Note: Short term predictions are relatively easy, it remains to be seen if this holds up over the long term. I have my doubts. – Anthony
Guest post by Frank Lemke
The Global Warming Prediction Project is an impartial, transparent, and independent project where no public, private or corporate funding is involved. It is about original concepts and results of inductive self-organizing modeling and prediction of global warming and related problems.
In September 2011, we presented a medium-term (79 months) quantitative prediction of monthly global mean temperatures based on an interdependent system model of the atmosphere developed by KnowledgeMiner, which was also discussed at Climate Etc. in October 2011. This model describes a non-linear dynamic system of the atmosphere consisting of 5 major climate drivers: Ozone concentration, aerosols, radiative cloud fraction, and global mean temperature as endogenous variables and sun activity (sunspot numbers) as exogenous variable of the system. This system model was obtained from monthly observation data of the past 33 years (6 variables in total: the 5 variables the system is actually composed of (see above) plus CO2, which, however, has not been identified as relevant system variable), exclusively, by unique self-organizing knowledge extraction technologies.
Now, more than a year has passed, and we can verify what has been predicted relative to the temperatures, which have really been measured (fig. 1).
Verifying the prediction skill of the system model from April 2011 to December 2012, the accuracy of the most likely forecast (solid red line) remains at a high level of 75%, and the accuracy relative to prediction uncertainty (pink area) is an exceptional 98%. Given the noise in the data (presumably incomplete set of system variables considered, noise added during measurement and preprocessing of raw observation data, or random events, for example), this clearly confirms the validity of the system model and its forecast.
In comparison, the IPCC AR4 A1B projection currently shows a prediction accuracy of 23% (September 2007 – December 2012, 64 months) and just 7% accuracy for the same forecast horizon as applied for the system model (April 2011 – December 2012, 21 months).
The two models, IPCC model and atmospheric system model, use two very different modeling approaches: theory-driven vs data-driven modeling. The IPCC model is based essentially on AGW theory by emission of greenhouse gases, namely CO2, the presented atmospheric system model on the other hand is a CO2-free prediction model. It is described by 5 other variables. The IPCC model shows a prediction accuracy of 7% and the atmospheric system model an accuracy of 75% for the same most recent 21 months of time…
The climate system is a complex system that consists of a number of variables, which are connected interdependently, nonlinearly and dynamically and where it is not clear, which are the causes and which are the effects. The simplistic linear cause-effect relationship “more atmospheric CO2 = higher temperatures” the IPCC model is based on is not an adequate tool to describe the complexity of the atmosphere sufficiently.
Read the complete post here:
http://climateprediction.eu/cc/Main/Entries/2013/1/21_What_Drives_Global_Warming_-_Update.html
No-one is or can be “expert on the future.” Any complex dynamic system –one with three or more mutually interacting variables– is necessarily chaotic/fractal, non-random but indeterminate, self-similar on every scale.
Regardless of modellers’ sophistication, any and all non-linear extrapolations are exponentially qualitative, not quantitative, meaning that “false precision” inescapably corrupts any extended time-series at its very root.
On this basis, conventional “hindcasts” too are terminally misleading. Realistically testing any multivariate scenario requires not “working backwards” but setting a remote start-date to begin a model-run from scratch, without interim “adjustments” or any skewed causative factors whatsoever. In principle, this simply cannot be done; and anyone claiming otherwise is if not a knave or charlatan, then a benighted fool.
Try running a global Hedge Fund for 3 – 6 months, and see how far conventional wisdom carries. Recall the celebrated case of Long-term Capital, rife with Nobelistas, whose verities evaporated overnight: Where illusions define imaginary problems, contingencies govern and exceptions rule.
Mark Bofill says:
January 24, 2013 at 7:05 pm
========================================================================
What risk? The claim has been made. The last 16 years have shown the claim is without merit. (Substitute ‘hypothesis’ for ‘claim’ if you wish.) What “brainer”? Why risk more?
