Guest Post by Bob Tisdale
Not too far in the future, a third-generation climate scientist will be asked by a college student about his grandfather’s work with climate models. Flabbergasted, the climate scientist replies…
My apologies to Mel Brooks and Gene Wilder, but I couldn’t resist.
The models used by the IPCC really are horrible representations of climate on this planet. One example: the modelers have to double the warming rate of global ocean surfaces to have the modeled global land surface temperatures come close to reality. That’s a horrible display of modeling. And there are many more. The following graphs were presented as Figure 9 in the post here.
PS: Just in case you’ve never seen Mel Brook’s Young Frankenstein, here’s a very short YouTube clip from the scene:
But wait! It’s worse than we thought!
The models are not just horrible representations, but intentional misrepresentations, IMO.
Sounds like a bad late night infomercial:
“But wait, that’s not all!!! Call now and we’ll DOUBLE the horrible representations, AND through in a free intentional misrepresentation AT NO EXTRA CHARGE!!! Call in the next five minutes and we’ll also give you a lecture from your moral superiors that is PRICELESS!!!”.
Josh?
Great post, Bob. Anything that can do a valid tie-in to Young Frankenstein has great merit.
On a serious note, check out Jennifer Marohasy’s work on exactly this topic. She and her co-authors are working hard on an alternative paradigm for cliamte science.
http://jennifermarohasy.com/
And they are getting some good results.
“So, what I will do in the limited time available today is tell you something about how artificial neural networks (ANNs), a form of Artificial Intelligence (AI), have potential as a new tool for a new paradigm, and in particular their application to rainfall forecasting. I will then be in a position to better answer the question for this session concerning rainfall and water availability in a warming world.”
…
“While AGW is a demonstrably failed paradigm, it will be replaced only when a critical number of practicing scientists start working on something new. New paradigms always have their own questions and their own tools. In the same way that GCMs underpin AGW theory, ANNs could underpin a new paradigm based on better quantifying the drives of natural climate cycles.”
http://jennifermarohasy.com/2014/07/the-need-for-a-new-paradigm-including-for-rainfall-forecasting/
Artificial neural networks (ANNs) are massive black-box statistical models that are very easy to over fit to the data, meaning they can match historical data well but do a horrible job of prediction. If you think today’s climate models are bad, don’t look to these as an improvement.
Yes and no, Barry.
Yes and no NOW, but at our current rate of advancement in AI soon our computers will be able to answer ‘maybe’
And from there it’s only a short step to ‘I don’t know’ and ‘could you repeat the question’
It seems she has found out that some parts of Australia will cool down over 17 degrees celsius / century, extrapolating from 10 year data of maximum temperatures apparently.
Isn’t that a bit alarmist, extreme or down right scientifically unsound projection?
Wow. That’s almost like having a massive volcanic eruption at Ayers Rock! It’s impossible without something profoundly catastrophic, considering there’s only been some 0.8C warming since the end of The Little Ice Age, 17C / century cooling is like changing the Moon’s in orbit so Australia never sees sunlight again.
Bob, You’ve got to remember that these are just simple climatologists. These are activists of the land. The common clay of the new Green West. You know…
All very fine and well, but I would prefer that they keep their fertilizer to themselves.
Thanks for the paraphrase, Robert. I’ll have to try to work a “Blazing Saddles” moment into a future post.
I believe the next line to what you presented would be, “You know…morons.”
I look forward to a post working in the campfire, bean-eating scene.
@hunter Great post, Bob. Anything that can do a valid tie-in to Young Frankenstein has great merit.
I prefer the Underpants Gnomes. Almost all Climate Change operates by following their business model:
1 – note some change in the natural world
2 – ?
3 – We’re all going to DIEEEE!!!!
I’m quite sure that #3 is actually still PROFIT!
Yup..
10 A&=CLIMATECHANGE(NATURAL)
20 A&=RND(A&+CLIMATE)
30 IF A&=0 THEN
40 panic$=”Wot me worry? Zero never happens.”
50 hassleguvforfunds&=a&+billions
60 GOTO 10
60 ELSE panic$=”We’re gonna DIEEEE!!!”
70 profit&=hassleguvforfunds&+billions
80 GOTO 10
I might recode this in assembler so it executes MUCH faster..
