM on 25 February 2016The Uncertainty Handbook: Download and Translations
I confirm SteveS's observations, but on a PC using Adobe Acrobat, Windows 10. I can't say I like the main font they use in the text. Several letters, eg a,r,v,u, x, are smaller/less tall than the others and I find it uncomfortable reading the words!
As far as the inherent uncertainty in climate science prognostications go, there has just been an important paper published in Nature which apparently firmly establishes and quantifies causality for CO2/CH4 increases and changes in GMTA ((global mean surface temperature anomalies) from 1850 onwards and previously back 800,000 years
http://www.nature.com/articles/srep21691 -
One Planet Only Forever at 01:10 AM on 25 February 2016Fossil fuel funded report denies the expert global warming consensus
Jeff T @ 5,
The Critical Thinking course I took as an option when I pursued my MBA identified understanding who wrote something to be one of the first steps in evaluating a source of information.
That understanding does not 'bias' a review by a conscientious responsible pursuer of better understanding. It helps establish the type of effort that should be put into the review. How much careful exclusion or deliberate deception to expect can make the effort to evaluate the information much more effective.
-
Joel_Huberman at 23:50 PM on 24 February 2016The Uncertainty Handbook: Download and Translations
I'm confirming SteveS's observation. Furthermore, Adobe Reader also shows the last two pages as being blank.
-
Models are unreliable
FrankShann - That's the entire reason for the various Representative Concentration Pathways (RCPs) evaluated in the models, to bracket potential emissions. Shorter term variations (which include emergent ENSO type events), volcanic activity, solar cycle, etc, are also incorporated, but as those aren't predictable they are estimates.
Which doesn't matter WRT the 30 year projections from the models, as the natural variability is less than the effects of thirty year climate change trends, and the GCMs aren't intended for year to year accuracy anyway - rather, they are intended to probe long term changes in weather statistics, in the climate.
-
ajadedmug at 18:44 PM on 24 February 2016The Uncertainty Handbook: Download and Translations
SteveS - Thanks very much: that is worth a deal, and is exactly the sort of response that I was hoping to get. I shall now amend my draft email to Climate Outreach and explain that the problem appears not to exist on a Mac for the most recent versions of OS and Preview. Then I shall send it :). CO can then judge whether it wishes to investigate further.
-
One Planet Only Forever at 15:28 PM on 24 February 2016Republicans' favorite climate chart has some serious problems
HK,
The real problem is not Capitalism, Competition, Free Markets, or Marketing (or Communism or Socialism). The real problem is the tendency for some people to want to personally get away with something they can understand is unacceptable.When people choose to pursue personal interest without responsible conscientious regard for the consequences of their chosen desires and actions, they can have a competitive advantage in any game (Those economic and political things are all games made-up by humans. And the problem gets worse when those callously self-interested types win the right to make-up the rules or make-up how the rules will be enforced).
And the misuse of the powerful science of Marketing can allow those type of people to get more undeserved advantage through the drumming up of undeserved popular support from others who are easily impressed, willing to believe unjustifiable claims in support of unacceptable actions.
That is what is going on, and not just in the US.
-
FrankShann at 14:25 PM on 24 February 2016Models are unreliable
A very tentative suggestion...
In addition to @972 and @973: to convince most people, climate models need to be able to forecast *future* temperatures, and not just accurately describe *past* temperatures (I am avoiding the word "predict").
Unfortunately, you then need to forecast all the forcings for a given future year, as well as have a good climate model. Errors in your forecasts of the forcings may be at least as great as the errors in your climate model.
What about taking the current Best Available Model (BAM-2016) and saying "we forecast that global mean surface temperature (GMST) will be Z if the CO2 is V, El Nino is W, aerosols are X, volcanic activity is Y etc (with reliable sources for the values of the forcings defined in advance, such as the appropriate values from www.esrl.noaa.gov/gmd/ccgg/trends/global.html#global_data for CO2)? The model would be published in advance and the results updated each year. New models could be introduced over time (such as BAM-2020 in 2020) while still calculating the forecasts from previous models (such as BAM-2016).
Perhaps this has already been done, or models such as CMIP5 cannot be applied in this way, or some important forcings cannot be pre-specified in a suitable fashion, or there are other problems. As I say, this is merely a tentative suggestion.
-
SteveS at 13:53 PM on 24 February 2016The Uncertainty Handbook: Download and Translations
For what it's worth, I had no problems with the english version on a Mac running OS X El Capitan (10.11.3) and Preview 8.1. Pages 19 and 20 are blank, but that seems like it could be intentional.
-
ajadedmug at 12:25 PM on 24 February 2016The Uncertainty Handbook: Download and Translations
BaerbelW - I thought that a more prudent first step would be to establish whether other Mac users in the Skeptical Science community are seeing this issue. If I email Climate Outreach, they may waste time investigating a software issue that is in reality solely related to my MacBook or to Macbooks running older versions of software and OS.