You’re free to risk your own well being or pay out of your own pocket but not mine. Would you use Government to force me to do so? If so, we’re back to my original comment.
Forget about it. Mosher only trolls here these days. One post per thread…then poof gone, while everybody else flails.
Gunga Din says:
Your comment is awaiting moderation.
January 24, 2013 at 9:33 pm
===========================================================
Dang it!
I meant to copy this.
Gunga Din says:
January 24, 2013 at 4:19 pm
Steven Mosher says:
January 24, 2013 at 3:36 pm
(Me:)Think of all those drug commercials in the US. Lots of disclaimers. (Who knew an eye drop might kill you?8-) Sure, they have those disclaimers to legally CYA but they need to CYA because it could happen. Before a drug hits the market it’s been thoroughly tested. But they don’t know everything. The human body is incredibly complex.
“Climate” is incredbly complex.
###############
(Mosh:)yup. that is why it is weird that people think it is safe to geo engineer the planet by dumping C02 in the air.
===========================================================
(Me again:)Or Olvatine in the ocean? 😎
Either way, we don’t know enough. If it wasn’t for the politics involved and the cost in $$ and freedoms that the policitians are after, let the honest scientist continue to … well … observe and apply the scientific method to any hypothsis that may arise from those observations. The Hansens and the Manns haven’t put out any ‘disclaimers’ that said they might be wrong or the consequences if they are. But the Gores and the Obamas are cashing in.
“People think it is safe to geo engineer the planet by dumping C02 in the air.”
People are just living. The only people I’ve heard of that are trying to “geo engineer” anything are those that have picked CO2 out of the chaos that makes “climate” and are trying to limit or remove it because they feel that what Man adds is somehow not “natural”.
Steve Oregon says:
January 24, 2013 at 1:54 pm
This climate forecast model is a real travesty.
Plus it got me thinking.
How will alarmists cope if warming never returns for the rest of their lives?
A few more years of the same will be bad for them. 6, 7 or 8 will be painful.
But 10, 20 or 30 years of a non-warming planet will be catastrophic for their funny little fictitious world.
Never be surprised how people can rationalise their beliefs. Eleven of the last 12 years have been the hottest in 34 years. We hear all sorts of excuses, including it is cooling, the temperature has plateaued, the measurements are wrong, the people measuring it are corrupt etc. What they won’t admit is that it is getting hotter. They ignore or make up various stories to explain the reduction in the Arctic ice cap, the rising sea, the retreat of the glaciers, the number of broken heat records broken being nearly three times the number of cold records broken.
So I imagine they will cry you a river of tears when it doesn’t get colder. They will be crying because they missed the opportunity to do something about it when they could.
Pamela Gray says:
January 24, 2013 at 1:33 pm
“The modeler that gets it right will use ENSO patterns of oceanic circulation and SST (a much slower lagged effect)”
==========================================
I respectfully disagree with you and Tisdale on this. This years aborted nino is the classic example. There is a bigger oscillation that transcends ENSO. To some extent the thermohaline circulation is like the loose end of a fire hose in the Pacific. It is quite likely that the northern Pacific has been a deepwater deadend since the Cretaceous. Its current deadend status is evidenced by the antiquity of the deepwater off California.
Bob Tisdale’s work detrending clearly shows that the 1997/8 El Nino was not the whopper everyone thinks it is. It was actually less severe than the previous big one in the eighties. It was a hinge that swung the loose end of the firehose to the eastern Pacific.
OK. I didn’t make my point very well.
You’ve seen those commercial put out by a law firm about some disease or side effect? (‘If you or your loved one has ever suffered from …”) Do you know what they’re doing? They are shopping for clients. They have MDs on their payroll that no longer practice medicine but make their living testifying in court. So it is with CAGW “climate scientist”. They are now being used by those who’s agenda (money, political power. ideology etc.) stands to profit by their testimony. That has all to often corrupted their testimony.
