Currently, I’m just adding tests to my graphs. For that reason, the discussion will be brief merely explaining the additional informaiton added to one of the graphs from Observations V. Models: “Model Weatherâ€. where I discussed a method to see whether the earth weather trend falls inside the spread of “weather trends” using the estimate of variability of model trends based on repeat runs for matched periods from that model.
I told readers I’d add the multi-model mean. I’ve actually added 3 versions of the multi-model mean shown in bold below. My “preferred” version is labeled “MT”.
(Careful readers will note that for convenience I am drawing data from a 2nd hand source, and these trends end in June or July.)
The three versions of the comparison to the multi-model mean are:
- M1: This trace represents the 95% range of weather for a forecast data point assuming the 11 models used to create the forecast all replicate the probability distribution of “earth weather”. I computed the multi-model mean for the test period over the 11 models with more than 1 run. I estimated the spread due to the “weather” in each model by taking the square root of the mean variance. (I’ll call this the pooled mean s.d.) That weather is shown in slate blue. I estimated the uncertainty in the model mean using the ratio of the standard deviation of the model mean over the 11 models to the square root of the 11; this is shown in grey. I then extended the “weather” uncertainty intervals by pooling the 95% spread of weather and the standard error in the sample mean. These are shown in slightly lighter blue. If earth weather was measured perfectly, and that range was a forecast we would expect trends in the forecast period to have a 95% probability of falling in that forecast.
Carrick correctl observed that a full comparison requires inclusion of measurement uncertainty in the earth trend. I estimated that and added it to the uncertainty intervals. The lower bounds of the trend including measurement uncertainty are shown with open circles. (See illustration below for more details.)
- MT: This trace represents the 95% range of weather for an as yet unseen realization under the assumption that all models used to compute the A1B SRES replicate the probability distribution of “earth weather” but we were only able to estimate the standard deviation of trends due to “weather” using the 11 models with repeat runs. The only difference between this and the previous test is the multi-model mean is a bit higher and “22” was used to compute the standard error in the mean. This is the multi-model mean actually used in the AR4, and for that reason, I consider it the “preferred” mean to test.
When the earth trend falls below the open circles, this indicates that the earth trend as “weather” falls outside the range that is consistent with the multimodel mean– assuming we accept the pooled variance of weather in models as the best estimate.
- ME is called “M enhanced”. This is the result one would get if one estimated the uncertainty using the spread of trends in each model (i.e. what the models actually estimate as uncertainty in trends due to weather”) and then added the spread in the mean trends (i.e. scatter in the location of open circles) into the estimate in the “weather trend”.) This could be justified if the difference in the model trends was due entirely to statistical error and not due to any real difference in trends. In that case, the uncertainty intervals computed under ME enhanced would on average be equal in size to those computed under M1. But we would have a larger number of samples and so get a more precise results. However, in this case, it’s actually pretty easy to show that trends in the forecast region differ from model to model. (I have not shown this test, but I’ve done it.) So, this method results in a spread that is not an estimate of the variability in 138 month trends due to “earth weather”. Rather, it is simply an estimate of the ±95% range of runs including both the effects of “weather” and the fact that some models show more warming and some show slower warming. (This is, btw, the set of assumptions that published papers used to insist the current trends are consistent with the models. They may be consistent in the sense that some models show less warming than the mean and/or are very variable. But this method does not contradict any diagnosis that the ensemble is biased warm.
Note: All graphs assume probability distributions for trends are Gaussian (i.e. Normal).
Guide
There were questions on reading the graph in prior posts. This might help:
Did you subtract the linear trend from each model run before determining the variability associated with weather? Global warming is not weather.
Frank: I computed M trends a 138 month period. I then computed the standard deviation over the M trends using the normal method of computing a sample mean. That normal method subtracts out the sample mean for the batch– and uses the sqrt(M-1) in the denominator as required to deal with the fact that the sample mean is substituted for the unknown true mean.
There is nothing fancy in this– it’s just like computing the standard deviation of anything. So, for example the standard deviation of three mens heights (69, 70,71) is 1.209153. The average is 70. Just substitute “individual man’s height” with a “run”. That’s what I did.
Lucia, Totally off topic. Who should I vote for in my municipal elections:
Tuxedo Stan:
http://www.cnn.com/video/#/video/us/2012/09/22/dnt-cat-runs-for-mayor.ctv
or Dawgfather, PHD (Professional Hot Dogger)?
Clearly, you should cast a write in vote for Tuxedo Stan and not the Dawgfather. Cats make better mayors. I think Earl looks like a good running mate for Stan.
As you can see in the video, the other cat mayor “stubs” has been doing a good job. We need more cats in office.
Thanks Lucia! That certainly lifts the burden of indecision. And here I was going to listen to my know-it-all neighbor with the English accent.
AJ–
If he knows anything he will tell you to vote for the cat. If suggests voting for the Dawg, you will need to begin calling him your “no-nothing neighbor”. There is no question about this. Always vote for cats.
If you vote for the Dawg, the cat will certainly tell you no no no no no.
Even the MET is coming slowly around on the problems with the models.
http://www.dailymail.co.uk/sciencetech/article-2217286/Global-warming-stopped-16-years-ago-reveals-Met-Office-report-quietly-released–chart-prove-it.html
“..he [Phil Jones]insisted that 15 or 16 years is not a significant period: pauses of such length had always been expected, he said….Yet in 2009, when the plateau was already becoming apparent and being discussed by scientists, he told a colleague in one of the Climategate emails: ‘Bottom line: the “no upward trend†has to continue for a total of 15 years before we get worried.’…”
Lucia,
A bit OT, but there is quite tempest over a recent graph showing that warming has been negligible for over 15 years. Now the graph is being ‘explained’ by the authors. Wasn’t there some period of time you were discussing that makes a threshold for falsificatoin of warming trends?
What does this recent graph offer, if anything, to determining the reliability of the predictions regarding warming?
hunter–
I’m in the middle of fiddling with the script to collect a bunch of stuff into a subroutine so I can the extend to using a harder-to-cherry pick metric. (This would be the sum of (dif_trends/sigma_trend , dif_means/sigma_means). That metric can only checked *outside* the baseline– but it’s harder-to cherry pick. But because I am coding, I can’t whip out a fair since 1997 graph very quickly. But the response to the daily mail graph is “on the one hand/on the other hand”.
Daily mail here: http://www.dailymail.co.uk/sciencetech/article-2217286/Global-warming-stopped-16-years-ago-reveals-Met-Office-report-quietly-released–chart-prove-it.html?ito=feeds-newsxml
On the one hand:
1) With respect to daily mail title claim: ”
Global Warming Stopped 16 years ago”, 1997 is a cherry pick.
2) The trend since 1997 absolutely does not prove or even support the headline.
3) Low positive trends since Jan 1997 is consistent with continued warming.
On the other hand:
1) Met office response: Mostly bad. They shold have just shut up after the 2nd paragraph under Q2. (http://metofficenews.wordpress.com/2012/10/14/met-office-in-the-media-14-october-2012/ )
Misleading things in under Q1
Of course if we include periods back in the early 2000s and late 1800s when GHG’s were low, 15 year periods of no warming are not unusual. That is: 15 year long periods of no warming are entirely usual when there is little or no warming due to green house gasses — like way back in the 1800s.
2) In later periods, nearly all 15 year periods of no warming terminate in the cooling ‘dip’ after a stratospheric volcanic eruption. (E.g. Pinatobo, Agung, Fuego, El Chichon). These events are newsworthy. People notice them when they happen. Scientists aren’t going to watch the temperature record and then do forensic reconstructions to “discover” that an eruption occurred in… oh… 2008? )
3) Nothing about those periods of no warming suggest that we could have 15 periods of no warming if the current forced trend is 0.2C/dec.)
Misleading Q2 which specifically asked about 0.2 C/decade
First: It depends what you mean by “uncommon”. But anyway, it is only “not uncommon” to see these if
a) you look cherry pick among models to estimate the variability of 10-15 year trends. (e.g. pick ECHAM with the wildest weather of most models. ) or
b) pick a model whose mean trend is less than 0.2C/dec.
c) Estimate “earth “weather as the spread of “weather + spread of trends in models”. (This is the spread shown in ME. MT would be closer to right.)
d) do several of theses at once.
If you estimate earth weather fairly and you are testing whether the trend is 0.2C/dec, 15 year trends with 0C/dec are pretty uncommon.
I’ll probably post about this tomorrow.
Misleading under Q3:
Volcanic eruptions? Is the Met office seriously trying to suggest the low trend since 1997 might have been caused by some dramatic volcanic eruption we somehow failed to notice happening when it happened?
As for solar or emissions: Well… yes. The 0.2C/dec might be wrong because people making predictions didn’t know what emissions would be. But if someone asked if it’s wrong, the fact that one might wish to know the reason why it was wrong wouldn’t turn the forecast into “right”.
More misleading: They show the graph of records. Sure… those are the records. But if warming did stop, that graph wouldn’t look any different than it does. We’d be near the peak….. Recent years would be warmest.
The Met office has decided to change the subject and instead of rebutting “warming has stopped” (which the Rose of the daily mail claimed) they are trying to rebut “warming never happened) which Rose did not claim. That Met office strategy is either a) stupid or b) dishonest. If the former, they are misleading themselves in believing they have rebutted. If the latter, they think their readers are too dumb to notice their subject change and they are willing to try to take advantage of their readers stupidity.
I apologize if this has been covered before, but shouldn’t you reject(ignore) models whose variance (from it’s trend) is different from the real temps variance (from it’s trend) before comparing trends? If you compare a “noisy” model with the real temps you’re allowing the model to subsume a range that includes the real temps. If it’s noisier than reality than how can it represent reality? Taken to the extreme a really noisy model will always include the real temps wouldn’t it? Conversely, too tight of a variability will have a high probability of rejection even if it is close to the real temps.
BarryW
Whether or not you should compare a particular models trend to observations depends on what question you are trying to answer. If your question is “Does model A’s trend agree with observations?” you have to compare model A’s trend to observations. There is any alternative.
You can also ask other questions. Like: Is it appropriate to assume the variability of N-month trends can be estimated using model A? And is it appropriate to use these when trying to figure out if the observation and trend agree? It may be the answer that you don’t think the variability of N-month trends from that models is appropriate– but then you need to figure out a way to do the estimate.
In this particular case, there really isn’t any good estimate of the “correct” magnitude of variability of “N month” trends for the earth if we want that variability to exclude the extra variability due to non-linearities in the forced trend. We would need to earth realizations under matched forcings to estimate that. So, the only alternatives is to do the comparison using several possible ways to estimate that variability. Using the variabilith in a particular model is one of those ways. I’m not claiming it’s the best way– but it is a way that is defensible in some manner. And we can report what that answer is.
lucia
Re: Type B (systematic) measurement uncertainty
You addressed:
From the size of the circle, that “measurement uncertainty” appears to be a typical average reading error.
Have you included potential systematic error in the measurement error? This could include the full range of anthropogenic errors ranging from differences/changes to time of day of measurement, satellite sensor drift, shifting sites, Urban Heat Island effects on through algorithmic “homogenization” errors, and (biased) adjustment bias etc. Watts et al. (2012) draft observe:
Consequently, may I suggest that there is a Type B component of “measurement uncertainty” (“component of uncertainty arising from a systematic effect”) that appears to be substantially larger larger than conventional evaluations of measurement error. e.g., from my cursory reading of the draft Watts et al. (2012), this systematic measurement uncertainty could be on the order of 0.05C to 0.1C/decade, depending on the final outcome of their paper.
For definitions and algorithms, see NIST Essentials of expressing measurement uncertainty.