It is recommended therapy to reset ones funnybone after mankind’s sins have been perpetrated on the rest of humanity.
Young “Frunkansteen” is my cure for all the evil the bad guys do and really “puts the candle back”
Models are indispensable in the engineering professions and other areas of applied science where there are criteria that have to be met. In these professions, models are not allowed to fail because of the imperatives.
It is different in climate science. Every single GCM gives an egregious product, yet the climate scientists refuse to acknowledge their failure. There are no standards nor criteria that have to be met in this field except the need to support ideological motives.
Comparing climate models and predictive models/tools used in other disciplines is somewhat apples and oranges but instructive nevertheless.
In my profession (satellite design, science space instrument design, space environment and effects) models are a routine part of getting it right. In the models we use, the physics is relatively well established.
For instance, a thermal model for space application of a design uses well-established physics principles of radiative and conductive heat transfer and known worst-case boundary conditions with runs over worst-case bounding environments for the entire life cycle of the design. On top of that relative certainty of the physics, we add design margin and ensure that the thermal model contains enough nodes to accurately represent the physical design. Then we test to the worst case, and we correlate the model to the actual observed response to these conditions. Then the model is good enough and trustworthy enough to be used for predictions of that specific design.
The same sort of process occurs for predictions using a finite-element structural model, dynamic system response using attitude control models, space environment radiation effects models, reliability models, data bus timing and traffic models, etc. Margin thrown in of 25%, 50% or even 100% allows our designs to work. And, we know the error bars of the model results. When the model correlation is not working, it’s back to the drawing board.
As for climate models, and I’m certainly not an expert there, it seems to me that much of the basic physics is missing, the Earth atmosphere model is much too coarse (i.e. misses processes like your basic thunderstorm and the associated energy transfer), or the physics is just plain wrong (e.g. water vapor feedback sign and magnitude). I don’t think the physics in climate models is in any way settled. That’s a huge difference with models used in other disciplines and making policy using Global Circulation Model output is a huge mistake, IMHO.
Bruce
In the design of a scientific study, the worst mistake that can be made is to omit identification of the events underlying the models. One of the consequences is for the system that was supposed to be brought under control by the study to remain uncontrollable. This mistake was made in the design of the study of the gtobal warming phenomenon.
I have no problem with this article but I remain a little bothered that climate crimes against humanity by the AGW crowd can so easily be turned to potty humor. I’d like to think better of these AGW pranksters and their work but I can’t – I have no grounds for doing so. They’ve brought this on themselves. Punk away, Bob.
Insert Abby Normal joke here.
What do the models show for projected deep ocean warming?
“It must be there! We can’t find it anywhere else!”
I wonder how this graphic
would change if you updated it to include the most recent 12 months? These most recent 12 months are the warmest in history (according to the dataset used on your graphic on the right-hand side graph.
http://www.washingtonpost.com/blogs/capital-weather-gang/wp/2014/10/20/after-record-warm-september-2014-is-on-track-to-warmest-year-noaa-says/
While we still have three months to see whether the calendar year will set a record, the past 12 months — October 2013 through September 2014 — was the warmest 12-month period on record, at 1.24 degrees above the 20th century average temperature.
Even so, the observed warming of .28C per decade at PREVIOUS warming rates shows that future warming projections will be much worse with a likely land-area warming of over 4C by 2100 under current warming regimes. Since the previous ice age was 4.5C colder than today, this warming will produce massive and catastrophic effects.
perhaps this is why the department of defense has determined that climate change is a current national security threat.
Is that why melting glaciers are revealing tree stumps from past forest?
Perhaps Mann ran a tree-ring analysis and determined that the trees that produced those stumps were ice sculptures?
The Medieval Warm Period didn’t exist? The Minoan? The Roman?
Perhaps you left out a word or two needed for clarification. Please, clarify.
PS Nice to see your name again.
Gunga Din – me too. I needed a laugh today…. 🙂
I recall several years ago hearing someone from the U.N. state, “Last summer [probably 1998] was the hottest in 5000 years.” Although I was aware of the Medieval, Roman and Minoan optimums, I accepted his remark at face value – and then wondered, “Gee, that’s interesting. Wonder what caused it back then?”