My MacBook runs Mac OS X Lion 10.7.5 (11G63) and Preview Version 5.5.3 (719.31). As a test comparison, I downloaded the latest Adobe reader to compare, and the documents were then perfectly readable. I then deleted the Adobe Reader and tried with Preview again. This time more of the text was viewable, but there are still major gaps. This is something that I have never seen before with Preview. I have tried opening numerous other .pdf files that I have already viewed previously with Preview and there were no problems.
I have written a draft email to Climate Outreach and will wait 48 hours to see if anyone responds before I send it, if I send it. I am probably untypical: most PC or Mac users probably use Adobe reader. I don't, because Preview has worked fine for me over the last eight years...until yesterday. Given most people with a computer will use an Adobe reader, this Preview issue is probably more curious than significant, and may be limited to people like me who are running older versions of operating system and software.
-
scaddenp at 12:25 PM on 24 February 2016Models are unreliable
Good points. If you want to look at how past model "predictions" of the future have gone, then you could of course look at how, say, one of the earliest models, Manabe and Weatherall 1975, used by Wally Broecker, is doing now. See here. It would be nice to update this chart. There are other predictions (eg FAR) also in the "Lessons from past climate predictions" series that are worth looking at. Of course any serious evaluation has to compare actual forcings against what the prediction assumed if you want to assess skill at climate modelling rather than guessing emission rates. Broecker overestimated emissions but also looks to have underestimated sensitivity.
-
sailingfree at 11:09 AM on 24 February 2016Models are unreliable
KR@968: Thanks, your statement
"The proper question is "Is there evidence to support the assertion that CMIP5 models predict temperatures from forcings?" The answer to that is certainly 'yes', as validated by their ability to reproduce past temperature evolutions on both the global and regional levels."
helps clear it up for this layman.
I do side with FrankShann, though, as it has not been clear to me to what extent past temperatures are inputs to the models. (After all, before Copernicus, anyone could predict a daily sunrise, knowing no physics at all.) It must be made clear that the models do not use past temperatures to predict the future temperatures. (like, Gee, I see a sine wave and a slope, so that should continue in the future.)
A passage like "We wrote the models during 2005 to 2010,(or whatever) and used the conditions and temperature of 1880 and the known forcings from then, like sunlight, CO2, and volcanoes, and then ran the models and successfuly predicted (or descibed) the temperatures from 1880 to 2010." would be more clear to me.
A doubter would still wonder, since we knew in advance the temperatures to describe, to what extent we tweaked parameters to get the right temperatures. So you would have to show him how the tweak was for proper humidity, wind, or what ever, not just to get the desired temperature.
-
FrankShann at 09:32 AM on 24 February 2016Models are unreliable
KR @968. My question accurately represented Dana's post. He said "climate models have done an excellent job predicting how much temperatures at the Earth’s surface would warm", and immediately below that showed the graph of CMIP5 and temperature for 1880-2010. Dana did not say "from forcings", so neither did I. The title of the graph includes "using CMIP5 all-forcing experiments" and that was shown - but every person I asked said that the graph did not provide evidence that CMIP5 has done an excellent job predicting temperature - because the vast majority of people use predict to mean forecast. Even Tom Curtis agrees that there is "no doubt that forecast is the archetypal definition of 'predict'".
*If Skeptical Science aspires to explain climate science to the general public, it needs to use words in a way that is understood by the general public (or flag the specialist meaning).*
-
sailingfree at 09:15 AM on 24 February 2016Satellites show no warming in the troposphere
64 Glenn, Thanks. My point is, good data or bad data, a denier would not care, as long as he could show data for the "goldilocks layer".
-
sailingfree at 09:06 AM on 24 February 2016Global warming stopped in
1998,1995,2002,2007,2010, ????
The Earth's surface is in a warming trend.
The stratosphere is in a cooling trend.
So would not there be a "Goldilocks Layer" in between, somewhere in the mid-troposphere?
That Goldilocks Layer is close to the middle troposphere, which is a recent favorite of Christy.
So in the Goldilocks Layer, we have no warming, over the whole globe since the beginning of the satellite data.Who could deny that?
-
BaerbelW at 06:53 AM on 24 February 2016The Uncertainty Handbook: Download and Translations
ajadedmug - I'll check with folks at Climate Outreach who created the handbook and let them know about your issue - unless you'd like to let them know yourself via a comment on their own page for the downloads?
-
ajadedmug at 06:35 AM on 24 February 2016The Uncertainty Handbook: Download and Translations
I downloaded the English and German versions to a MacBook and viewed them in Preview. The text looked like it was full of gaps and missing letters. At first I thought this was just going to be a witty first-page pun on 'uncertainty', but both versions were like this from start to finish. I had no similar problems using Preview to read The Debunking Handbook. Has anyone had similar issues reading The Uncertainty Handbook in Adobe's PDF viewer (Mac or PC)? Or has my MacBook flipped :)?