NIST Technical Note 1297 1994 Edition, Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results. Note:
As I understand it, this would expand the uncertainty in the “data”. Incorporating this full uncertainty would result in longer periods required to validate or disprove the IPCC models relative to the data. Conversely, it would take more effort to identify and validate anthropogenic contributions to temperature. More “fog” to wade through!
PS My guestimate is not quantitative. A formal evaluation would need to be made and incorporated into the full Expanded Uncertainty. I understand you to be using a coverage factor k of 2. Before we are forced to undertake public expenditures of hundreds to thousands of trillions of dollars, may I suggest it would be prudent to increase the coverage factor k to 2.56 or 3 for 99% to 99.73% probability.
David–
I’m estimating the uncertainty for GISTrend based on the uncertaintys GISS thinks are contained in the data, those in HadCrut based on HadCrut etc.
I don’t think satellite errors are a large factor for these sets.
As far as I am aware, I am not using ‘coverage factors” at all. What do you mean by this? If you mean that when I discuss 95% confidence intervals, I report 95% confidence intervals: Yes. When I say I am showing 95% confidence intervals, I show 95% confidence intervals.
If someone wants to use a different confidence interval when doing a statistical analysis to test the null hypothesis “we should not undertake public expenditures of hundreds to thousands of trillions of dollars” they can use use those when they display the results of that analysis. But I’m not testing that null. I’m testing the more pedestrian null of
“the model mean trend = the true trend”.
I really don’t think this is such a strong well supported null that anyone needs to insist we wait for a weather event that falls outside the 99% or 99.73% confidence interval before we consider it to be worthy of rejection.
lucia
Re: “the uncertaintys GISS thinks are contained in the data,”
Thanks for the clarification.
Sometime may I suggest comparing what would happen using the best data found by Watts et al.( 2012) to GISS etc. (without the large systematic uncertainties in conventional published data.) I think that would show faster rejection or higher probability of rejection of IPCC’s models vs the best measurements.
Re: “I understand you to be using a coverage factor k of 2”.
Yes, your within 95% confidence intervals is what I was referring to by coverage factor of k=2. Expanded uncertainty = mean +/- k * standard deviation.
The measurement uncertainty would have its own k* SD separate from the model uncertainty. See NIST Expanded Uncertainty
I support your 95% rejection for this stage of evaluations.
(I assume common +/- 1 sigma (68.27% confidence interval) is probably more clearly rejected.)
I was just suggesting that when we are pushed to spending $ trillions rather than $ billions, we should insist on much more stringent evidence, e.g. > 99% including the systematic measurement errors uncovered by Watts et al.
“Sometime may I suggest comparing what would happen using the best data found by Watts et al.( 2012) to GISS etc. (without the large systematic uncertainties in conventional published data.) I think that would show faster rejection or higher probability of rejection of IPCC’s models vs the best measurements.”
not likely.
1. Watts et al 2012 is science yet. Code and data please.
2. It only covers the US.
3. The averaging technique they use is presently undocumented. If they did what I think they did, it is biased and would have large uncertainties. Not what you want when comparing models to data.
4. Models are not rejected because they dont match data. The mis match is studied. The causes are identified and the models are improved. Ask anyone here who has worked with detailed physics models.
SteveMosher
Agreed. We can only reject a very specific hypothesis. If the results of a run– or set or runs– or whatever– are inconsistent with data, then either
a) one looks into the data or
b) one investigates why runs don’t match “reality”.
(This is not an exclusive or. Both usually are done.)
That said: If it appears runs don’t match “reality”, you don’t proceed under the assumption they match “reality”.
But what if reality is wrong and is simply masking the true effects of atmospheric carbon with unfair, unknown aerosols and/or other unmodeled behaviors? If reality doesn’t follow accepted science, is reality a denialist?
OK that was silly.
But what if you could hypnotize a sample group of climate scientists to believe The Consensus is that climate sensitivity is 1.4 to 1.6 rather than 2.0+, what explanations would they offer for why certain models grossly overpredict temp increases. The compare those explanations to those of an unhypnotized control group…would the explanations be different in substance and points of emphasis?
Mosher, USCRN is supposedly a high quality temperature record that is significantly cooler than NOAA on average.
Are there any other high quality thermometer sets (which GISS is not) similar to USCRN that would meet your approval?
Steven Mosher:
Sometimes they are abandoned if they aren’t able to be used for what they were developed for. I’ve seen that happen plenty of times.
(The software graveyard is full of high-dollar products developed by government labs or corporations that never actually work.)
Steven Mosher
Re: “4. Models are not rejected because they don’t match data.â€
That raises the issue of Falsifiability in science.
When you say “The causes are identified and the models are improved†in common parlance, the models were “rejected†or “falsified†and then corrected or replaced. e.g., Copernicus gave way to Kepler, which was clarified by Newton’s gravity, which was extended by Einstein’s General Relativity.
The challenge with IPCC models is that they are in continual flux. How then can they be “validated†or “rejected� cf Lucia above.
Even with the updates, I understand Lucia to be showing that the IPCC model means and most individual models are now running ~2 sigma (>95%) hotter than the short and long term evidence. That gives me no confidence on their predictions being “reliable” a century and more into the future.
To borrow from Wiki,:
“George Edward Pelham Box[,] Professor Emeritus of Statistics at the University of Wisconsin, and a pioneer in the areas of quality control, time series analysis, design of experiments and Bayesian inference[, once wrote:]
‘Remember that all models are wrong; the practical question is how wrong do they have to be to not be useful.’ — Empirical Model-Building by Box and Draper, p. 74.”
In terms of what makes them useful (as opposed to needing to get buried):
I would define needed for a GCM as being able to reliably predict 30-year temperature trends (this is what is needed for policy related issues of climate impact).
We might also decide they need to be able to reproduce short-period climate variability, in order to be able to make test a prediction in 10-years. Too much short-period climate variability can lead to a heightened response of climate (possibly just transient sensitivity though).
This latter seems to be a active research question.
[The GCMs may not be useful if it takes 30 years to validate them.]
Lucia
I had posted a link to your post on climate models performance from last month and got the following reply from one of the geniuses who post at unScientific American. I thought it might make you giggle.
“It is pointless to analyze GCMs with a simple univariate ARMA time-series model as done by your blogger, Lucia. Is that all you can give us is a link to an anonymous blogger who might know something about time-series analysis, but knows nothing about GCMs. If Lucia has any significant expertise in statistics (other than taking an undergraduate class in time series analysis), Lucia would know better than to abuse the data in this way. I am sure that is why he blogs anonymously, so his colleagues do not discover his fraud.â€
I hope it made you giggle.
Rob–
Out of curiosity, which post did you link them to? I’d like to move your comment to the appropriate post.
Lucia
I linked to “Adding Multi-Model Means to Model v. Observations Graphs”. I agree I should have posted the comment there, but was not sure you’d see it.
Rob–
Don’t worry about it. I’m just going to move it so your comment attachs to the post you were discussing. Otherwise, readers can’t know which post you mentioned over at Sci Am.
If you are engaged in a live pissing-contest, you might want to let them know no “univariate ARMA time-series model[s]” were used in that post. Those spreads are based on standard deviation of repeat runs in models.
Comments re-opened.
I doubt he’s going to come, Lucia. But at least it’s worth a shot, and a nice olive branch by you.
We’ll see. He’s commented back at Scientific America, but so far hasn’t come here.
I don’t expect him to and at this point, other than for football practice, I couldn’t care less. I’m pretty sure if he can’t navigate the KNMI site to figure out what the models provide, that he doesn’t have enough knowledge nor ability to critically think to hold his own here.
He should stay at ClimateProgress (or whatever similar site he actually blogs at) where unthinking anti-intellectuals are actually welcomed rather than scorned.
This is in response to Cramer “over there”
http://blogs.scientificamerican.com/observations/2012/11/14/climate-change-denier-likely-to-lead-congressional-science-committee/#comment-14846
Lucia’s comments on this thread have been superficial and she has ignored my questions — acting like Sisko (Rob). [I am mostly referring up to the point when I said Lucia == Sisko.] I even acknowledged that I was wrong.
What question did I ignore?
I arrive here in comment 70. You addressed me in 76 asking if I came to defense Sisko/Starky. I answered that. You suggested I look at your comment in 73. I responded that I was confused because according to the blog, Bird posed 73. You posted no questions to me. In 76 you also accused me of being Rob/Starkey at that point. This amused people– their commenting really can’t be helped.
But as you can see above, I’ve engaged the questions you’ve asked me. If you have more, ask them. But please do so clearly so I know what they are. (Also, if you have the impression that I failed to answer a question, check to see that it wasn’t moderated away. Bird suggested some of your comments were moderated. If they vanish before I see them I can’t really answer them.)
I note you wrote this:
Cramer:
“[I am mostly referring up to the point when I said Lucia == Sisko.]”
It appears based on your comments that you are reading comments at my blog and then posting your questions here.
It would be more convenient if you just posted them at the blog where you are reading the comments
http://rankexploits.com/musings/2012/adding-multi-model-means-to-model-v-observations-graphs/
Cramer:
Unfortunately (or should that be fortunately), I have other committments and will have to restrict my comments to points that are more substantive. I’m here to learn new things–please provide me some knowledge of value. Examples:
Like you, I have time constraings. Since you are able to read comment at my blog, I’ll post the remaining portion of my answer to items in your comment over at my blog. You and others can read it here
You can read it here:
http://rankexploits.com/musings/2012/adding-multi-model-means-to-model-v-observations-graphs/
Now, on to substance.
I can’t begin to guess “how” you misunderstood. You can download the monthly data from runs used to create the projection in the AR4 at KNMI. The monthly data are available by filling out a form here:
http://climexp.knmi.nl/selectfield_co2.cgi?id=someone@somewhere
Load the link. TAS is surface temperature. select that, for the model you want. Otehr than that, follow directions for filling out the form, keep clicking, download the data and get it. It’s very simple– particularly for someone like you who claims you could tell by looking that the data are of decadal resolution. 🙂
Which Tamino analysis? There is a Foster and Rahmstorf paper, It has nothing to do with the CMIP3/AR4 model data. (Moreover, if you look at the graphs you can see that the nominal trend of 0.2 c/decade is outside the range consistent with data in some of his grapsh. He just doesn’t say so. And– it gets worse if we update to include data since the time that paper was posted. Ther are other problems with that paper, but mostly, with respect to the graphs Rob/Sisko linked, it’s merely looking at an entirely different question from the one I was looking at in my blog post. )
By Sakguch, do you mean
Koichi Sakaguchi, Xubin Zeng, Michael A. Brunke. (2012) Temporal- and Spatial-Scale Dependence of Three CMIP3 Climate Models in Simulating the Surface Temperature Trend in the Twentieth Century. Journal of Climate 25:7, 2456-2470
Online publication date: 1-Apr-2012.
That would appear to be a different topic. But maybe you mean a different paper? Or maybe there is something specific in that paper that relates to my figure? If so, you are going to have to say. Because I’m not a good mind reader.
lucia:
It’s funny you say this. Over on Climate Etc., willard has apparently decided I’m both Judith Curry and Dikran Marsupial.
Standard warning applies: willard’s comments are known to cause headaches in small doses.
Yes, that’s the question I have been trying to get you to answer. What question were you looking at in your “Observation V. Models” blog post?
If you’re simply claiming that your “Observation V. Models” analysis shows that the observed temperature trends over the last 10 to 15 years are lower than those predicted by the AR4 models. Yes, I agree; you are correct. I have never claimed otherwise. But those facts are obvious.
Sisko (Rob Starkey) used your analysis to argue that sea level would not rise more than 1 ft by 2100. He believes that your analysis proves that AR4 models overestimated predictions for 2100. The CMIP3/AR4 multi-model data at KNMI shows a 2.3K increase by 2100.
This could definitely be too high but I don’t believe your analysis proves that. If you also don’t believe it proves that, then take it up with Sisko (Rob), not me.