You are really not getting the models are total doo doo in both methodology and integrity of data input. Anyone can pick any particular dataset from any particular time period to make any particular claim they want to make or make-up as the case may be. Compare to the methodology and data used to make bridges, airplanes, and ships and the failings become obvious.
Due to “model blindness” (a new psychiatric term to describe an unhealthy obsession with models) you go on and forecast the temperature of the entire planet 90 years from now with the same certainty an evangelist has in knowing their spot in Heaven. With the advanced stage of model blindness hitting, you wildly predict the temperature difference is going to be catastrophic. Apparently you have modelled the entire planet ecosystem to the last amoeba, molecule of water vapor, rain forest, animal, insect and plant species.
Since you are so adept at looking 100 years into the future perhaps you could let us know if we will have mastered interplanetary travel by 2100. How about solar powered bicycles?
” These most recent 12 months are the warmest in history (according to the dataset used on your graphic on the right-hand side graph.”
Well they will have fudged it up the tiniest amount over the previous record 12 month average… They don’t dare to fabricate a Marcotte-Joe Romm-style fantasy hockey stick… so… indistinguishable from “hiatus”… but enough for the system journalists to scream bloody murder (none of the headlines say HOW MUCH warmer.)
Hi jai mitchell. Welcome back. Where have you been? I haven’t heard from you for a while.
You wrote: “I wonder how this graphic…would change if you updated it to include the most recent 12 months?”
How’s about I add 13 months of data and models to really make the graphs up-to-date? I wouldn’t want anyone to think I was hiding something. Just happened to have them lying around.
Here’s the land-surface-air temperature-only (oceans masked) model-data comparison:
And here’s the sea surface temperature model-data comparison:
As you can see, there haven’t been any drastic changes in the warming rates even with the temporarily high sea surface temperatures in recent months. The models still perform horribly.
BTW, just in case you want to complain about the start date, thinking I’ve somehow cherry-picked it, the Reynolds OI.v2 sea surface temperature data start in November, 1981.
Cheers.
“Even so, the observed warming of .28C per decade at PREVIOUS warming rates shows that future warming projections will be much worse with a likely land-area warming of over 4C by 2100 under current warming regimes. Since the previous ice age was 4.5C colder than today, this warming will produce massive and catastrophic effects.”
Well, it is already 2015… nearly. Let’s assume (for giggles) that the 0.28C warming rate is correct (but wait, it ISN’T, because the UAH and RSS satellite data do not show anywhere NEAR that rate of warming). But anyway, again, for giggles, lets assume that it will continue to warm at an equivalent rate of 0.28 degrees per decade for the remainder of this century. That would mean the total amount of warming by the year 2100 would likely be 8.5 * 0.28 = 2.38 degrees.
However, it is WORSE THAN YOU THOUGHT, because the rate of warming during THIS CENTURY ONLY (i.e. from 2000 to present) is only about 0.07 degrees/decade. So, if it continued warming at THE CURRENT RATE we would only see a warming by 2100 of 8.5 * 0.07 = 0.6 degrees. I find this scenario to be much more likely, given that the rate of warming over the past ~15 years is almost non-existent.
In order to achieve 4 degrees of warming by 2100, WARMING WOULD HAVE TO BE ACCELERATING, which, clearly, it is not. In fact, warming has essentially stopped dead in it’s tracks since the big 1997/8 El Nino.
Absolutely. Having a global temp range of -77.62C to 42.38C instead of -80C to 40C is truly “Catastrophic.”
Does the chart on the right really look that bad?
Only because the “observed” land station “data” have been so shamelessly “adjusted”.
The GISS (satellite) data shown in blue is one of the sets not “adjusted” and not subject to UHI effects, maintenance issues, and other siting problems. The “hottest year on record” refers to dubious surface data.
You are wrong.
GISS is not satellite data.
UAH and RSS are.
GISS land temperature is homogenized, extrapolated to places 1200 miles from the next thermometer, and GISS constantly cools the past and warms the present. The most corrupt of them all.
What we see is that the most heavily corrupted land temperature series – GISS – corresponds to the warmist models; while the oceans on the left side do not.