-
Fairoakien at 05:52 AM on 24 February 2016Fossil fuel funded report denies the expert global warming consensus
My concern is that NOaa is trying so hard to maintain the human caused global warming thesis that they have turned to changing facts.
NOAA has funded two studies to deny that there was ever a MWParound 1100 AD that lasted for about 100 years.
NOAA funded a study to falsely lalim that the danes did not settle Greenland because of warming of the Atlantic and did not leave because the climate cooled.
Most recently NOAA submitted a report denying the flattening of the warming curve betwee 1998 & 2015 by using the most subverted statistics. NOAA internal staff disapproved the report.
NOAA has recently changed the source of temperature readings, which coincidentally supports their prior report that there was no paise in the warming curve and indicates a big jump in the 2015 temperatures.
Sorry to say but NOAA is losing credibility and is more susceptible to climate deniers by providing non-science backed support for their position.
Oh, many of those climate scientists that support the human caused global warming have degress in climate fields, e.g, biologists, physicists, engineers etc.
Moderator Response:[PS]
Please note that posting comments here at SkS is a privilege, not a right. This privilege can and will be rescinded if the posting individual continues to treat adherence to the Comments Policy as optional, rather than the mandatory condition of participating in this online forum.
Moderating this site is a tiresome chore, particularly when commentators repeatedly submit offensive or off-topic posts. We really appreciate people's cooperation in abiding by the Comments Policy, which is largely responsible for the quality of this site.
Finally, please understand that moderation policies are not open for discussion. If you find yourself incapable of abiding by these common set of rules that everyone else observes, then a change of venues is in the offing.Please take the time to review the policy and ensure future comments are in full compliance with it. Thanks for your understanding and compliance in this matter.
Take note in particular of accusations of fraud etc. in policy. If you wish to take issue with the published science then present papers and data to support your viewpoint rather than resorting to conspiracy theories. If you are uninterested in facts that do not support your wishes, then science is not for you and certainly this site is not for you. There are many other blogs that will appreciate your comments.
-
John Hartz at 02:46 AM on 24 February 2016Models are unreliable
From the handy-dandy SkS Glossary...
Climate model
spectrum or hierarchy
A numerical representation of the climate system based on the physical, chemical and biological properties of its components, their interactions and feedback processes, and accounting for all or some of its known properties. The climate system can be represented by models of varying complexity, that is, for any one component or combination of components a spectrum or hierarchy of models can be identified, differing in such aspects as the number of spatial dimensions, the extent to which physical, chemical or biological processes are explicitly represented, or the level at which empirical parametrizations are involved. Coupled Atmosphere-Ocean General Circulation Models (AOGCMs) provide a representation of the climate system that is near the most comprehensive end of the spectrum currently available. There is an evolution towards more complex models with interactive chemistry and biology (see Chapter 8). Climate models are applied as a research tool to study and simulate the climate, and for operational purposes, including monthly, seasonal and interannual climate predictions.
Definition courtesy of IPCC AR4.
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 Pres
-
PhilippeChantreau at 02:37 AM on 24 February 2016Models are unreliable
Splitting hair about semantics while the planet is accumulating heat at a rate that was only seen before during cataclysmic mass extinctions. Perhaps that's what is truly wrong about this species of ours...
-
Models are unreliable
FrankShann - I agree that the question you pose (regarding "evidence to support the assertion that CMIP5 predicts temperature") will lead people to answer 'no', but that is because as stated it's an incomplete question.
The proper question is "Is there evidence to support the assertion that CMIP5 models predict temperatures from forcings?" The answer to that is certainly 'yes', as validated by their ability to reproduce past temperature evolutions on both the global and regional levels. And that indicates that their projections, conditional predictions, about the future are useful.
If you pose a poorly worded or incomplete question, you shouldn't be surprised by a nonsensical answer.
-
FrankShann at 00:31 AM on 24 February 2016Models are unreliable
Tom Curtis @966.
*If Skeptical Science aspires to explain climate science to the general public, it needs to use words in a way that is understood by the general public (or flag the specialist meaning).*
You seem very confident of your expertise on many subjects outside climate science. What precisely is your expertise on survey techniques? The survey questions suggested by you were of very low quality indeed. I didn't pretend that my survey was definitive, but your statement that the information is "worthless" is incorrect (it is often risky to assert absolutes in science). In fact, you have very little information about how I did the survey. I asked university staff because it was convenient (as I said), and because an educated sample was more likely to know that "predict" has alternative meanings (so was more likely to refute my hypothesis). You say, "I have no doubt [another absolute] that forecast is the archetypal definition of 'predict'" - which is consistent with the meagre evidence from my survey (where respondents took "predict" to mean "forecast").
You asserted (without presenting any evidence) that "your suggestion, 'description' will [another absolute] cause confusion as to whether or not GMST is a dependent or independent variable in the models". Who says? My point (which you overlook) is that "predict" is just as likely as "describe" to cause confusion as to whether GMST is a dependent or independent variable - both terms can be used to describe independent variables and your ignorance of the term "descriptor variables" is merely evidence of ignorance (and suggests that "describe" may be less likely than "predict" to be mistaken for an independent variable).