Regarding Tamino and Sakaguchi, you have the wrong analysis and wrong paper. I provided the links both last night and this morning — it is interesting that you overlooked them [i.e. Why should I take this dialogue with your seriously and expect any more than Sisko-style tribal behavior??? And that’s the exact behavior that I been observing here with all remarks]. I will not post the links a third time here.
You also discussed your “Tamino” analysis in 1st paragraph of your “Observation V. Models” analysis. It was titled “Arima11 Test: Reject AR4…” So I find it interesting once again that you did not know which Tamino analysis.
Cramer
You should have asked it. 🙂
The question is: Is the weather treated as a realization outside the range of “weather” inside an ensemble of model runs (with models from the AR4.)
1) I read comments at SA and I don’t think Rob claimed that. Could you get a quote.
2)My analysis says nothing about sea rise.
Where did you provide links? Not in the comment where you asked me the question. That comment is here:
http://blogs.scientificamerican.com/observations/2012/11/14/climate-change-denier-likely-to-lead-congressional-science-committee/#comment-14846
It contains zero links.
Could you provide a link to the links so that I don’t have to hunt for it?
I haven’t even seen them once.
My response to you there has been deleted. Possibly the links you posted are deleted. Bird thinks your question to me– which you complained I didn’t answer– was deleted. You might want to rethink your complaints and take them up with whoever deletes comments related to your conversations over at SA.
But if you want me to know what papers you are talking about you are going to have to
a) Describe Tamino’s as more than “Tamino” and
b) Give a title or something for Sakaguch.
If your question is “what Tamino analysis did I examine in my blog post called “Arima11 Test: Reject AR4…” That would be the Foster and Rahmstorf paper. But I would have had to be a mind reader to know that based on your question 89 over at SA.
Cramer–
BTW: Your memory of what you said and what Sisko said at SA could use some refreshing.
In comment 13
http://blogs.scientificamerican.com/observations/2012/11/14/climate-change-denier-likely-to-lead-congressional-science-committee/#comment-14724
you wrote:
In this statement, you sure as shooting give the impressing to be saying that my analysis does not show ” that the observed temperature trends over the last 10 to 15 years are lower than those predicted by the AR4 models.”
You are suggesting the results shown are due to cherry picking.
2. Sisko never says my results are about sea level rise.
3. You provide no links to anything by Tamino or Sakaguch– and in fact don’t even provide any titles or dates.
Are you sure you are hitting the “submit” button after you write long posts? Maybe all your posts are being auto-banned for some reason? Who knows. But you seem to be getting angry because you think people don’t answer your questions– which seems to happen because they are invisible to others. You seem to be angry that I don’t chastise Siski/Rob for saying things it turns out he did not say. And you seem to be angry that I somehow don’t read links you think you posted but which I never saw and in fact do not appear on that thread.
Possibly if you had taken my advice and switched to commenting here were posts don’t vanish after you write them, I would be better able to answer your questions or read the links. But alas, Christine Gorman /SA or whatever seems to ‘disappear’ stuff.
Cramer’s memory does seem rather selective, doesn’t it? I posted this on the SA blog
Since there is a tendency for comments to disappear, I’m cross posting here in slightly modified form.
…….
Cramer:
What!? This comedy of errors on Cramer’s part continues.
Here is the full AR4 model ensemble, available from KNMI explorer, neatly packaged for anybody who wants to see it:
https://dl.dropbox.com/u/4520911/Climate/ipcc.ar4.zip
These are annual averages of the full model for the period 1900 to 2099. More detailed output is available from KNMI.
Then this:
Cramer:
http://en.wikipedia.org/wiki/Lucia_Liljegren
Not anonymous.
I love this one too:
Cramer:
She provides full code and the data are available, so yes they are completely reproducible. The actual models given are also disclosed.
This assertion by Cramer, which is easily demonstrated to be false, and to have been made by the full knowledge of Cramer that he didn’t know whether it was true or not.
Cramer [again repeating the same claim:]
Labeling speculation as fact is a form of falsehood.
If there is code she hasn’t made available that one needs, your can email her, generally she will provide a copy.
It seems to be a bit of the pot calling the transparent crystal black to say that her work isn’t reproducible when you never even looked at it, had no actual knowledge of what she has done, never verified whether or not the code and accompanying documentation was available (I could the blog posts towards documentation of course), never spent more than two minutes even pursuing the KNMI site to find out that your absurd claims are totally unfounded.
Cramer also claims:
No, that wasn’t the point of the posts at all. Had you you bothered to click on the link provided, you would have learned that the point was to test AR4 models against observed data. (And to study the methodology for doing that.) And again you present speculative statements as fact, which is unquestionable a dishonest thing to do.
I don’t know why anybody who sees this thread would keep any respect for Cramer. I suggest he change his moniker.
Here is my most recent comment. It is more apropos to this thread. The previous comment was included to preserve some of Cramer’s rather wild claims, in the event that Scientific America decides to excise them. I think context is important.
…….
Cramer:
You don’t run simulations on KNMI, you pull simulation results from them… like the series I linked to above.
You then use the Monte Carlo method to simulate short period climate noise using the model ensemble provided. One easy technique that I use is to compute an Hann-tapered RMS-averaged Fourier transform of any of the model output, or if I need it, the data themselves, then randomize the phases and inverse Fourier transform.
Do this 10,000 and you have 10,000 instances that allows you to characterize short-period climate noise on the estimate of temperature trend. An other method (Lucia uses) is ARIMA based, but it is fully standard (the software package for ARIMA is part of the distributed R-statistically-oriented language).
Either way, this all involves standard statistical methods.
Cramer:
It doesn’t matter whether Tamino is bogus to discuss why it doesn’t apply. I presume you are referring of course to Foster & Rahmstorf, ERL 2011, which derives a temperature related series that purports to be related to the climate-related contribution to temperature by regressing temperature against several other measured physical quantities, MEI, aerosols and TSI, that are known to affect short-period temperature, and arrive a residual series,
Residual Series = T – a1 * MEI – a2 * Aerosols – a3 * TSI,
that is much flatter (less noisy) than the original three series.
The main point here is “T†is temperature, the “Residual Series†is not temperature, because it’s a quantity derived from temperature. One name might be “long-period climate indexâ€, there are others you could come up with, but one wrong name would be “temperatureâ€. It’s not temperature.
And warming relates to temperature, not to derived variables from temperature.
Regarding Sakaguchi (note spelling), I presume you are referring to
http://www.agu.org/pubs/crossref/2012/2012JD017765.shtml
It relates to hind-casting skill of models, not forecasting. So you’re looking at the period for which you have data available during the time period you were developing the model. Since the models are tuned, see this:
http://www.agu.org/journals/ms/ms1208/2012MS000154/2012MS000154.pdf
the degree to which the models can replicate the data is an important metric of how well they were able to replicate then-known measurements.
We sometimes refer to this in model building as “the verification processâ€. If a model fails to verify, it probably isn’t of much use in forecasting.
Simply because a model passes verification testing doesn’t mean it has any forecasting skill. To test this, you have to look at the period after which the models were developed and see how well the model forecasts (“predictionsâ€) compare with data that were collected after model assumptions were frozen in.
What Lucia finds is what basically everybody else finds, which is that the models predict too much warming. As I’ve pointed out above this has been a pattern for the models. As the models increase in spatial/temporal resolution and new data became available, the sensitivity of the models has been steadily declining from Hansen 1984 (4°C/doubling) to 3°C/doubling in AR4.
It will be interesting to see if it remains at 3°C/doubling or will continue this pattern and drop in AR5.
On a related note, if you start with:
Residual Series = T – a1 * MEI – a2 * Aerosols – a3 * TSI,
and replace MEI, Aerosols and TSI with three other band-limited signals, and fit for a1, a2, and a3, you’ll find that the “Residual Series” will be missing these frequency components, i.e., will be “flatter”.
That is an automatic result.
Since Cramer is discussing warming, I thought maybe a definition of warming was in order 😉 This is extracted from my comments on the SA thread:
Cramer = willard
For clarity and in all seriousness Cramer; please note that:
“Tamino” = Dr. Grant Foster of “Foster and Rahmstorf”
^^ this is a true fact which you can confirm by examining his email correspondences in the liberated Climategate dossier.
Carrick–
It’s interesting that
1) Bird’s comment discussing oral sex was not deleted
2) At least some Cramer’s comments appears– including some containing complaints that I do not provide answers are left showing
3) At least some of my answers to Cramer’s questions are deleted.
The fact that (3) occurred certainly tends to give the impression that I am not answering Cramer’s questions. If they were deleted before Cramer had a chance to see them this would certainly be so. I don’t know if any of Cramer’s comments were deleted– but he seems to have the strong impression he asked me questions that I do not recall reading, leaving links I never saw– neither of which currently appear.
Lucia,
Do I have to read the comments for you? And both at SciAm and your very own blog??? Try using the search function in your browser. For example, search “Tamino” and “claim.” Or try another browser. Maybe SciAm is not optimized for all browsers (but I don’t know how that would eliminate comments and links).
The Sisko claims were made over two SciAm articles:
http://www.scientificamerican.com/article.cfm?id=is-global-warming-happening-faster-than-expected
http://blogs.scientificamerican.com/observations/2012/11/14/climate-change-denier-likely-to-lead-congressional-science-committee/
Some comments have also been deleted. For example when Sisko links to your article on 11/01/2012 he said he had posted the links before, but that comment was deleted:
103. Sisko in reply to Bird/tree/dinosaur/etc. geek 01:34 PM 11/1/12
http://www.scientificamerican.com/article.cfm?id=is-global-warming-happening-faster-than-expected#comment-103
(note: links do not take you to exact comment — you have to browse your own way to comment 103)
——————-
I asked you many times “What claims are you attempting to make?”
see 73. Cramer 7:13 pm 11/15/2012
http://blogs.scientificamerican.com/observations/2012/11/14/climate-change-denier-likely-to-lead-congressional-science-committee/#comment-14809
——————-
Sisko has definitely made claims about forecast to 2100.
Why would anyone not talk about climate forecasts to 2050 or 2100??? Do you think we should only be forecasting to 2020?
67. Sisko in reply to Trent1492 09:19 AM 10/31/12
http://www.scientificamerican.com/article.cfm?id=is-global-warming-happening-faster-than-expected#comment-67
——————-
Tamino:
I am only using the description that you used on your own blog. Search for Tamino in your own blog.
http://rankexploits.com/musings/2012/arima11-test-reject-ar4-multi-model-mean-since-1980-1995-200120012003/
Tamino: 1 match.
Foster: 0 matches.
Rahmstorf: 0 matches.
http://rankexploits.com/musings/2012/arima11-mc-corrected-gistemp-trends-inconsistent-with-0-2cdecade/
Tamino: 46 matches
Foster: 15 matches
Rahmstorf: 12 matches
——————–
That’s all I have time for now. I don’t know why I have to do this work for you. I get the feeling you are screwing with me. I would rather think that than believe you are dilusional. You seem to want to make this all about semantics (an attempt to waste my time by having me continually repeat myself — I only have so much patience.)
For example:
You wrote and entire analysis on Oct 3 where you keep referring to the “Tamino” method and never once mention Grant Foster in your original blog post (only latter in the comments).
Carrick,
I don’t have time to go through all your attacks.
Regarding resolution, I was referring to the models themselves, not the output data from the models. Anyone can run a high frequency simulation for a model. The AR4 model simulations could produce output forecasts for every hour over the next century, but that does not mean the output data has anymore accuracy. However, high-frequency data is used (such as at KNMI) for interpolation purpose only, that does mean those “interpolated” points are accurate predictions for each month or year.