There is no genuine risk from providing affordable energy. CO2 is no more a pollutant to be limited and removed than H2O is a pollutant to be limited and removed.
Steven Mosher says:
January 24, 2013 at 3:36 pm
” yup. that is why it is weird that people think it is safe to geo engineer the planet by dumping C02 in the air.”
Ah, precautionary principle Mosher again. The oceans outgas CO2, Mosher; how will you stop them from doing so?
About any human activity besides sitting there is geoengineering. How will you stop humans from building houses, Mosher?
Steven Mosher says:
January 24, 2013 at 3:36 pm
yup. that is why it is weird that people think it is safe to geo engineer the planet by dumping C02 in the air.
Oh, turn it off already, you’re boring. We dump F’-all CO2 in the air in comparison to what nature “dumps” and historically has dumped in the air. There is no evidence at all that the CO2 from coal is any more harmful than the CO2 from termite farts. Plants eat it up just the same. CO2 levels have been way higher in the past than now, and it wasn’t Armageddon. If the Alarmists were to realise their goal of reverting back to pre-industrial CO2 outputs, the West will join the third world in famine, misery and death. SO what’s YOUR solution to your perceived problem? Besides more bed wetting, I mean?
Moe says:
January 24, 2013 at 9:48 pm
“Never be surprised how people can rationalise their beliefs. Eleven of the last 12 years have been the hottest in 34 years. We hear all sorts of excuses, including it is cooling, the temperature has plateaued, the measurements are wrong, the people measuring it are corrupt etc. What they won’t admit is that it is getting hotter. ”
I’m not sure, but I tend to think you’re NOT being sarcastic.
Moe, yes, it is getting hotter, but only in the GCM’s. In real world Germany I got an effin -5 deg C and ice and snow outside, this is not right! In 2000 I was able to walk around in a T Shirt. DO SOMETHING, MOE! CO2AGW, why have you forsaken us?
So over the period of prediction, which is too short to produce a reliable conclusion, the actual temperatures vary much more than the predicted temperatures.
So it’s not looking good so far.
In fact the data mining model used does not seem to be much more useful than a linear extrapolation of the previous few years temperature trend.
“radiative cloud fraction” as an input makes no sense. For all anyone knows, this may closely track CO2 concentrations, and so in reality CO2 is an input into this model.
You have to know what drives changes in “radiative cloud fraction” for this to be of any use. What’s the ten-year forcast for this parameter?
A few remarks from the comments and the post.
1. This is true out-of-sample (ex-ante) forecast. The system model was developed in May 2011 algorithmically by software on data of 6 variables (ozone, aerosols, reflectivity, SSN, CO2, HADCRUT3) from Nov 1978 – Apr 2011 (378 months). The time constraint is simply given by the availability of satellite data (TOMS) used for the first three variables. This model (actually it is a set of nonlinear dynamic models that are interconnected) has been predicted then until Nov 2017 (79 months). This prediction has been published in my original post in Sep 2011 and on Climate Etc. in Oct 2011. Neither this initial prediction nor the models have been modified since then. So the model has proven that it worked for 21 months so far.
You find further information about the system model in my original post:
http://www.climateprediction.eu/cc/Main/Entries/2011/9/13_What_Drives_Global_Warming.html
There are also forecasts for ozone, aerosols, SSN etc.
2. I consider it a medium-term prediction, because it predicts 79 months ahead relative to the 378 months of observation data. But this is just 6 years and in the context of climate this is short-term, of course. This is what this model can do: forecasting few years ahead. Longer forecast horizons are not possible given the short observation time of 32 years. It is a decadal-scale forecasting model, which also the UK Met Office is doing research on.