I did not intend to make such a big deal of it, but I still think that saying "Climate models have done an excellent job describing how much temperatures at the Earth’s surface have warmed" when combined with the 1880-2015 CMIP5 graph is less likely to cause confusion than Dana's wording. It still refutes Christy's misleading graph, which is far more important than my minor quibble about semantics.
As we seem to agree that climate change is an important problem that needs to be addressed urgently, is it time to call a truce over "predict" versus "describe" so we can spend more time fighting the deniers rather than squabbling with each other?
-
Jeff T at 22:25 PM on 23 February 2016Fossil fuel funded report denies the expert global warming consensus
Although "Who wrote it and what is their expertise in the field?" is often relevant, is it really the first question one should ask about a manuscript? Asking that question first leads to evaluation by source rather than evaluation by content.
-
bozzza at 21:10 PM on 23 February 2016Fossil fuel funded report denies the expert global warming consensus
Seeing how important naming systems are to science I think it fair to say that consensus is ver important in science.
-
shoyemore at 19:53 PM on 23 February 2016Fossil fuel funded report denies the expert global warming consensus
This paper finds "robust and replicated evidence that
communicating the scientific consensus on human-caused climate change leads to significant and substantial changes in perceived scientific agreement among conservatives, moderates, and liberals alike"Deniers often argue that "consensus is not important", but clearly it is, and the Heartland Institute know it. The report is just another attempt to muddy the waters.
-
barry1487 at 17:47 PM on 23 February 2016Fossil fuel funded report denies the expert global warming consensus
It could be added that Richard Tol criticised the Cooke et al paper, and said that he had no doubt there was a strong consensus (high 90s).
http://www.theguardian.com/environment/climate-consensus-97-per-cent/2014/jun/02/republican-witness-global-warming-consensus-real
Unwilling to read the Heartland document. I wonder if they mentioned this Tol's efforts in their 'literature review.'
-
barry1487 at 17:24 PM on 23 February 2016Tracking the 2°C Limit - January 2016
Thanks for the reply, Rob. 6 mo does look better. Did you analyse the fits statistically?
-
Tom Curtis at 16:23 PM on 23 February 2016Models are unreliable
FrankShann @965,
First, you do not test for reading comprehension by asking questions that can be answered by simply parroting the text. Ergo, questions that do so do not test for confusion or lack of confusion introduced by the terms used. And, yes, the precise wording of my questions woud be confusing to the general public, but a I presumed I was conversing with an intelligent person, I did not undertake an appropriate rephrasing required if I were to conduct an actual survey.
Second, you now appear to be indicating that in your convenience survey you showed members of the general public Dana's article, and asked them whether, from the use of the word predict, they would conclude that CMIP5 models were developed prior to 1880? That is certainly not how you described it before. Rather, you took a survey of people with a highly specialist knowledge on a few academic topics, and tested for the archetypal meaning of 'predict' in their usage. The survey population was not representative of the general public - particularly so with their usage of statistical terms. The survey was not double blind. And the survey did not test for whether or not Dana's wording would in anyway cause confusion. As supporting evidence for your position, it was worthless. That you cannot recognize this shows your knowledge of linguistics to be as abysmal as my knowledge of medicine.
I have not ever said that temperature was a predictor variable in the models. Indeed, I have said the exact opposite. Nor would I describe anything as a 'descriptor variable' which is a term that has no meaning that I am aware of.
What I would say is that both CO2 emissions and GMST are equally desribed by climate models. One, however, is an independant variable. The other a dependant variable. Saying that CMIP5 models 'describe GMST' leaves us completely in the dark as to whether GMST is dependant or independant.
-
FrankShann at 15:23 PM on 23 February 2016Models are unreliable
KR @962. I am not suggesting that we wait decades. I agree that urgent action is needed now, and should have been taken years ago. I am quibbling about a very minor matter - what the average person thinks about a graph of temperature from 1880-2015 as evidence to support the assertion that CMIP5 predicts temperature. That is, I suggest that it is not a good idea to use "predict" with a specialist meaning on a site aimed at people who are not climate scientists (and without pointing out that predict is being used with a non-standard meaning).
I was suggesting only that Dana consider altering one word, and substitute describe for predict, so that the text would read, "Climate models have done an excellent job describing how much temperatures at the Earth’s surface would warm". This does not alter the thrust of Dana's post in any way.
John Christy's misleading graph purports to show that CMIP5 does not model past temperatures well, and so cannot be trusted to predict future temperatures. Dana's graph provides strong evidence that CMIP5 is an excellent model of past GMST, which suggests it is very likely to be good at predicting future GMST (especially as previous CMIP models have predicted future GMST well even though they were not as good as CMIP5 at describing past temperatures).