Regarding the 10,000 run simulation, I know KNMI does not offer simulations. That was the whole point of my rhetorical question. Lucia did not even mention in her “Observations V. Models” that she was using KNMI model data. Why? To me it sounded like she was running her own simulations. How? I don’t know why anyone would use a monte-carlo analysis for such a simple analysis.
Cramer:
“Regarding the 10,000 run simulation, I know KNMI does not offer simulations. That was the whole point of my rhetorical question. Lucia did not even mention in her “Observations V. Models†that she was using KNMI model data. Why? To me it sounded like she was running her own simulations. How? I don’t know why anyone would use a monte-carlo analysis for such a simple analysis.”
1. She doesnt mention KNMI becuase she has been doing this for about 3 years as a monthly exercise. So those of us who follow her are well aware of the data sources.
2. Why use a monte carlo? If you understood the problem you would not ask such a question. Go back and read all the work that has been done on the problem. In part, Monte carlo was used to settle various debates commeters here were having.
People who know that the problem is not a simple cook book look the formula up in a stats book.
Let’s go through each of Cramers comments and note the errors.
“Sisko, how many times are you going to link to Lucia’s inadequate analysis at rankexploits.com?”
1. Assumes facts not in evidence. Assumes that lucias analysis is inadequate without showing it.
“Lucia cherry picked the data. Even she does NOT make the claims that you do about her analysis.”
1. The data is the data produced by the IPPC
2. The time period selected is the SAME time peroid
that people use to say the warming is worse that expected
3. The time period is actually the correct time period to use
to test predictions of the models
“Why should anyone expect Lucia’s simulations over 138-month period would produce trendless data? Yes, the observed temperatures over the last 138 months have been relatively trendless, but this does not disprove the models.”
1. Lucia ran no simulations. The modelling groups did.
2. She never makes the claim that models are disproved.
If you wanted to that, pointing to failed sea ice projections
would be a better choice.
“Lucia’s results are not reproducible. I do not know exactly what models she ran or how she ran them.”
1. They are exactly reproduceable. Ask for the code. The claim that something is not reproduceable implies that you have tried and failed. Look the IPCC model results are not reproduceable.
I tried to run ModelE on my PC and it died. ha.
2. Did you ask?
Next Comment
“Sisko wrote, “The link I provided to the analysis performed by Lucia clearly showed that GCMs have forecasted a higher rate of warming than actually occurred as of this point in time.â€
Wrong.”
1. Cramer is wrong. Sisko is correct
“The link Sisko provided shows an analysis of the 138 months of trendless observed temperature data beginning January 2001 (i.e. datamining). Lucia ran unspecified simulations with 11 climate models that produced temperature trends. ”
1. Wrong. Lucia did not run simulations. never claims that she ran simulations.
2. 2001 is NOT datamining
A) to test predictions you use the start date of the prediction (2001)
B) The claim that the climate is “worse” than the models
relies on the same start date.
“Sisko’s conclusion is that since her simulations for that CHERRY-PICKED time period showed trends, but the observed data was trendless, then the models are wrong.
Sisko is also concluding 138 months of observed temperature data beginning in January 2001 is trendless, then climate change is not happening faster than models predict.
Is this the best that skeptics can provide?”
1. The models predicted .2C warming per decade starting in 2001.
2. That has not happened.
3. Draw what conclusion you like, but support it with math.
many conclusions can be drawn. The models are too hot.
shit happens. the observations are wrong. the models are
wrong. The actual forcings need to be accounted for. Many
explanations. Lucia and the vast majority of us here draw no
conclusion. We point to possible conclusions but since we
cant run the models ( that nasty reproduceable thing you complain about ) we can draw any firm conclusion. The same
holds when observations are worse that the models predict.
Next:
“Maybe if Sisko can’t defend her analysis, he could get her to post her own defense here at SciAm (he’s already engaged her about my comments). First, to be science rather than pseudoscience, her results must be reproducible. That is, she must give all the details on how she ran her simulations.”
Nice; I suppose Cramer will apply this rule equally to all science.
I’ll suggest that he start by contacting all the groups that ran simulations and ask them for there tuning data. or their forcing data. or their code. Perhaps he has missed the last 4 years of the debate when people her have been asking for reproducebale research from climate science and have had to resort to FOIA to get information.. only to find that the data doesnt exist anymore.
Next:
Sisko, please read what I have already said. I have pointed out valid points.
1) Lucia cherry-picks the data: 138 months of trendless observed data starting Jan-2001. You can not cherry-pick a decade to refute a 100 year forecast (i.e. 2100 AD). Similarly, I can more easily predict US GDP ten years from now, but I probably can predict what years we will be in recession.
2) Most importantly, Lucia does not give any details about her simulations. Therefore, it is NOT scientific. It’s only faith.
Lucia is not making outrageous claims that forecasts of sea level rise by 2100 are wrong because she cherry picked 138 months of data. YOU ARE.
If she provides all the details of her simulations and re-opens her comment thread, I will be happy to review her methodology and comment. She must give where she obtained the models (not all models are publically available, especially the fully-implemented versions rather than dumb-down versions) and/or provide the code/software. I really doubt she actually ran 11 fully implemented models.
That’s your problem, Sisko. Science is more than simply having FAITH in the a peer’s results. The methodology needs to be specified and reproducible along with the results. That’s what it means by “peer reviewed.â€
#################################
well you all should be able to pick apart the problems point by point. What is amazing is that Cramer voices the very complaints we have had ( and taken legal action over) about some in climate science. I wonder how he would react to Nic Lewis findings WRT Forest. I bet he would make excuses.
Next:
“Bird, I would suggest you refer to Sisko as Sisko as not to confuse people. “Rob Starkey†is also probably an alias. As a true skeptic I do not believe anyone. Pseudo-skeptics like Sisko believe anyone with the same ideological bias, just like Sisko believes the anonymous blogger, Lucia, who doesn’t even provide the details of the simulations in her analysis (not to mention her cherry-picked data).”
True skeptic. He doesnt believe anyone. Well skepticism invloves more than just knee jerk disbelief. In includes.
1. Avoiding generalization ( all psuedo skeptics believe anyone
with the same ideological bias )
2. Checking facts: Lucia is not anonymous. She is not a skeptic
3. Researching completely before spouting off: She didnt run simulations or claim to. The start date is not cherry picked. It is the start date of predictions. The date you use to show
A) the models are worse than observations
B) the models are better than observations.
Most importantly a skeptic must be skeptical of himself. Cramer is not.
Next:
“Lucia simulations:
That is not what Lucia has done. What she has done is attempt to backtest a very small sample of recent trendless data. And you claimed that this is evidence that these models are useless for making forecasts to 2100. Currently, Europe is in recession. If my models could not predict that, it does not mean that my models can not forecast the GDP ten years from now within a specified confidence interval (data can be is noisy — timeframes are important!). For your info, the climate simulations that Lucia ran can not be applied to times frames shorter than three decades. For higher resolutions you need a supercomputer running more detailed models. This is why she will not provide the details of her simulation, because she knows they are bogus.”
1. Cramer still doesnt get what Lucia has done.
2. he makes a false claim about tests not being applicable
for timeframes less than 3 decades. He should beat up
Trenberth.
3. He assumes to know what is in Lucia’s mind. Lousy skeptic
Cramer believes himself, when he claims to be a skeptic
who believes no one. Motive seeking is the first clue
you use to spot a fake skeptic like Crame
“Your sea-level rise claim (read entire thread):
http://www.scientificamerican.com/article.cfm?id=is-global-warming-happening-faster-than-expected#comment-103
Sea level rise:
Have you heard of acceleration?”
Not related to Lucia
“Bible:
I never said anything about the bible. You could be an atheist for all I know. But there is proof that you have FAITH in Lucia, because she has provided ZERO details on her simulations and you accepted them as the truth.”
1. Rob has been here long enough and read all the details that you avoided to read.
Yes, this back-and-forth is all about semantics. And I have ran out of patience. I am very forgiving of people who make grammatical and spelling errors in a comments dialogue. Things obviously can be misinterpretated. But many people make an honest attempt to ask questions and try to understand what is being said (and they say they do not understand what the author is attempting to say as I have said). Others seem to use these errors (e.g. dangling prepositions) as a basis for attack by feigning a false interpretation.
For example, Lucia attack on Nathan for being unclear (whether he actually was or not is not the point).
lucia (Comment #106523) November 16th, 2012 at 6:57 pm
However, in original blog posts (SciAm-Christine Gorman and here), more care should be taken. This discussion has been about Lucia’s “Observations V. Models.” And in her very first senctence she forgets the verb:
We what? We model? We fudge? (I know what she meant, but I not going to feign a bogus response).
and she doesn’t even reference her use of KNMI model data (doesn’t matter if she’s been doing for 3 yrs, it’s a reference).
Talk about be unclear.
Yes, as Lucia describe at SciAm, she has very good tactics to chase away “trolls”, but now all she has is a tribe.
“Resolution of models included in AR4:
http://www.agu.org/pubs/crossref/2012/2012JD017765.shtml
This research claims CMIP3 and 5 models (which includes the AR4 models) can not be used for periods less than three decades.
What scientist is claiming that the AR4 models could accurately predict the Lucia’s 138 month period? Otherwise your position is simply a straw man.”
1. As carrick notes this is about hindcast not forecast.
2. The scientists who claim that the models are accurate within
short time periods would be those who calculated the error bars around the model mean. Go read Ar4. If you want something
more recent. read trenberth.
And my last comment applies to Steve Mosher. I am not going to go through all your INTENTIONAL misinterpretations of what has been said.
The whole debate was about Sisko claiming that sea level would not rise more than 1 foot by 2100 and Lucia’s analysis gives the “real facts” that prove it.
If you want to DEFEND that, go right ahead and waste your time.
See:
http://www.scientificamerican.com/article.cfm?id=is-global-warming-happening-faster-than-expected#comment-103
and the rest of the comments made over the last few weeks.
Next:
“Sisko, are you saying Lucia Liljegren’s analysis was not backtesting? Hmmm. She used 138-month observed data starting Jan-2001.
The basis of my claim is that data is noisy. The models are used for forecasting longer periods.”
1. the models are not used to ‘forecast’ longer periods. The models are used to forecast. At shorter time periods the data is noisy. That means you have wider error bars. That noise can be characterized. That is what Lucia is doing.
2. Lucia is not the only scientist looking at shorter periods. You should do some reading.
“The basis of your claim seems to be that, hmmm, we have had zero increase in temperatures over the last decade, therefore GW is not happening for the longer term.”
Not related to Lucia
“Yes, your last statement shows that you do not understand acceleration. You ask “when will the rate of sea level rise increase?†It’s not a matter of when, it’s continuous. It’s rising faster now than 20 years ago, and it will be rising faster in 20 years than now. Your the first engineer that I know without an understanding of calculus.”
Interesting. not related to Lucia, but your analysis here probably needs some due diligence. Show your work.
Next
“This is pointless discussion. You have very little knowledge (you are not an aeronautical engineer). I have respect for Dr. Liljegren. She seems to be open-minded and a straight shooter. She is not making the same claims that you are about her analysis. If she provides the details of her simulations, I will be happy to discuss them with her (even on her website).”
Cramer discovers that he should have practice more skepticism WRT his previous opinions of Lucia. He should be aware that many aero engineers hang out here. I dont think Rob is one of them.
Next:
“Sisko, I don’t know your ideological bias. I just know that you have one because you are not a skeptic — your main objective is simply to have faith in things that confirm that you are correct in your beliefs.”
Cramer practicing fake skepticism WRT his own beliefs.
Ironic I think.
“Sisko,
You do not question AGW, but you do not believe sea levels will rise more than 1 ft/century or that temperature will rise anymore than 0.1C/century? And you also don’t believe CO2 is relevant to global warming. Did you know that the A in AGW is for anthropogenic?