3. The system model was developed by inductive self-organizing modeling, which is knowledge extraction from data. It shares some concepts of Neural Networks but it goes beyond them. It is not a Neural Network. Self-organizing modeling means that the model is completely undefined at start, nothing else is given than observational data (and a learning algorithm). During self-organization an optimal complex model is evolved, autonomously, i.e., the composition of the model structure, the parameters, the relevant inputs, the individual transfer functions. Self-organization performs STRUCTURE IDENTIFICATION and parameter estimation, Neural Networks and traditional statistics only perform parameter estimation.
More on this here:
http://www.knowledgeminer.eu/about.html#som
4. Why SSN? SSN, TSI, or the sun‘s magnetic field data are all highly correlated. They represent sun activity. Also remind that we use more or less noisy data anyway. Practically, it makes no difference which of these variables is used. I also tried TSI and the results are very similar.
Sunspots forecast here:
http://www.climateprediction.eu/cc/Main/Entries/2012/3/22_Sunspot_number_prediction_(3).html
5. Additionally, our climate system is essentially influenced by external, cosmic climate drivers such as the Earth Orbit Oscillation in centennial time frames, the multidecadal tri-synodic Jupiter/Saturn cycle, or the well-known orbit eccentricity Milankovitch cycle, which causes glacial and interglacial ages on Earth. These cosmic climate drivers are responsible for most of the variation of solar radiation received on Earth, resulting in medium- to long-term warming and cooling trends, independently from the sun‘s own rather small changing activity and radiation.
http://www.climateprediction.eu/cc/Main/Entries/2012/10/8_What_drives_climate_in_the_long_run_A_new_paper.html
DirkH, if the -5 degrees you are experiencing is a record breaking cold temperature, then there are three warm records broken at the same time somewhere else. The ratio of hot records broken to cold records is about 2.8.
The problem is what you are experiencing is on a little speck of the Earth and we should be looking at the world as a whole.
@Steveta_uk
Radiative cloud fraction is different from CO2. See here:
http://www.climateprediction.eu/cc/Main/Entries/2011/6/23_Prediction_of_Radiative_cloud_fraction.html
@YEP
„Data-driven models are good exercises to go through when analyzing a complex, dynamic, non-linear system. But nothing is theory-free, other than simple vector autoregression. Choosing the 6 variables, for example, had to be based on theory.“
No. Things are different here. The chosen 6 variables and also their time lags of up to 120 months (which sums up to several hundred inputs, actually) are just POTENTIAL inputs. Self-organization develops an OPTIMAL COMPLEX model out of the given data. This includes composition of the model, self-selection of relevant variables and their dynamics, validation. No theory involved. However, if there is well-known a priori knowledge (theory) it should be used also for data-driven modeling.
@Ian
„Has the model been used to make “hind-casts”? If so were they accurate? If not will hind-casting be attempted?“
No, no hind-casting but real forecasting. The reason is simple: Satellite data only go back until 1979. So there is no data that could be left out for hindcasting.
@Truthseeker
„Let us assume that they have correctly identified the most significant variables (which do not include CO2 – IPCC and alarmists please note) and can get a good correlation for past data. The problem is still predicting the values of those variables. Maybe they can use models to predict the variables they are using in the model to predict climate. Of course then they will have other variables to predict, which will mean other models to predict the variables they need for the models to predict the variables they need to predict the variables they need to predict the climate. Then they will need models … ad infinitim …“
No. It is an interdependent dynamic system of 5 variables (plus CO2, but it is independent). That‘s it. No more.
http://www.climateprediction.eu/cc/Main/Entries/2011/9/13_What_Drives_Global_Warming.html
Moe says:
January 25, 2013 at 3:25 am
DirkH, if the -5 degrees you are experiencing is a record breaking cold temperature, then there are three warm records broken at the same time somewhere else. The ratio of hot records broken to cold records is about 2.8….
>>>>>>>>>>>>>>>>>>>>>>>>>>>
Well Moe, the warm record sure as heck is not being broken across the ocean in North Carolina! It is 15°F ( -9.4°C) The record min is 17 °F
Notice how record breaking warm is always a sign of climate change but record breaking cold is not?