@964 Tom, we differ about the question - your questions are not what Dana wrote, and they would be completely unintelligible to the general public. Despite your theoretical speculations, people who are not climate scientists did not think that the graph provides evidence that CMIP5 has done an excellent job in predicting temperature; they took predict to mean forecast.
We also disagree about describe. Saying that the model predicts temperature does not imply that temperature is a predictor variable in the model, and saying that the model describes temperature does not imply that temperature is a descriptive variable in the model. You can't have it both ways, although I suspect you may well try.
*Again* If Skeptical Science aspires to explain climate science to the general public, it needs to use words in the way they are understood by the general public (or flag the specialist meaning). If those who run the site will not accept this, then we are all worse off because the site will be less effective and it is important that it succeed.
-
Tom Curtis at 13:52 PM on 23 February 2016Models are unreliable
FrankShann @963, I can agree with you that 'predict' is not an ideal word in the context. The problem is that 'described temperatures' would be even worse. 'Retrodict' would be better except that for part of the data todate, it is in fact predicted (whether we take that from 2005, ie, the date of the last historical input, or from 2012, ie, the date experiments for inclusion in the IPCC AR5 needed to be completed, and which therefore are the results actually presented).
Having said that, your convenience sample was asked the wrong question. To truly test whether Dana used the wrong word, you should have given them an example of his sentences of equivalent, and asked:
Based on this sentence and graph
1) Was GMST an independant or dependant variable in CMIP5 models;
2) Were CMIP5 models constructed around 1880 or around 2010; and
3) Did CMIP5 models successfully or unsuccessfully model GMST.
Based on Dana's sentences, if they lead to significant confusion about any of these three points, there was a problem with his use of the word. If not, then not.
I have no doubt that forecast is the archetypal definition of 'predict' just as "unmarried, marriagable male" is the archetypal meaning of 'bachelor' (or was in the 1950s, I suspect the gender specification has now been dropped, or is in the process of being dropped). The later, however, does not cause confusion when we talk of bachelor degrees, or knights bachelor, and did not cause confusion when we first started hearing about 'bachelor girls'. We humans are smart enough to modify the meaning of words from the archetypal value based on context and without confusion. (Computers, not so much.) As a result we use that capacity for flexibility of communication when no word has the exact semantic value we require. We do it all the time, and typically seamlessly.
And that is all that Dana has done.
His problem was that there is no ideal word in the context. But a non-archetypal, but quite common usage of 'predict' worked well. I am sure he would welcome a better word, but none has been suggested. In particular, your suggestion, 'description' will cause confusion as to whether or not GMST is a dependant or independant variable in the models.
-
Rolf Jander at 13:51 PM on 23 February 2016Fossil fuel funded report denies the expert global warming consensus
The frustrating thing is that scientists need to waste their time saying that they all agree on the problem. refuting the attack on the consensus must be taking some time away from actually doing vital research and communicating the finding. That must be part of the objective of these attacks.
-
FrankShann at 13:11 PM on 23 February 2016Models are unreliable
Tom, I am intending to be descriptive - of what people who are not climate scientists think about using a graph of temperatures between 1880 and 2015 as evidence that CMIP5 has "done an excellent job predicting how much temperatures at the Earth’s surface would warm". I told a convenience sample of my university colleagues that "CMIP5 is a climate model developed from 2008 onwards" and asked, "Does this graph provide evidence that CMIP5 has done an excellent job predicting temperature?" None thought that it did.
I repeat, if Skeptical Science aspires to explain climate science to the general public, it needs to use words in the way they are understood by the general public (or flag the specialist meaning).
-
Models are unreliable
FrankShann - As Tom Curtis points out, conditional predictions are indeed part of the definition (a basic part of physics, as it happens), and that's exactly what climate models provide. Trying to focus on only a single one of the multiple definitions in common usage is pedantry.
As to validation, the fact is that GCMs can reproduce not just a single thread of historic GMSTs, but in fact regional temperatures, precipitation, and even to some extent clouds (although with less accuracy at finer and finer details, and clouds are quite challenging). Those details are not inputs, but rather predictions of outcomes conditional on the forcings. _That_ validates their physics - and justifies taking the projections seriously.
We certainly do not need to wait decades before acting on what these models tell us.
-
Tom Curtis at 10:28 AM on 23 February 2016Republicans' favorite climate chart has some serious problems
FrankShann @21, dictionary definitons are descriptive lexicography (see comment here), not presciptive lexicography. It follows that they can only guide usage, not prescribe it. In using the Oxford defintion, you are prescribing that Dana drop a less confusing term ('predict') for a more confusing term ('describe'). Specifically, many people would interpret 'CMIP5 models desribe GMST' as indicating that GMST was prescribed for CMIP5 models in much the same way as forcings are. It would certainly leave them confused as to whether it was a dependant or independant variable. Nor would it be particularly useful to describe the CMIP5 GMST output as a dependant variable, for that certainly is jargon. In fact, short of expanding the article by a carefull discussion of prediction, retrodiction and projection, and how it applies to CMIP5, I cannot think of a better word (in terms of avoiding misunderstanding) than the term used by Dana in this context.