Bird,
Lucia never claimed that her analysis shows that AGW does not exist as Sisko has. She is actually quite intelligent (unlike Sisko). I would have to read more of her blogs before I come to a position that she is just another closed-minded denialist.”
Cramer begins to see the light. Maybe he is not so closed minded.
Of course everyone who follows this debate knows that Lucia is a luke warmer and we believe in AGW. Cramer should read more and comment less.
Next
“Sisko,
I don’t know if you like to spin things or are dilusional. What else would “Arctic sea ice decline: Faster than forecast†mean except as you said, “a comparison of the actual observed trend vs. what had been projected in prior analysis.â€
Here’s the 2011 paper:
IPCC climate models do not capture Arctic sea ice drift acceleration: Consequences in terms of projected sea ice thinning and decline; P. Rampal , J. Weiss , C. Dubois , J.M. Campin; Journal of Geophysical Research , vol. 116, 2011.”
weird. Now we see that it is ok to compare forecasts with the observed values. Lets see if I get this right. You are allowed
to compare forecasts of Ice with observed ice to say that
“Its worse than we thought”. BUT you cannot compare observed temperature with forecasted temperature, and say “its better than expected”
Cramer= fake skeptic.
Mosher is cramer actually Willard? Willard usually at leasts asks correct/non-projection questions.
Next:
“Lucia said, “I’ll go write a post that Cramer will say uses a cherry picked start date. Then maybe he can suggest one he’d rather see.â€
You seem to have missed my point about model resolution. When I look at the model data at KNMI, it obvious they are of very low resolution — like three decades, not one.
Arguments based on your analysis of 138-month period starting Jan2001 seem to be of same nature as that made by the Daily Mail: “Global warming stopped 16 years ago, reveals Met Office report quietly released… and here is the chart to prove it.â€
1. Carmer appears to have no knowledge of simulation.
2. The temporal resolution of the typical GCM is 30 minutes.
3. Files are posted with daily, monthly etc values.
4. My guess is he is computer illiterate. Kinda like thinking that Rob Starkey is Lucia.
last comment:
“Is anyone going to discuss anything rather than trivial points, such as my off-the-cuff comment about the possibility of Lucia (on this SciAm thread) being Sisko? The reason I made that comment was base on my perception that Lucia is intelligent (and I am not saying that I have revoked that position). Lucia’s comments on this thread have been superficial and she has ignored my questions — acting like Sisko (Rob). [I am mostly referring up to the point when I said Lucia == Sisko.] I even acknowledged that I was wrong.
Unfortunately (or should that be fortunately), I have other committments and will have to restrict my comments to points that are more substantive. I’m here to learn new things–please provide me some knowledge of value. Examples:
Point me to where I can learn how to run 10,000-run simulation at KNMI or tell me how I misunderstood. Tell me why the Tamino analysis is bogus or why the Sakguchi, et al study does not apply to CMIP3/AR4 model data at KNMI.”
###########
bottom line. Cramer stepped on his dick. unlike a good skeptic he did not research before opening his mouth. he still thinks you can run simulations at KNMI. But above he says the following
“Regarding the 10,000 run simulation, I know KNMI does not offer simulations. That was the whole point of my rhetorical question.”
Jesus christ kid do your homework. I suspect you know nothing about simulation. You clearly open your mouth before engaging your brain. you went to KNMI and didnt understand what it is.
You skimmed posts at Lucia’s. You have no clue whatsoever that carrick will eat you alive in an argument and has a record of scientific achievement that most of us would kill for.
Stay if you want to be schooled. Most of us enjoy learning what we dont know. Drop the crap about ideology and show your work or go back to your moms basement, which is what SciAmer amounts to.
Opps.. compile job is done.. gotta run
Mosher,
Are you claiming that the model data at KNMI has a resolution of 30 minutes? (rhetorical)
What did I say about INTENTIONALLY misintrepreting things?
See my comment to Carrick.
The CMIP3 model data at KNMI has a resolution of probably no better about 3 decades. All other points are no better than interpolations. That’s why the trends are so smooth. I can run a simulation with a frequency of 1 second, but who cares (except to use it for interpolation–are you going to try to INTENTIONALLY misinterpret this too?)?
It’s funny how accurate you think these models actually are (being facetious). But both you and I know they are not acurate — you just like BS’ing if its about tearing down someone else.
Here’s a wattsup blog post about Sakaguchi et al paper where you actually comment.
http://wattsupwiththat.com/2012/09/18/climate-models-shown-to-be-inaccurate-less-than-30-years-out/
I won’t call you dumb like you called me. I know your smarter than that, but you just have a MO of attacking the enemy no matter what. And that’s ashame.
——-
And BTW Steve, I would stop your tribal diatribe because I doubt I have read anymore than 10% of your INTENTIONAL misinterpretations and assumptions about what I said and who I am.
But if it make you feel better, have at it.
Get it right Mosher. I never thought you could run simulations at KNMI. My question was rhetorical because Lucia failed to list her source of data. (Wow! I don’t care how long she’s been doing this analysis). She the one that said she was running simulations. If she brings her analysis up to publishing standards, I will take a more detailed look to see why she has to do simulations for such a simple analysis that looks at model data that is so smooth that you could luge on it.
Cramer –
“The CMIP3 model data at KNMI has a resolution of probably no better about 3 decades.”
Perhaps you could give us your definition of “resolution” in the context of these simulations. You keep using that word. I do not think it means what you think it means.
Cramer:
This is typical of verbal bullies like yourself. Make a series of wild accusations, then treat the mere explanation of how they are false as “attacks”. Boohoo. You wrote the statements that have been reproduced faithfully that now indict you as intellectually dishonest. Very little is required on my part in interpretation, I’d suggest you refocus your defense of your own words away from other people *cough* INTENTIONALLY mis-representing something that is properly quoted and place in as good a context as anybody could put it.
When you are able to explain how you, in an intellectually honest manner, arrive at the conclusion that Luca was Rob Starkey or that her code was not reproducible or any of another bits of tripe you’ve written we can move on. Or you can admit to your many non-trivial errors, and we’ll move on.
Keep reminding yourself we can read. Maybe that will help.
Cramer seems never to have learned the first rule of holes…
Mosher, along with all the other comments that shows insecurity (Carrick will eat you alive, drop the ideology, show your work), ends with “Opps.. compile job is done.. gotta run.”
Why include the info that your compile job is done? Is that suppose to impress me? Who are you trying to impress?
I know of you and Lucia. I know your educational background. I haven’t determined if you even have PhDs–I would guess not. I know you are both academic rejects (Lucia from Iowa State, you from UCLA if I remember correctly). And the “work” that I see on this site explains that.
Stay on the fringe, that’s where you belong.
—–
BTW, SciAm is not a scientific journal. It’s a news magazine — it’s a simple place to go to read NEWS from other fields outside my area of expertise. I use many different sites for very quick science news. Sciencedaily isn’t that bad. Occasionally, unfortunately, I get drawn into the comments.
Yes, climatology is NOT my field of expertise. But I do have a BS Chemical Engineering (Illinois), MS Statistics (Chicago), PhD Mathematics (Courant). Maybe if you google that combination, you can determine my identification — Cramer isn’t even close. Trust me, you will know me. And guess what, I have already “eaten you alive.”
Carrick,
That’s your case about me?
“Lucia was Rob Starkey or that her code was not reproducible.”
An off-the-cuff remark not any different then “Cramer = willard” by Mosher.
And Lucia does not reference her source data, but it’s somehow suppose to be reproducible.
What a joke. I’m done wasting my time.
http://medical-dictionary.thefreedictionary.com/superiority+complex
It’s tragically comic to see Mosher arguing with a recalcitrant pea-brain like Cramer.
In a perfect world Carrick and McIntyre would rightly tell us what to do – even their corruption by ultimate power would be better than the utterly ignorant shite that Scientific American disseminates.
Exactly John M.
The quotes means I was quoting Mosher:
Cramer: And Lucia does not reference her source data, but it’s somehow suppose to be reproducible.
Can you describe exactly what effort you’ve made to reproduce that which you say is not reproducible before proclaiming it so?
Cramer:
She does reference her source code, you obviously didn’t look and you are now asserting that you did.
That is tantamount to lying.
By the way, Cramer, I put together this zip file:
link
as a result of a “non-existent” referencing of the data she used.
🙄
Suggestions: Don’t assume this was a single post on Lucia’s part. Don’t assume we’re all st*pid. Do assume that making off the cuff remarks will get you eviscerated here. Do find the Categories section on the right hand side of the main page of the blog. Maybe started with Data Comparisons.
Stop trying to defend the indefensible. You won’t win arguments that you can’t win.
Cramer–
Goodness. That’s a whole bunch of stuff there. But I think you might need to refresh your memory and read what you claimed.
You are providing a whole bunch of links that shows Sisko made claims about sea level and but none where he claims my analysis says says anything about sea level. In fact, it appears he made no such claim– as I wrote previously. If you can locate anything where Rob claims my post shows sea level rise is slower than predicted, please quote the text. Because you keep linking an entire article and I can find absolutely no text where he claims my analysis says the sea level is rising slowly.
As for your continued claim about having written something in comment 73: This is how comment 73 displays on my browser:
I’ve told you comment 73 was by bird several times now. That’s how it displays for me. Moreover, on my browser, bird posted 71 at 7:00 pm 11/15/2012 and I replaied with 72 at 7:31 pm 11/15/2012. Moreover, the source code for the page agrees with my browsers interprestation. If your browswer shows “73. Cramer 7:13 pm 11/15/2012” that is a mystery.
I’m also mystified by your inability to find references to “Foster” or Rahnstorf at this blog. Here are the results when I do the query:
Foster: 90 results.
I can also find lots of references to Rahmstorf.
If you want to have a substantive discussion about something meaninful we can. Or not. Do you have any substantive questions about my post? And can you state them directly?
By the way: believe it or not, rhetorical questions are not permitted at this blog. They result in miscommunication. I had no reason to believe your question about how to download material from KNMI was rhetorical. You had repeatedly complained I didn’t answer your questions so I answered it.
Now: Is there something you want to know about my post? Or not? Cuz it’s Saturday night. I need to sit on the couch and knit.
Carrick,
Suggestion:
Take your issues up with the SciAm troll, Sisko (Rob Starkey). He linked to that “single post on Lucia’s part” and made some outrageous claims about it proving that sea level would not rise more than 1 ft by 2100.
You seem to be forgetting Sicko’s claims every time you have read something that I wrote, because when I was writing, it was only in the light of those outrageous claims. And that’s why I have continuously asked Lucia what claim she was making. This has been mostly about Sisko, not Lucia.
I have assumed nobody to be stupid (I even wrote that I thought Lucia was intelligent a few days ago at SciAm).
Why should I have investigated more info back to 3 yrs ago on this blog??? I agree that the AR4 models significantly underestimate the temperature trend for the last decade. I don’t believe that has much to do with predicting sea level rise by 2100.
So, please tell me just what I am trying to defend.
Cramer
It’s a game Lucia plays. She keeps her scope so narrow to be meaningless, and refuses to expand it.
So she’ll investigate whether the model trends match the real world trend over a 30 year+ interval, and when she sees they don’t match, she claims they’re wrong, and we can’t trust them.
She then stops. And does no more. She doesn’t investigate why, nor what it means, nor whether it is important.
Many many authors have published reams of material on this, and mostly they support the notion that the models are ok. And as you point out their resolution is not designed to look at the detail she’s investigating.
I would suggest simply ignoring all the other remarks about mistaken identity and whatever and just stick to the basics of what models can be fairly expected to do.
Lucia,
I do not know what is going on with the comment numbers. Please try another browser. What you are saying is 73, I see as 72. But it shouldn’t be that difficult since I did provide the time stamp info. Please look for them, if you are interested. If not, you can believe whatever you want about Sisko and his trolling at SciAm.