You know some people are confused because GISS is NASA and GISS says Goddard institute for SPACE studies; but NASA knows exactly what it’s doing; they earn 1.2 bn USD per year for creating Global Warming propaganda.
I stand corrected by DirkH. My comment about satellite date vs surface data still stands, but the graph in question does not reflect RSS or UAH.
Anyone who hasn’t seen “Young Frankenstein” needs to retain a “life coach” immediately, and start catching up ASAP.
I have a life couch. Does that count? I slept on in when I was married.
You are not married now? If that is indeed the situation, you are some kinda catch.
Certainly has been a long time already. I was introduced to the anti-human alarmist campaign at a Jefferson School conference circa 1990, where the movie Four Pillars or such was shown. Of course the campaign started much earlier.
Oh, I wonder if Tim Ball has grandchildren. He was researching climate before Andrew Weaver was conceived.
” These most recent 12 months are the warmest in history (according to the dataset used on your graphic on the right-hand side graph.”
I wonder how they reconcile this when the USCRN network is showing cooling? There is almost a deg and a half of disparity. (1.47 Deg C)
Just food for thought….
They’re a result of the high sea surface temperature anomalies, caused by the blob in the extratropical North Pacific:
http://bobtisdale.wordpress.com/2014/08/16/on-the-recent-record-high-global-sea-surface-temperatures-the-wheres-and-whys/
Data. Well that’s the problem, isn’t it?
Here we are, hundreds of years after the Scientific Enlightenment, and we can’t even agree on what the “actual” data is – not even for the last 30 years (aka: the satellite era).
As an old CFO (and I mean old), I know painfully well that if you can’t get agreement on basic data, it’s a waste of time discussing implications of the data.
Its worse than you thought – they take the actual raw data and make all kinds of “adjustments” to it, none of which are really justifiable. They would be better off simply using the raw data, but only from stations which actually met good siting criteria (a lot of sites do NOT).
Bob Tisdale: “Thanks for the paraphrase, Robert. I’ll have to try to work a “Blazing Saddles” moment into a future post.”
I think I have just the scene… Sitting around the campfire after ingesting a plateful of beans…
The ramifications are horrific, considering that methane is a much more powerful green house gas than CO2.
Sorry, I’m usually more erudite in my comments.
Is it not the case that observations remain inside the model range, albeit at the low end? http://www.met.reading.ac.uk/~ed/bloguploads/FIG_11-25_UPDATE.png
On that scale, global temperatures (black line) look likely to move back inside the 5-95% range once 2014 is added (currently on target for more or less in the middle of the vertical green ‘met office 2014 forecast’ line).
If so, then are the models so bad, really?
How do they compare, for instance, to the forecasts made by David Archibald and Don Easterbrook over the same period?
DavidR,
Statistically, “Model Range” is totally meaningless. If you take 35 models, all of which don’t accurately reproduce the climate, run them all 100 times, AVERAGE THE 100 MODEL RUNS (which is a statistically invalid operation to begin with), and then PLOT THE SPREAD of the 35 models all averaged out over 100 runs for each of the 35 models….
All you really have is 3500 pieces of spaghetti (“homogenized” into 35 “lines”), none of which actually mean anything!
There was a recent article here on WUWT by Professor Robert G. Brown (Duke University) that does a FAR better job of explaining this than I just did, but I think I got the gist of it. He frequently comments here as Rgbatduke, so perhaps he can add to what I just said or provide the link back to his article. It was fantastic, but I didn’t have the patience to find the link for it 🙂
I agree that the MEAN of the model ensemble is meaningless.
But the RANGE has a great deal of meaning.
Either the reality is falling outside the range, which translates to:
…”We don’t have a clue!”
Or the range incorporates the reality only by being so wide that its statistical significance translates to:
…”We don’t have a clue!”
So you see, the RANGE not only has meaning, it is robust; reaching the same conclusion under more than one set of assumptions.
The ‘range’ encloses a finite number of projected outcomes. If observations are contained within that range of outcomes, then the model range isn’t invalid. It could be invalidated in two ways, from what I can see:
1. Observations fall outside the model range.
2. Observations remain outside the 5-95% range (whether at the low or high end) for more than 10% of the projected period.