I don't particularly care about your statistical experience, given that you are making a linguistic point. Nor, so far as I can see do you have any justification for complaint about the response to your suggestion. I have made suggestions for improvements in the past on SkS articles. Some have been ignored. Some have been hotly debated, and a few have been accepted. That is what I would expect from a group of independently minded people who make up their own minds about the validity of arguments.
-
Tom Curtis at 10:03 AM on 23 February 2016Models are unreliable
FrankShann @960, you quote as your source the Oxford English Dictionary but my print version of the Shorter Oxford gives an additional meaning of predict as "to mention previously" ie, to have said it all before. That is equally justified as a meaning of 'predict' by its Latin roots which are never determinative of the meaning of words (although they may be explanatory of how they were coined). The actual meaning of words is given by how they are in fact used. On that basis, the mere fact that there is a "jargon" use of the word, means that 'predict' has a meaning distinct from 'forecast' in modern usage. Your point three refutes your first point.
For what it is worth, the online Oxford defines predict as to "Say or estimate that (a specified thing) will happen in the future or will be a consequence of something". That second clause allows that there can be predictions which do not temporally precede the outcomes. An example of the later use is that it could be said that "being in an open network instead of a closed one is the best predictor of career success". In similar manner, it could be said that forcings plus basic physics is the best predictor of climate trends. This is not a 'jargon usage'. The phrase 'best predictor of' turns up over 20 million hits on google, including in popular articles (as above). And by standard rules of English, if x is a good predictor of y, then x predicts y.
As it happens, CMIP5 models with accurate forcing data are a good predictor of GMST. Given that fact, and that the CMIP5 experiments involved running the models on historical forcings up to 2005, it is perfectly acceptable English to say that CMIP5 models predict GMST up to 2005 (and shortly after with less accuracy based on thermal inertia). On this usage, however, we must say they project future temperatures, however, as they do not predict that a particular forcing history will occur.
As a side note, if any term is a jargon term in this discussion, it is 'retrodict', which only has 15,000 hits on google.
As a further sidenote, you would do well to learn the difference between prescriptive and descriptive grammar. Parallel to that distinction is a difference between prescriptive and descriptive lexicographers. The curious thing is that only descriptive lexicographers are actually invited to compose dictionaries - while those dictionaries are then used by amateur prescriptive lexicographers to berate people about language of which they know little.
The only real issue with Dana's using 'prediction' is if it would cause readers to be confused as to whether the CMIP5 output on GMST was composed prior to the first date in the series or not. No such confusion is likely so the criticism of the term amounts to empty pedantry.
-
Republicans' favorite climate chart has some serious problems
Well said, OPOF!
Like you, I’m fundamentally critical to the capitalist system too. I’m a little hesitant to call myself a socialist as they also have said and done a lot of stupid things, but an economic system that depends on non-stop growth on a limited planet can’t be sustainable in the long run. It seems that most politicians believe that the Earth is bigger now than it was 100 years age, while everyone with some scientific knowledge know that its surface area of 510 million km² hasn’t changed much for 4.5 billion years.
But let’s get back to the Christy chart:
In my opinion the chart isn’t directly lying about anything, it’s just misleading and therefore lying in an indirect way. And the most misleading part of it is what I’ve already highlighted, namely comparing estimates by a climate model for the surface with observations of the "bulk atmosphere" reaching up to 50,000 feet.
The blue curve in my graph is the average of all the RATPAC-A data sets from the surface to the 100 millibar level (close to 50,000 feet) and indicates a warming of roughly 0.3 °C since the late 1970’s. That’s pretty similar to the balloon data presented in the Christy chart. No direct lies there!
The red curve in my graph is the average of the data sets from the surface to the 400 millibar level and indicates a warming of roughly 0.6 °C since the late 1970’s (the trends given in the graph cover the whole period from 1960). If we insert that curve into the Christy chart, we will see a far better agreement with the climate model, especially if we consider the uncertainty!
So, no direct lies, but the Christy chart is more or less like comparing a climate model estimate from Norway (there I live) with observations from Africa! -
FrankShann at 07:56 AM on 23 February 2016Models are unreliable
KR @959. Thank you for taking the trouble to respond to my posts in www.skepticalscience.com/republicans-favorite-climate-chart-has-serious-problems.html and for explaining how GCMs are developed. However...
1. The word "predict" means to state what will happen in the *future* (Latin prae- "beforehand" + dicare "say"), or to state the existence of something that is *undiscovered* (such as Einstein's prediction of gravitational waves). My source is the Oxford English Dictionary. Any other meaning of predict is jargon and will be misinterpreted by most readers (especially if, as in Dana's post, it is not flagged as being used in an unconventional sense). For Skeptical Science to "explain what peer reviewed science has to say about global warming" to the general public, it has to use words in the sense understood by the general public (or clearly flag the use of jargon).