The only thing I ever wanted to know was what are you trying to claim? Which you still never answered. You did however say that you believed Sisko’s claim about 1 ft sea level rise by 2100 was low. So I guess we got that far.
I guess you can allow whatever you want on your blog. I respect that.
Nathan
“So she’ll investigate whether the model trends match the real world trend over a 30 year+ interval, and when she sees they don’t match, she claims they’re wrong, and we can’t trust them.
She then stops. And does no more. She doesn’t investigate why, nor what it means, nor whether it is important.”
1. I’ve never seen her claim they are “wrong” none of us would, except to say that all models are necessarily wrong.
2. Ive never seen her say you could not trust them. The issue is can you use them for a given purpose.
3. You cannot investigate why a model doesnt match reality without the ability to run the model. you might have ideas or speculations, but they are just that.
Basically nathan you operate using bad faith. you attribute ideas to people that they have never expressed. you are as bad as any fake skeptic on Goddards site or wuwt. back to your moms basement, troll
Cramer
“The only thing I ever wanted to know was what are you trying to claim?”
This is the problem, Cramer, you’ll never get to the nub of her complaint. These are just ‘investigations’ they don’t have any particular claim or point. Unless it’s ‘falsifying the IPCC projection’, but even that is pretty meaningless as the IPCC projection she tests isn’t the actual IPCC projection.
Originally, when she started with this ‘Lukewarmer’ theory it was that ECS was too high in the models… But that seems to have taken a back seat now to ‘weather noise is underestimated’.
Cramer you numbering is off. Just cut and paste your comments.
I already did that for you, it doesnt take long.
Yes, Nathan, this has been a complete waste of my time. I came here to get views from others, and all I found was tribal behavior.
Cramer:
I’ve given you a pretty long list of indefensible statements on your part. You don’t need to invoke anything Rob Starkey said as a defense of your own outlandish claims. As far as I know, the only error he made related to claims about arctic sea ice, and, as that has been already hashed out, I see no reason to revisit it.
If you don’t want to defend your statements, the cure is simple: Stop defending them.
Mosher,
Really? She doesn’t falsify the models? Oh, she falsifies the IPCC projection, that’s right… So yes, can’t say anything about individual models.
She doesn’t think that we can’t trust them for ‘planning’?
“The issue is can you use them for a given purpose.”
So what is that purpose?
Fergal.
part of the reason for my post was to put all of cramers comments here so he could stop playing the stupid game of ‘search for what I said”
You can all see his comments in order. You can all see his failure to do a modicum of investigation before spouting off. fake skeptic who believes his own BS. Lucia must be a denialist. The crap about KNMI.
lunchmeat. I was hungry. I’m still hungry. next.
Cramer, don’t leave. Just stick to the models discussion and ignore the trivial rhetoric
Nathan:
Let’s be really blunt about this, shall we?
You don’t like the results of the analysis, so you nitpick on things to find wrong with it.
As I’ve said above, the point here is to test AR4 models against observed data, and to study the methodology for doing that.
These are questions I and other people find interesting. If you don’t, well frankly you can just go f**k off.
Hi Steven,
I will reply to your comment #106604 since it contains reasonable statements. I do not mind be told I am wrong.
I don’t believe that the numbers being off is relevant to the solution. I provided the time stamps. I am not going to cut and paste ten comments. Then she will just say those comments did not exist. I have spent enough time on this. I feel I’m being played so I do not plan to spend much more time on this.
Mosher
Here is Carrick’s take on Lucia’s investiagtions:
“What Lucia finds is what basically everybody else finds, which is that the models predict too much warming. As I’ve pointed out above this has been a pattern for the models. As the models increase in spatial/temporal resolution and new data became available, the sensitivity of the models has been steadily declining from Hansen 1984 (4°C/doubling) to 3°C/doubling in AR4.”
So, basically this is a statement that they are wrong and untrustworthy, no?
And then a claim that ECS is too high.
BUT as is usual on this blog there is a complete failure to actually investigate ECS. Knutti and Hegerl, 2008 show that the recent temps are a poor way to determine ECS as the models show pretty much the same response for low and high ECS.
ECS is determined from a range of investigations.
Nathan:
You can choose to ignore the nontrivial nature of Cramer’s errors, as that is par for the course with your group.
I can promise you I and others won’t.
I’m going to do a Mosher here and identify Nathan as this deluded sock puppeteer: http://www.butterfliesandwheels.org/2011/his-elbow-slipped/
Nathan
“Really? She doesn’t falsify the models? Oh, she falsifies the IPCC projection, that’s right… So yes, can’t say anything about individual models.”
1. err no she doesnt falsisfy the projection. There are many reasons why you might get a 3 sigma event as she has pointed out.
sometimes a 1/20 thing happens. That was made clear at the OUTSET of this series years ago. when you lived in your moms basement.
“She doesn’t think that we can’t trust them for ‘planning’?”
1. I don’t think she has ever made a statement about the issue. Personally I think you can use them for planning certain things.
even though they are wrong. Its called using the best information you have. Ask your mom about it.
“The issue is can you use them for a given purpose.â€
So what is that purpose”
1. Dont be stuck on stupid. The person who wants to use the model specifies the purpose. So, for example, I would say that the prediction of a 1 meter rise in sea level is useful for long range planning. The amount of undeveloped land on us coasts that is below 1meter and undeveloped his > 50%. I would suggest that one not plan to build there until we have a better handle on uncertainties. Simple.
next
Carrick
I am asking him to go straight to the heart of the issue.
That’s the important thing, not whether he incorrectly assumed she referenced things.
Nathan:
To paraphrase…
All models are wrong. The question is whether they are so wrong as to not be useful.
The fact they fail to validate means there is still work to be done. That in itself doesn’t mean they’re worthless, but it does put them under some stress to improve. Which is good, it means we’re learning something new.
In a moment of confirmation bias, somebody might dismiss temperature as a not important variable. But I’d invite you to cut and paste the exact text from this article that you think makes this claim rather than just paraphrasing it (or at least cite the page, column and paragraph where you think it is made).
Temperature is one of the most important metrics for model performance because of its relevance to human activity. I would hope that Knutti wouldn’t make as silly a claim as you suggest he is making.
Cramer – again, how much time and effort did you expend trying to produce the analysis above before proclaiming it unable to be reproduced? This is a simple question.
Nathan
“What Lucia finds is what basically everybody else finds, which is that the models predict too much warming. As I’ve pointed out above this has been a pattern for the models. As the models increase in spatial/temporal resolution and new data became available, the sensitivity of the models has been steadily declining from Hansen 1984 (4°C/doubling) to 3°C/doubling in AR4.â€
So, basically this is a statement that they are wrong and untrustworthy, no?
#####################
no a statement that they are untrustworthy would be something like this “they are untrustworthy” Now carrick may believe that.
But you should ask him. I would guess he would say something more nuanced than that. Further he is observing a simple fact. the mean estimate of sensitivity is falling. Go figure.
“And then a claim that ECS is too high”
that may in fact be a cause of the mismatch. You havent read Ar5. You havent read any of the latest papers on ECS. the trend in the science is toward a lower number.
“BUT as is usual on this blog there is a complete failure to actually investigate ECS. Knutti and Hegerl, 2008 show that the recent temps are a poor way to determine ECS as the models show pretty much the same response for low and high ECS.
ECS is determined from a range of investigations.”
please update your reading from 2008 to 2012. I’ll suggest that you become a reviewer for Ar5. That will give you access to plenty of stuff. Yes there are multiple lines of evidence.. all trending down.
Nathan:
People making numerous false claims then not owning them is the heart of the matter to me. If I catch a person repeatedly making false statements, and then refusing to admit to it, I am not take anything that person says seriously from that point.
Reliability is a major issue in discussions in science.
Steven Mosher:
I see your Jedi mind reading tricks are more effective than Nathans. (Excuse me “ENGLISH reading tricks.”)
That is exactly all I was saying.
One of Nathan’s little tricks is to state something somebody else says then adds spin to it that wasn’t there originally. Clever boy for thinking nobody ever noticed that.
“TerryMN (Comment #106620)
November 17th, 2012 at 7:32 pm
Cramer – again, how much time and effort did you expend trying to produce the analysis above before proclaiming it unable to be reproduced? This is a simple question.
#########
he spent no time. He learned a stupid pet trick. Say a result is not reproduceable. that makes you sound sciencey. of course, since he never gets out of his moms basement, he didnt know he was in the thick of it with people who actually spend time reproducing the work of climate science. fake skeptic, doesnt doubt his own knowledge or ability
Mosher
Most of her earlier posts were entitled “Falsifying the IPCC projection” back in 2008.
She claimed to me that it was important to investigate the IPCC projection as it was used for planning…
You are right in saying that is not a reasonable use.
““The issue is can you use them for a given purpose.â€
So what is that purposeâ€
1. Dont be stuck on stupid. The person who wants to use the model specifies the purpose. So, for example, I would say that the prediction of a 1 meter rise in sea level is useful for long range planning. The amount of undeveloped land on us coasts that is below 1meter and undeveloped his > 50%. I would suggest that one not plan to build there until we have a better handle on uncertainties. Simple.”
OH MY GOD, what? She was looking at temperature models for the world. Has she ever investigated the models for anything else?
What is their purpose? I would suggest it is to try and estimate the future global temperature. Their resolution isn’t relly high enough to give you much elseok, so your original statement, before you meander off into some other realm was:
“2. Ive never seen her say you could not trust them. The issue is can you use them for a given purpose.”
so she doesn’t say you can’t trust them, but that the issue is whether they can be used for that purpose.
OK, so her claims become that they’re not suitable for the purpose. How is this different from saying that they’re not trustworthy?
ya carrick, you would think that anybody who read this blog and the exchanges we have with our good friend Nick would know that its all about the nuance and detail.
and further when we go too far, we admit our mistakes. imagine that.
I like the way Cramer and Bird pretended to be busy with real science. Known as the false time constraint.. too busy with my beakers to respond…
Mosher
So what is the estimate now?
Mosh – that’s my surmise as well; that he spent no time or effort before making that assertion. But I would really like to hear from him whether he did and if so – what were the steps he took before giving up and making the accusation. Or if he ignores the question. Gives a window to character and tenacity…
Nathan you still dont get it.
The purpose is not to estimate the global temperature. The PURPOSE is what the policy maker wants to use the information for.
See my example.
The tie between modelling and policy is underscored by the new interest in regional simulation. basically, policy makers dont know what to do with a global estimate. they want regional data.
How much warmer will it get here? what will happen to farming? crops? floods? thats the planning and purpose dimension.
read more. comment less.
I’ll suggest you read the MET/CRU site where they provide climate simulations for a variety of users.
Carrick,
Mosher has already dismissed it, but here is the text:
Figure 4 The observed global warming provides only a weak constraint on climate sensitivity. A climate model of intermediate complexity3, forced with anthropogenic and natural radiative forcing, is used to simulate global temperature with a low climate sensitivity and a high total forcing over the twentieth century (2 °C, 2.5 W m−2 in the year 2000; blue line) and with a high climate sensitivity and low total forcing (6 °C, 1.4 W m−2; red line). Both cases (selected for illustration from a large ensemble) agree similarly well with the observed warming (HadCRUT 3v; black line) over the instrumental period (inset), but show very different long-term warming for SRES scenario A2 (ref. 101). For simplicity, ocean parameters are kept constant here.
Mosher
So the policy makers ARE using GCMs for whatever purpose they want?
So the issue is whether they are fit for purpose, which is whatever the policy makers use them for…
I think you need to articulate what you mean more clearly.
Nathan:
There goes Nathan again adding false pejorative nuances to what other people have said and not expecting people to notice it.
Lucia isn’t perfect, I don’t think she would make that claim, howver, I do think she would say she has learned as she has gone on with her investigative process.