@Nicola Scafetta
„The only problem with the above figure in the post is that it seems that the latest temperature dot for Dec/2012 is located at Jan/2012.“
I have checked it, it is correcly shown at Dec 2012. Thanks.
“The simplistic linear cause-effect relationship “more atmospheric CO2 = higher temperatures” the IPCC model is based on is not an adequate tool to describe the complexity of the atmosphere sufficiently.”
This statement is the gist of my skepticism.
When I first heard the argument that a trace gas at 350ppm drives the climate. period.
Too simple, too cut-and-dried. I became a skeptic. Some time later I found WUWT, and realized I wasn’t alone.
Hey Moe, Hey Moe. CO2 causes global warming. Mean average temperature rises. period.
That’s the theory. Extreme temperature records go back 150 years or so. A very short time.
Breaking records does not relate to average global temperature.
The Earth’s temperature has not risen in 15 years. While CO2 has gone up. Get it – no correlation. Never mind causation. (Where’s the link for your Stat of “2.8”?)
Gunga Din says:
January 24, 2013 at 9:33 pm
========================================================================
What risk? The claim has been made. The last 16 years have shown the claim is without merit. (Substitute ‘hypothesis’ for ‘claim’ if you wish.) What “brainer”? Why risk more?
You’re free to risk your own well being or pay out of your own pocket but not mine. Would you use Government to force me to do so? If so, we’re back to my original comment.
——————————————————————————-
Whoa there. I didn’t say anything about using the Government to force anybody to do anything. My position is quite the opposite, actually; if CAGW were in fact proved beyond a shadow of a doubt I’d still rely on mitigation via the free market, capitalism, and innovation as the best bet to see us through.
But you’ve got to call a spade a spade. There is always risk when you mess with complicated systems, of unforeseen consequences if nothing else. Life is FULL of risk, as I’m fond of saying, this isn’t a wiffle ball world where everything is guaranteed to be safe. I don’t think the guys who try to model climate (or even predict the weather for that matter) have difficulty succeeding because they’re incompetent fools, but because it’s a damn hard thing to model or predict. Maybe it isn’t a tractable problem at the end of the day. But a consequence of this is that if something is hard to predict, it’s hard to know (as in know with scientific certainty, not just know as in good enough for me know) how the system is going to react to being perturbed. Do I THINK CO2 is going to hurt anything? No. Do I have some reasonable evidence of it? Yeah, I think so. Am I scientifically certain of this? Heck no, I haven’t done the work.
Was Mosher standing up a straw man with the Geo engineering thing? Sure. I don’t know Mosher at all so maybe I’m giving him too much credit, but it seems to me he’s plenty sharp enough to know that. Who argues that we should burn fossil fuels for the sake of geo engineering, well, nobody. But people (myself included) do argue that it’s plant food and that it may well be beneficial to the environment and mankind. I took his comment to heart because I’d privately accepted the argument that CO2 is going to be good for the environment without subjecting the that theory to the same skeptical rigorous standard I subject AGW to.
~shrug~ Just to explain where I was coming from. I might be nineteen kinds of wrong here, I often am. I therefore value comments like yours (and Mosher’s, incidentally) that help me see it when that’s the case.
Regards.
GAH!!! – I’d still rely on ADAPTATION, not mitigation. I was wondering what my foot was doing sticking out of my mouth after finishing my last post…
Frank Lemke says:
January 25, 2013 at 4:14 am
“No. It is an interdependent dynamic system of 5 variables (plus CO2, but it is independent). That‘s it. No more. ”
Frank, I think you should see if those same inputs can also generate the CO2 output. That would help sort out cause and effect. If that alone won’t do it, add the human co2 signal. This will clarify how much of the rise is human based.
Having spent a lot of time modeling, and running simulations, I like what you’re doing here, it’s time this was done.