2. Dana presented a graph of CMIP5 modelling of global mean surface temperature (GMST) from 1880 to 2015. By definition, the CMIP5 estimates from 1880 to 2000 are not pre-diciton (before-stated), they are "hindcasting" (see point 4, below). Yet Dana implies that this graph is evidence that CMIP5 has "done an excellent job predicting how much temperatures at the Earth’s surface would warm" without any hint that he is using "predicting" to mean something other than "stating that a specified event will happen in the future" (OED).
Additional (peripheral) comments...
3. Even with a jargon definition of predict, experience in experimental science has shown that even randomised trials are subject to bias if they are not double blind. Development of the CMIP5 model was not blinded to GMST, so it is subject to bias. Consequently, the only rigorous validation of CMIP5 is how well it predicts future climate.
4. Look at the first sentence in the Intermediate section of this thread. "There are two major questions in climate modelling - can they accurately reproduce the past (hindcasting) and can they successfully predict the future?" In Skeptical Science's own words in this very thread, hindcasting is distinct from prediction.
-
One Planet Only Forever at 01:12 AM on 23 February 2016Republicans' favorite climate chart has some serious problems
HK@14,
I would say you are justifiably skeptical of the motives of informed and knowledgable people like Dr. Christy and those who choose work like his as their preferred presentation of what is going on.
I am almost certain that Dr. Christy is well aware of the difference between the trend of the lower and higher atmosphere and 'abuses' that understanding to misrepresent what he actually better understands. Others like him also try to focus only on satellite data that they understand to be potentially unreliable (highly uncertain compared to other measures of what is going on), and which would require many more years of gathered data to undeniably show how wrong their preferred presentations are.
I am also quite certain that there are some people among us who only care about enjoying and winning as much as they can get away with in their lifetime. And they understand that they can obtain a competetive advantage by deliberately trying to get away with activity they actually understand is contrary to the advancement of humanity to a lasting better future for all. They have almost no interest in the consequences of their actions, and are particularly callous about the consequences in 'a future they will not be alive in'. And some are even worse, having a short-term focus on getting away with winning just 'one more year (or one more fiscal quarter) of getting away with what they understand they should not be allowed to get away with'.
And I am equally certain that the politicians and media pundits who choose to use the deliberately deceptive abuses developed by the likes of Christy understand the unacceptability of their preferences, including fully understanding how unacceptable it is for a more fortunate member of current day humanity to contiune to obtain more personsal pleasure and profit from the known to be damaging and actually ultimately unsustainable burning of fossil fuels.
Hopefully one of the better understandings that grows from this climate change challenge is that popularity and profitability (in business or politics) cannot be relied upon or trusted to develop a lasting better future for humanity, because of the power of the science of deliberately deceptive marketing.
History is full of examples of understood to be damaging activity developing and being prolonged by unjustified drumming up of popular support for known to be unacceptable things. This climate change challenge is an obvious example of how some people are encouraged to behave as unacceptably as they can get away with by the competetive capitalism system and the fatally flawed rules of the game made-up by those among humanity who only care about themselves in their region of the planet in their lifetime (or worse yet just caring about getting away with stuff for one more, 4 year election cycle, or even worse focused only on a 3 month period of time).
-
Republicans' favorite climate chart has some serious problems
FrankShann - I have responded to your model discussion in the appropriate thread.
-
Models are unreliable
FrankShann - Yes, to 'predict' involves a result that wasn't input to the model, but given that GCMs don't have temperature observations as inputs, rather the forcings and the physics, even a retrodiction is still producing results that weren't inputs. Now as to 'tuning' models, what occurs in real life (as opposed to rhetoric) is that when models differ from observations at any scale, including regional variations, relative humidity, ocean currents, etc., the physics for that portion of the model are investigated for errors in the physics. Then the models with (hopefully more accurate) physics are run to see how well they reproduce observations. They are not tuned by temperatures, as erroneous physics re-tuned to a specific output will become even more erroneous, but rather to physical observations at all scales.
Purely statistical models do get tuned, but GCMs are physical models. And the many apparent attempts to dismiss models based upon efforts to faithfully reproduce physical observations, casting them instead as attempts to get a specific output temperature, are therefore incorrect.
The primary results of GCMs are projections, which is to say conditional predictions - if forcing change X occurs, the climate will evolve as Y over time. The CMIP5 model runs projected certain temperatures given specific forcing estimates, and those do diverge from observations - but then so do the observed forcings diverge from the forcing estimates. When we check those conditional predictions using actual forcings, to see what the models show in that case, we find that they are actually quite accurate, that the observations fall well within the bounds of model variability. And thus the results of the models are indeed "predictions". Conditional predictions of the relationships between forcings and climate evolutions.
-
Republicans' favorite climate chart has some serious problems
I would strongly suggest that these discussions of models, reliability, and the semantics of "predict", "project", etc be moved to the models are unreliable thread where these issues have been extensively discussed, as this thread concerns issues with Dr Christy's chart and the (odd) choices made in creating it.