As far as I can tell, neither of those two things have yet happened, and if 2014 to date is anything to go by, they seem unlikely to happen in the immediate future.
DavidR says: “If so, then are the models so bad, really?”
It depends on what metric we’re looking at. You can’t overlook this.
Bob,
With respect, you’re comparing the ‘average’ of the models with observations; not the ‘range’. What your chart shows is that ‘on average’ the models predicted faster warming than has been observed. I’m not suggesting that isn’t the case.
What I’m saying is that it’s a fact that observations remain inside the model range and, indeed, look likely to move back inside the 5-95% range once 2014 is added. This means that some models must have projected a rate of warming that was even slower than what has been observed to date.
DavidR:
It’s best to avoid saying that the models “predict” by saying that they “project” ( reason at http://wmbriggs.com/blog/?p=7923 ).
When it comes to defending climate models, another Mel Brook’s Blazing Saddles quote is apropos:
“We’ve got to protect our phony-baloney jobs, Gentlemen!”
Concerning Jim Hansen and GISS adjusting and homogenizing recorded temperatures I think of a scene from another of my most favorite movies.
…. YOU have meddled with the primal FORCES of NATURE!
And YOU… WILL…. ATONE !!! ….
Am I getting through to you Mr. Hansen?
(Howard Jensen, Network, 1976)
I always thought that Ned Beatty should have been nominated for an Oscar for that brief scene.
You aren’t the only one.
He was nominated for Best Supporting Actor.
Ned Beatty says that he was brought in as a replacement in mid shoot. The first actor wasn’t hitting the right notes with the Jensen “Forces of Nature” speech. He did one day of work. One day of rehearsal and shooting at the Board Room of the New York Public Library. He gets nominated the Oscar for Best Supporting Actor.
The lesson he learned:
“Never turn down a day’s work.'”
Competition was tough that year: Burgess Meredith and Burt Young, (both Rocky). Lawrence Olivier (Marathon Man), and Jason Robards (won, All the Presidents’ Men).
I love the Jensen speech and Beatty’s delivery, but I would have voted for Burgess Meredith as Rocky’s boxing coach.
White Elephant: The range of temperatures on this planet of our go from -80C to +40C. The charts here go from -0.2C to +-0.6C.
Anomalous Analysis isn’t a bad thing to do, but keep the chart of all data points on your wall when you do it. When you are studying the ant crawling down the bark on the twig of a branch of a tree in the national park connected to a national forest which is part of the state, it is a really good idea to step back every once in awhile and assess what the whole is doing.
Averages are your friend. Averages will kill you. Average are dead. Long live averages.
There are those that will laugh at what I write because they get it. There are those that will think me insane.
Both are correct. Both are wrong. The dance will go on though no matter what I do. The rain will fall and smack me with its enthalpic reality. The wind will blow from the differential change. Some will see the rain as a blessing of life. Others will despair at the gloom of the clouds hiding the sun.
Any chart that has units of C with regards to an analysis of climate change is guaranteed to be an average gone wrong. Rains enthalpy is lost. When these charts start to be in units of enthalpy, the scientists will have made the first step towards starting to understand. The enthalpies will still be wrong. That is the nature of the data we use. They are an approximation of the state of things. Enthalpies give a better place to start evaluating the state. Once you convert to enthalpy, you can’t convert back, not with the unknowns involved in the climate discussion.
Charts in Anomalous Temperatures are the output of erroneous energy balance science.
I don’t know that much, but I know that any scientist who doesn’t get rosy faced discussing anomalous temperatures doesn’t have a clue about the uncertainty of the data he is discussing. The error bars on the charts above are bigger than the the entire range of the data displayed.
Rubbish. There is nothing wrong with the climate models. The earth has just been bought out by the fossil fuel propaganda machine to conceal the warming.
Sorry I should have added Surely the question is not whether it is warming but whether it is warming in each decade at a rate consistent with the CO2 levels at the time. Aren’t they supposed to be scientifically trained to understand this when a mere engineer can do so?
I agree that decadal measurements are a good way to look at it. If the current decade (2011-20) doesn’t turn out to be warmer than the last one (2001-10) then I’d agree that the impact of CO2 has been overestimated.