Nathan on the other hand, as far as I can see, is still stuck on zero. It’s what he earns for being snitty instead of having an open mind and thinking new thoughts.
For somebody who really thinks everybody here is dumb, he sure spends a lot of time spewing pure nonsense.
Mosher
You are drifting away from your original statement.
So here it is again:
“2. Ive never seen her say you could not trust them. The issue is can you use them for a given purpose.â€
ok so this was in regards to GCMs… so what is their given purpose? Are you saying that the GCMs are used in local planning?
Carrick
“Lucia isn’t perfect, I don’t think she would make that claim, howver, I do think she would say she has learned as she has gone on with her investigative process.”
Yes, I am sure we have all learnt things along the way.
I was asking why it was important for the GCM’s to match recent data or for the IPCC projection to hold (@ 0.2C/decade for the next two decades), and she said it was used for planning.
Mosher may be attemping to claim the same.
I don’t think I am stuck on zero. Whatever ‘zero’ is.
Mosher.
What is the current ECS according to your reading?
Nathan, thanks for the text and I’m relieved that Knutti didn’t say what you claimed they said.
As usual, you have your valance backwards in your interpretation:
Comparison of backcast to measured temperature is a weak test in the sense it’s easy to replicate historical temperature patterns (hind-casting). It is a weak test in the sense of a “failure to reject” problem, and is the result of the classic model tuning problem.
What we’re talking about here something entirely different, which is “failure to validate”.
Were it a weak test of models to compare observed versus forecast temperatures, then failure to validate would be an even more serious problem for the GCMs.
I am pretty sure that Knutti wouldn’t describe forecast vs. observed temperature as a weak test of GCMs. It is, in fact, probably an overly strong one. (That is: A model need not be rejected as usable simply because it fails to validate against global mean temperature.)
Seriously, Nathan, if you could even keep your basic facts straight, it would streamline arguments enormously. If for no other reason than you’d be writing 1/3 of the comments you write now, realizing that you’d made an implausible leap of logic along the way…
Nathan:
I’ll address that:
Climate sensitivity is 3C
Carrick
“The observed global warming provides only a weak constraint on climate sensitivity.”
“Knutti and Hegerl, 2008 show that the recent temps are a poor way to determine ECS as the models show pretty much the same response for low and high ECS.”
Are these really that different?
Carrick, they’re showing that what we observe is consistent with pretty much all the reasonable/central estimates of ECS.
What I am saying is that if you are attempting to use modern temps as evidence of low ECS, then you are probably wrong.
Nathan:
For most of us, I would agree. What would you name as the most important new thing you’ve learned in four years?
Carrick
HOORAY! WE AGREE!!!
🙂
Nathan:
Weakness in comparison of historic to backcast temperature is a result of knowledge of historic temperatures, and as I said results from model tuning. Since we’re comparing forecast temperatures to as then unknown future measured temperatures, there is no way this comparison can suffer from the same fault, short of a time machine.
So yes, these are completely and utterly different things.
Ummmmmm
That’s tricky, you mean only in the reading of blogs on climate change?
That’s tricky. I’ll have to get back to you…
Nathan:
I don’t know that you can really say they are wrong.
I think you can say they are making a JUDGEMENT without complete information, but that doesn’t mean the judgement is wrong (we’ll have to wait and see won’t we?).
My take is, I expect GCM sensitivity drop again as they adjust the models to fit the newest data.
Now—nuance time: That doesn’t mean I think they are correct to adjust it down. I don’t really have a bloody idea. If I did, I’d be king of the world (so you are really lucky I don’t). I just think it’s plausible that they will adjust it down. I think it is equally plausible in AR6 they will shift it back upwards.
Carrick
Ok, so how can you say that the modern temps are suggestive of lower ECS?
And in fact if you think that it is 3, then we have no disagreement and are in agreement with Knutti and Hegerl.
BTW you seem to be suggesting that Knutti and Hegerl have somehow made their model give the 20th Century results regardless of what ECS they use as an input. Am I right in that?
Carrick
“My take is, I expect GCM sensitivity drop again as they adjust the models to fit the newest data.
Now—nuance time: That doesn’t mean I think they are correct to adjust it down. I don’t really have a bloody idea. If I did, I’d be king of the world (so you are really lucky I don’t). I just think it’s plausible that they will adjust it down. I think it is equally plausible in AR6 they will shift it back upwards.”
hmmm you seem to see recent temp data as the most important for determining or adjusting ECS?
ECS is constrained by a large number of different lines of evidence, I don’t see how five more years of temp data will do much to better constrain ECS. Or are you suggesting they will learn more about the LGM or something?
Also, why do you think it is 3, if you think they will adjust it up and down over the next few years?
Nathan:
Nuance is important here.
What I am talking about is fitting models to data. And if your most recent interval has, for whatever reason, lower than expected temperatures, that will tend to push the estimate of the ECS lower than it would if it were to occur in the middle of the time series.
That’s generally accepted as true, even if people sometimes argue that in some really technical sense it isn’t. It’s not “curve fitting” in the sense that you are optimizing parameters using a goodness of fit measure, but in every other sense it still results in the same outcome: Namely models that fail to give realistic 20th century temperature series will be rejected as unrealistic. So modelers tend to “shop” for historical forcing series that match up with their particular numerical model to give the best (or at least “good”) reproduction of the measured temperature.
See this post by Lucia.
It’s got a nice bit on hindcasting in it.
I think that’s the best value given current state of knowledge. Add more knowledge and the value *should* change.
Given that the last decade ran “cool” I expect it to drop. I wouldn’t be surprise if we get another period of “warmer than expected” and that pulls it up again (but probably not to 4°C/doubling). I don’t have any idea what the true temperature sensitivity is going to be, though I’d be really surprised if it were lower than 2°C/doubling or higher than 4°C/doubling.
Mosher, I never claimed I was too busy with real science to respond (a false time contraint). I believe that was Bird, whom I’ve never exchange more than a few words.
I did say this was wasting my time (and I can’t believe I’m still commenting). I am not even at home in New York. I’m far outside of Chicago in Illinois attending a funeral and taking care of some personal family things.
You’re the one that said your compile job was done — as if that would impress me.
And now I am going out for dinner. Is that also a false time contraint?
And carrick talks about false statesments.
Cramer–
I’m not fixating on the numbers. The comment you claim is there is not there under any number. I can see a comment with a time stamp before it and one after. But the comment you think you posted (and evidently can see) is not there when I load the page. It is not in the source code. It does not display in my browser. Period.
I’ve been reading using firefox. But it’s the same with Safari:
Look.. I’m perfectly willing to believe that comment 73 displays differently for you. But this is how it displays for me. It would be nice if you would stop insisting I am simply refusing to answer whatever question you think you posted in ’73’ and refusing to repeat information simply because you think you are protecting yourself form some sort of “tribalisms” response. Either SA is moderating, or their system is F*ed up or something. But your links, and various things which evidently show for you do not show on for me.
Why don’t you try refreshing yoru cache? Maybe it will disappear?
Nathan–
I’m not sure what your issue is. Models–to some extent– are used for planning in various ways. The global tempeperature rise is considered one of the more robust measures. I think it is worthwhile to know whether they are accurate when used to predict the more robust measures. If they are not, it is difficult to use their guidance on planning at any level.
“Planning” is a very, very broad term. One can “plan” to change incentives to encourage people to switch from GHG’s to nuclear or wind. One can “plan” to build a breakfront. One can do all sorts of planning. One can “plan” on the level of costs one might expect as a result of climate changes. One can make all sorts of plans.
So– yes– GCM results are used in some sense for planning. Were they not intended for such uses, there would be no synopsis for “policy makers”. Policies are created to carry out plans of various sorts and the information in the section for policy makers is intended to help them “plan”.
As usually, changing the subject– in this case by trying to change my meaning by adding an adjective I never used. As it happens, in English the word “planning” doesn’t necessarily imply “local”. Since part of your job responsibilities is copy editing, I would think you would realize that when one omits an adjective like “local” that omission might not be accidental. In this case, when I did not say “local” that was because I did not mean to suggest that the planning would be in anyway “local”.
As Carrick suggested: You might do well to go back and quote. This would force you to reread. Failing that, it would help others explain how you came to misunderstand by pointing out that “planning” can be done at levels other than “local”.
Mosher
And yet…. they seem to have plenty of time to visit comments and post ….
Obviously, there are people who are too busy to argue in comments at blogs. These people don’t argue at blogs.
BTW:
I am a bit puzzled about what the heck is going on at comments at Scientific America. Can anyone else take a screenshot of comments over there to see if they can find “Cramer 7:13 pm 11/15/2012”. Presumably, it should appear near comment 73. But it doesn’t appear for me and I’d be interested in knowing if it appears for anyone else.
Mosher… Carrick… Any chance you can take a screen shot. Because this is very puzzling!
Lucia, I see exactly the same commenting sequence that you do.
I don’t believe Cramer would see anything different, unless perhaps he hasn’t refreshed his browser.
Ok– so the question he believes he repeatedly asked me but I supposedly didn’t answer doesn’t appear for you either. Based on my discussion with Bird it appears he couldn’t see it back when I said Bird wrote 73.
Maybe Cramer can refresh his browser and take a screen shot? Tjat would spare him the trouble of retyping the questions (which the rest of us still can’t see). We could then all learn what the questions were and maybe I could answer them. Meanwhile– I still don’t have any idea what was supposedly in that 7:15pm comment. ‘Cuz it doesn’t show!
Here are the screen shots:
Part 1
Part 2
Cramer 7:13 pm 11/15/2012 is clearly missing.
Maybe Cramer can refresh his browser and take a screen shot?
—
Or – maybe he could show off some of his Mad Stats & Math Skillz(c) and show how the original post is wrong?
TerryMN:
I’d settle for proof it wasn’t replicable. 😉
Lucia:
If it’s still on his browser he should take the screenshot before refreshing. Otherwise I guarantee it will disappear.
I’d settle for proof it wasn’t replicable. 🙂
I already asked about that twice too… Didn’t want to be re-redundant 🙂
Cramer (Comment #106612)
The time stamp you gave for comment 73 (now 72) was
Yes. I am saying “Cramer 7:13 pm 11/15/2012” doesn’t display. It doesn’t display as 72. It doesn’t display as 73. It doesn’t display.
I’m sorry if you feel you are being played, but Carrick also can’t see it.
In case there is any doubt, Carrick ≠Chewbacca.
(NB: That may only make sense to willard.)
Well… I would ordinarily think so. But evidently, he can read later comments. so it’s very puzzling that he can read comments after whatever it is he sees as “comment 72”– but he is still seeing as comment 72 something that does not display for either you or me. (And which did not seem to display for Bird. Bird theorized it was deleted two days back when I mentioned I could not see Cramer’s question.)
It is very puzzling.
This is really bugging me. After all, Cramer is seems certain his comment with a 7:13pm time stamp shows. Here’s from the google cache:
Nope. Not there.
I don’t think he is making this part up (though he’s made plenty up by the point), but it simply may not have been there long enough to get cached. I’m guessing like your blog the blog editors get email notification, and it appears his reputation is preceding him (a bit).
Cramer (Comment #106599)
November 17th, 2012 at 7:04 pm
If you want to know what I am trying to claim in that specific post– not much of anything. That post is illustrating the results of computations. So I guess my only claim is: If we compute the means and standard deviations of the trends and for models and that for the earth, the appear as plotted in that graph.
( Are you underwhelmed? I don’t know about you, but I think it’s just fine for a blog post to communicate material that is less comprehensive than one might anticipate in a ph.d. thesis. Just my opinion. )
I would think someone with more than two brain cell in their head could have figured out what the post was about by reading the intro and so on. It begins:
(Gosh… rereading that, one might think I was having a conversation with my readers. Perish the thought!)