-
FrankShann at 22:03 PM on 22 February 2016Republicans' favorite climate chart has some serious problems
The core issue here is merely semantic. Modellers use "predict" to mean how well their model performs on past data, even when the model has been adjusted (tweaked) in the light of knowledge of the dependent variable. This meaning of "predict" is jargon. The general meaning (well over the 97% consensus mark) is that predict means to make a statement about something that is unknown (such as GMST in 2030-2040, or the presence of gravity waves). For example, I take information about the results of football matches over the last 10 seasons and make a model that "predicts" the winners. Then I try different variables or transformations and get better "prediction". But the vast majority of people do not regard this as prediction - for prediction, they require my model to say which teams will win *next* season. That is, GMST after 2015, not 1880-2015. I was suggesting (and still suggest) avoiding jargon and using the widely accepted meaning of "predict" (as defined in the Oxford Dictionary) - which means that CMIP5 describes rather than predicts GMST for the vast majority of 1880 to 2015.
Tom Curtis @19. I did not say global mean surface temperature (GMST) was "fed into" CMIP5, but I agree that I should have made it clear that I am not suggesting that global mean surface temperature (GMST) is an independent variable in the CMIP models. I *am* suggesting that the models have been adjusted in the light of how well they predict GMST (and other variables) over some or all of the period from 1880 to 2014 (the period shown in the graph). Knowledge of GMST during the period has influenced the development of the models.
Tom Curtis @20. I am not atacking Dana's post - it is very helpful indeed (as I said @3). Also, I heartily agree that climate models are remarkably useful predictors of future climate, and vastly superior to the denialist attempts. I merely suggest that Dana consider altering one word (predict to describe) so the post is more plausible to readers who are not statistical modelers (the vast majority of the population) so it reads "Climate models have done an excellent job describing [instead of predicting] how much temperatures at the Earth’s surface would warm" because this statement is supported by a graph plotting CMIP5 against GMST from 1980-2015 (and it still refutes John Chrisy's misleading implication that climate models do not describe past GMST well). The link you mention to the excellent Comparing Global Temperature Predictions article (which I printed and gave to my friends in 2011) occurs at the end of the post, far removed from the predict/describe statement.
I am disappointed at the response to my efforts to help. I am not a climate scientist, but I have have extensive experience with statistical modeling and scientific publication (I'm a member of the International Advisory Board of The Lancet). I tried to help because I think climate change is extremely important and that Sceptical Science is a very useful resource - and that it might benefit from advice from a non-climate scientist about how a post could be misinterpreted by other people who are not climate scientists. Perhaps I won't bother in future.
-
Tom Curtis at 16:15 PM on 22 February 2016Republicans' favorite climate chart has some serious problems
FrankShann @various, CMIP5 uses historical data to 2005 inclusive, but scenario data thereafter. The equivalent model experiments for the IPCC TAR and AR4 used historical data to 2000, and scenario data thereafter. I am not aware of the dates for the FAR and SAR, but clearly they predated the reports themselves (1990 and 1995 respectively). Hansen 88 used historical data only up to 1983 (from memory).
This is important because Dana wrote:
"Climate models are certainly useful, and are doing a pretty darn good job predicting global warming. Their predictions have been far more accurate than those made by climate contrarians, who keep telling us that the Earth will start to cool any day now, as we keep breaking heat records."
In doing so he links to an article discussing the predictions of all five IPCC reports, plus other 'warmist' predictions, along with those few predictions made by AGW 'skeptics'. He was not referring to just CMIP5 predictions as you suppose, and all of the predictions include several years after the last historical forcing data, including CMIP5.
The article summarizes some of the results from a number of articles discussing the predictions of just one individual or organization in any article. The results of the comparisons are summarized in this graph:
So, leaving aside all the technical points about predictions and retrodictions, and predictions and projections; your criticism of Dana is wrong. The worst that can be said of his claim is that it is fleshed out in this article, but another. That, however, is a necessity of communication. Even in the world of the internet, there is still no royal path to knowledge. And links are not just decorations.
And for what it is worth, the HadCRUT4 trend since 2005 is 0.127 +/- 0.22 C/decade, compared to a CMIP5 prediction of about 0.2 C/decade. As the difference between scenarios is inconsequential over that time period, the distinction between projection and prediction is nugatory. Ergo, the CMIP5 prediction still fairs fairly well, and certainly outperforms any 'skeptical' prediction over the same period.
-
Tom Curtis at 15:15 PM on 22 February 2016Republicans' favorite climate chart has some serious problems
FrankShann @18, GMST was not data fed into the climate models. It is a dependent variable. This can be seen by the variability of model retrodictions over the historical period (in this case for CMIP3):
Given the other things you say, your claim that "Global mean surface temperature (GMST) information *was* given to CMIP5 beforehand, so the graph is description and not prediction" is quite bizarre. I assume your claim was just poorly worded, and ask for clarrification.
Prev 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 Next