What I was doing in that post was:
1) adding information to a graph I had created to visualize whether the observed trend for earth temperatures trend falls inside the spread of trends for ‘weather’ in models from the AR4.
2) In the previous post, I had not included the multi-model mean. I added that.
Now returning to the “Tamino” question (now that we have sorted out that when you referred to “Tamino” back at SA, I was supposed to know that you mean the paper I mentioned in “post C” when discussing “post B” which I linked in post “A”– and Sisko linked post A at SA– and in fact, I correctly guessed that you were referring to the paper by Foster and Rahmstorf)…
Even though I make no bold claims in “post A”, it is nevertheless entirely possible to note that my post A is on a quite different different subject from Foster and Rahmstorf (aka “Tamino’s paper).
The subject of Foster and Rahmstorf is– largely– whether or not global warming has paused or slowed. And so “the question” would be– stated largely “Has global warming paused or slowed?”.
The subject of my post A is whether the rate of warming falls inside the range consistent with the AR4 models. So the “questions” would be– stated largely “Is the earths temperature trend consistent with projections in the ensemble of models in the AR4”.
Carrick
Odd. They left Bird’s comment about oral sex up. I admit to being puzzled by their standards for moderating comments.
Too bad he didn’t take a screenshot so we could read the comment as it was written. Then we might learn what sorts of things Scientific America moderates.
Hmmm. Very interesting. Yeah, their blog definitely has some bugs.
It was fine in Sleipner, so I opened IE when you said you had problems and it also works fine there.
Doesn’t work for Firefox and Chrome.
Good luck with whatever you’re trying to determine. I have had enough of the behavior of your friends–can’t see what could be gained by spending anymore time here. It’s your blog and you are free to create whatever environment you thinks works for you.
Cramer:
Pull the other one. Or better post an actual screenshot. I don’t believe you.
Windows Internet Explorer, Windows 7, looks exactly the same as any other browser.
Sleipner, exactly the same.
I linked the wrong file for the Sleipner browser. Corrected link.
I went and looked at the HTML source and there is absolutely no way that different browsers would display it different. The entry that Cramer claims simply doesn’t exist.
Screen shot
Actual html source of this moment.
Looking at it in a bit more detail, there’s a missing comment, #comment-14809. I suspect this comment has either been blocked or just not approved (yet?).
I’m guessing that Cramer can see it because he is logging in as himself.
In other words, Scientific America is a worthless forum to try and have any meaningful exchanges in, if they don’t have consistent rules for what is appropriate or not (it’s ok apparently to talk about oral sex for example) and if they don’t clearly mark for you when your posts are either not approved or have been blocked from public view.
What dreck.
After reading this and over at SciAM, I find the lack of charity in conversation and lack of civility disappointing. Not to sound too churlish, but cramer stated “I don’t believe that the numbers being off is relevant to the solution. I provided the time stamps. I am not going to cut and paste ten comments. Then she will just say those comments did not exist. I have spent enough time on this. I feel I’m being played so I do not plan to spend much more time on this” after demanding all this accounting from Lucia that one can read here and at SciAm; but demanding a host of demands and then can’t bother to help clear up a problem, but instead makes a claim of being the victim, seems a bit overboard.
I wonder how many such as I found the accusations at SciAm at the middle school level rather than an adult conversation one would expect from a decent mag/org like SciAm? I was really disappointed.
John F. Pittman (Comment #106684)
If their blog actually displays the comment Cramer insists he sees to Cramer but does not display it to anyone else, then it’s no surprise that conversation becomes churlish. After all, he seems to honestly believe that I am simply claiming the reason I did not respond to his comment nor see the links he says he posted is that those comments do not display on my screen. Which they don’t. And I’ve provided ample evidence confirmed by Carrick.
If the SciAm forum really functions so badly (as it appears it may) that would make it possibly the single worst forum in which to try to have any sort of conversation at all.
Carrick
The number for comments in the html are likely the number in the database. So, the third theory is that the comment between Bird and mine is by someone entirely different and appears under another post. But that theory is not consistent with the missing comment displaying for Cramer.
For example: Your comment is #106682, John’s is #106684. Where is #106683? It’s Girma (Comment #106683) on this other thread: http://rankexploits.com/musings/2012/trends-relative-to-models/#comment-106683. Of course I can easily find these because I get emails for all the comments and I saw Girma commented on the other post. If one wanted to discover if the ‘missing’ comment at Scientific America might display you would either need to get a helpful person at SA to help you or write a ‘bot to crawl through all the pages looking for the appropriate snippet of text.
Lucia, I thought of course of the possibility that it was on a different thread, then remembered what you so well summarized:
😉
It really explains a lot. Notice that the comment numbers are only 14808 and 14810. For such a “prestigious” website those are amazingly low numbers.
Also the utterly bad nature of the forum explains why you only find people like cramer and bird on it. It’s actually self-selecting for behavior at “the middle school level”, as John F Pittmann rather charitably (I thought) put it.
I’m with Carrick (#106682) that the missing post was likely a non-approved one. Certainly there are other blogs at which not-yet-approved comments appear to the submitter. All of my experiences with such, clearly indicate that the remark is pending approval; perhaps SciAm doesn’t clearly indicate this, or perhaps Cramer just didn’t notice it.
But even if this is a fault in the SciAm forum software, I don’t think that would make it “possibly the single worst forum in which to try to have any sort of conversation at all.” The problem is participants who are more interested in point-scoring than in communicating clearly. E.g. Cramer, upon hearing that Lucia couldn’t identify his 7:13 SciAm comment, could easily have posted a screenshot, or cut-and-pasted the comment. That would have conveyed the information he was trying to convey with his reference, and moved the conversation forward. Instead, he proposed (and repeated) rather far-fetched claims that Lucia was deliberately being obtuse.
At SciAm — or more narrowly, on that thread at SciAM — the signal-to-noise ratio is low because people prefer to discuss each other’s motives, qualifications and suspected peccadilos rather than engage in substantive conversation.
By the way, it does appear that my last three comments suffered from the same “disappeared” phenomena that cramers comments suffered from.
Here is comment 88
Here is what I see for comment 89.
Do you see this? If you don’t, then it does prove that being logged into a particular account “enables” certain comments that otherwise aren’t visible.
What a rathole of a website.
HaroldW, I think that it is a rathole self-selects for the type of long-term denizens you find on it.
By the way, try posting at the bottom of the thread and let us know (Lucia or anybody). If we can see that comment, then this is proof that an editor had to have seen the comment I see as 89 and deselected it for public view.
Since I saw some of cramers comments before they disappeared, I’m leaning towards that explanation now.
There’s a “bit” of a discussion of deleted comments on this thread to.
It appears bird etc gets a pass. Wonder if he moderates? That’d be really frightening considering his oral sex comment that was cleared and allowed to stay in.
I can see throwing out rants. I can’t see throwing out technical explanations to other people.
I have a comment to Cramer’s last visible comment. Does anybody see this on the sciam blog. Here is the sciam link.
I think it’s just a low-traffic site because the articles are so utterly bad that only ClimateProgess types would find them intellectually stimulating and because of the bizarre asymmetrical nature of the comment “disappearing”.
Can’t resist. The sort of comment that makes it through their gauntlet is amazing:
This new learning amazes me…
Anyway, slightly back on topic to Cramer’s “disappeared” comment.
It would be hard to believe that his comment was such as gem as to be worthy of repeating that would later be followed this not-very-trenchant observation:
Because only Rob Starkey can reopen comments on this thread.
Everybody knows that.
Carrick,
The last comment I see by you is #87 time stamped 12:23 pm 11/16/2012.
The very last comment I can see on that thread is this one:
I’m wondering if Sci Am might be assigning administration of that blog to some undergraduate intern…maybe a History of Science or Poly Sci major.
Thanks John. That confirms what I thought.
I think the comments are getting “disappeared” (unless you are logged in on the account that posted them).
If you think Christine’s thread is bad, hanker over to the Hurricane Sandy one. Only grade school level banter is allowed here, sorry. 😮
Carrick–
For me, comment 89 is by Cramer (http://blogs.scientificamerican.com/observations/2012/11/14/climate-change-denier-likely-to-lead-congressional-science-committee/#comment-14846)
Oh– I should note: I did read the text of the comment in your screenshot at one time. I later saw it was gone as are some of mine. (I think there is some conversation above about some comments vanishing and others not.)
I guess the reason my comments fail to display for me is I never log into Sci Am. I’ve linked to Twitter. So I don’t seem to experience the feature you and Cramer are experiencing. If my comment is moderated, it doesn’t display for me either.
Interesting…. So, likely as not, Cramer is still seeing his own comment. But it doesn’t display for the rest of us. Due to SciAm’s “system”.
Carrick,
The last comment I see from you on that thread is “#86. Carrick
11:33 am 11/16/2012”, which begins
After that is
87. Cramer 12:23 pm 11/16/2012
88. Bird/tree/dinosaur/etc. geek 12:48 pm 11/16/2012
89. Cramer 2:19 pm 11/16/2012
90. jkeenan913 11:03 am 11/18/2012
It would be remarkable to be able to follow a conversation when the identifying post #s change over time, not to mention the apparently frequent deletion of posts.
It is also not credible that those who post there regularly would not be aware of the idiosyncrasies of the blog.
And I agree with your comment about self-selection in #106690.
It’s pretty amazing that for an extremely low volume site like this, with the resources they have, that they can’t provide any feedback about comments that get moderated.
If you’re going to moderate “in thread”, I still like Steve McIntyre’s approach about the best. As far as I know he does these all himself, is unpaid (except for that large fossil fuel check of course) and of course gets as many comments a year as SciAm has gotten at least since this new blogging system was enabled.
See this exchange as an example:
Image.
Note this is comment 373182, and his comment numbers got reset when he switched to this new blogging system.
I agree Lucia. The posting at SciAm makes him a victim, while making it look like he is churlish and complaining , when correctly from what he sees, he actually is trying to communicate. This does have to go down as one of the worst moderations I have ever seen. And the lack of moderating the suggestive while deleting valid comments is just bizzare. I guess we can name SciAm – the Bizzarro Blog.
Well… it is pretty obvious that Scientific America doesn’t generate much conversation. After all, they have about 3 articles a day and the time between comment numbers in the data base is on the order of a 1/2 hour. If one were disposed to assume that blogs associated with long standing periodicals were “the place” to discuss things, the low rate of comments might be surprising. But… the fact is… often those magazines run blogs as advertizing form more magazine sales. It appears the level of articles at Sci American’s blog are particularly low– and likely lightly read.
HaroldW (Comment #106699)
Oddly, they might not be aware of it. They see their own comments even if moderated while those of other people’s vanish. They may develop the impression that they rank among the “valued” commenters who are never moderated while those moderated are the “troll”. And since for many, conversation becomes fragmented and confused, and they just go away. After all, why stay when your comments are deleted, comment numbers quoted by others are inconsistent with what you see etc.?
Meanwhile, frequently moderate-ese (i.e. those who were moderated like Cramer) might never realize they are often moderated. So… they merely develop the impression that the reason the comment numbers quoted by others don’t align with the display on the moderate-ese screen is others are making mistakes. In fact, it would seem that if the numbers quoted by other visitors don’t match the ones a blogger visitors doesn’t see it’s quite likely that…. well… some of that visitors comments are only visible to him! That is: his comments were deleted and … well.. it just may be he is deemed a troll by the moderator!
Anyway, based on discussion on this thread, it appears that Cramer may be frequently moderated at Sci Am. If he is hoping people are benefiting by his words posted there— he is mistaken. He is likely wasting his time posting because no one other than he can read many of his comments!
And he thinks I’m the one mistreating him. The irony. It burns.
John F Pittman:
Suffering however from self-inflicted wounds. He could have just posted a screenshot.