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
So, we can relax again. Their prediction is roughly what we have seen out of the window for years now. Nothing is happening at all. Another 30 years of a non-warming of the planet will be catastrophic for Catastrophic Anthropogenic Global Warming. It will be interesting to see who is the last man standing, as there will assuredly be one.
“The Global Warming Prediction Project…” These are the first 5 words and the name of the project. Hardly a “impartial, transparent, and independent project” if you are assuming there will be Global Warming to predict over the next 60 months or so. Still, I’ve been asking the question for some time – has anyone tried running the GCM models and shrinking the climate sensitivity factor to see if the predictions can be flattened down until it to fits into the subsequently observed range. If IPCC had used 1C/doubling of CO2, their forecast ranges would have at least included the observed record that came to pass.
Mosher “people think it is safe to geo engineer the planet by dumping C02 in the air.”
Have you honestly not, over the past decade or so begun to reconsider downwards the degree of influence that CO2 must have? I believe the IPCC has. What will it take to give you second thoughts? The planet has already tested CO2 geoengineering many times in the past.
Ah a neural network. Those do indeed work reasonably for short term prediction, but long term is usually not quite as good. But ok, lets see.
Steve Mosher wrote;
“yup. that is why it is weird that people think it is safe to geo engineer the planet by dumping C02 in the air.”
Yup, that is ALSO why it is weird that people think it is safe to geo engineer the planet by dumping;
Mercury (CFL bulbs)
Silicon Tetraoxide (Si PV Cells)
Lithium (batteries)
Radioactive mine waste from mining for rare earth elements (Advanced Magnets)
Etc.
Etc.
Etc.
On the planet.
The only difference between the alleged “geo engineering” that has happened in the past and the “geo engineering” you advocate is we DON’T KNOW THE UNINTENDED CONSEQUENCES (AND YOU CAN BE SURE THERE WILL BE MANY) OF ALL THE “GREEN SCHEMES” YET…………….
Cheers, Kevin
How can I take a predictive climate model seriously that does [not] factor in chickens and goats? As near as I can determine, accurate “readings” of chicken entrails and goat fur is at the foundation of AGW prognostication.
Eyeballing the graph, the prediction is just BAU (=BUT), as so many commenters have already pointed out. It makes no sense. Stick the same BAU forecast anywhere you like in the past, and in a very few decades it will be a long way off.
My guess is that their forecast will already be visibly too high by its end year (2017).
Make that “does not” factor in….
“the presented atmospheric system model on the other hand is a CO2-free prediction model. ”
We should be thankful for small mercies … CO2 has no influence in this model and it seems to orbit closer to reality as a consequence. This must be a shuffle in the right direction … now for those negative forcings 😉
Ha Ha they said “quantitative prediction” and “non-linear dynamic system of the atmosphere” in the same paragraph!
3 pages – but i can’t bear to open 2 and 3:
24 Jan: Politico: DARREN SAMUELSOHN and JONATHAN ALLEN: John Kerry: Mr. Climate
Obama’s choice of John Kerry as the nation’s top diplomat is the strongest signal to the international community — and the smart set in Washington’s political class — that the president is truly committed to striking deals designed to save the world…
“Obviously, he has enormous credibility. I think that’s going to help,” said Phil Schiliro, former Obama White House legislative director. “Combine that with the fact that the president has such a commitment to the issue, and it sends a good signal.”…
http://www.politico.com/story/2013/01/john-kerry-climate-change-86695.html?hp=t1
Y’all might like to compare this forecast with the timing and amount of cooling predicted in my blog
http://climatesense-norpag.blogspot.com/2013/01/global-cooling-timing-and-amountnh.html
“Conclusions1) It seems reasonably probable – say 60-40 that the NH will cool by about .35 degrees by 2035.
2) We should be able to check the accuracy of this forecast by 2018 -20.”
The structural shortcomings of the IPCC models are also discussed there- this new method certainly is better structured than the IPCC models which are little more than drafting tools to produce AGW propaganda.
“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.”
So it is based on:
A) sunspot numbers
B) radiative cloud fraction
C) ozone concentration
D) aerosols
E) ?past? global mean temperature
(A) is solar.
In reality, (B) is not solely but heavily influenced by solar activity / solar-modulated GCR flux (as in the illustrations with water vapor and cloud variation in http://s7.postimage.org/69qd0llcr/intermediate.gif — click to enlarge), though such as the ENSO impacts it too.
(C) is also not solely but largely influenced by solar activity, as in UV.
That means the model is largely solar-based in the end.
While interesting, some weaknesses, though, include:
Sunspot numbers can break down as a metric when the probable Grand Minimum develops, since one time of 0 sunspots can be much different from another time of 0 sunspots in solar activity (and hence in GCR flux, etc.), as the last minimum already started to demonstrate, let alone the future.
And, particularly, as big questions for any prediction:
1) What exactly is the assumed future of solar activity, as in how is it being guessed for year-by-year future detail? You have to predict the future sun to start to predict future terrestrial temperatures, and different predictions of future solar activity vary. I would want to see not just the outputs but the inputs to the model assumed over future years; that even includes aerosols, like is it effectively implicitly assuming no substantial volcanic eruptions between now and 2017?
2) How are ocean oscillations like the ENSO being handled? Really predicting the ENSO in detail would be necessary for predicting future temperatures closely.
In initial looking at their website, I haven’t yet checked all the slower-loading parts yet, though, so some of what I am looking for might already be answered perhaps.
A note to all.
When reply to a comment, please at least copy/paste the name and timestamp of the commenter. If you don’t, then sometimes no one knows the thought or the person who “thunk it” that you are responding to.
(With those two bits of info then we can scroll up to the original comment.)
[The moderators strongly recommend also adding a “blockquote” before the quoted words, and a “/blockquote” after the quote. Mod]
“Give me four variables and I can model an elephant. Give me five, and I can model its trunk.”
And what about two variables?
For example, adequate reconstruction of HadSST2 temperatures of tropics (30S-30N) by linear regression could be achieved by consideration only 2 factors – ENSO (Nino34 index from HadSST2) and step climate regime changes in 1925/1926 and 1987. During these steps the mean value of temperature rises, over which natural variability associated with ENSO occurs.
Calculations are in this simple Excel file:
https://www.dropbox.com/s/63jb7fd3v4c8vmb/Tropical%20SST%20with%20shift%20%28for%20figure%205%20and%206%29.xls
Correlation coefficient for monthly mean time series – 0.86. Another remarkable moment is that coefficients of linear regression can be fitted by small amount of data – for example, from 1910 till 1940, for adequate reconstruction of the whole period.
Reality of climate regime shifts in 1925/1926 and 1987 is proved by many independent studies that are summarized in our preprint: http://vixra.org/abs/1212.0172
So lemme guess, the model mimics the recorded history of the HADCrud data that was used to construct it to 98% of something.
So now go and input the history of some other like time period, that HAS NOT been used to create the model; do you still get a 98% match to any 79 month period of data from the last 10,000 years or what.
They say the model is confirmed; how does that work.
There are completely fake models of the fine structure constant that predict the answer to about 8 significant digits; yet they have no input data from the physical universe.
So 98% good doesn’t confirm diddley squat.
Present levels of CO2 in the air are very low by historical standards. Biology says the level is near the starvation level for plants. Those claiming that adding more CO2 is a bad thing is a statement that does not pass the sanity check. Anyone who understands atmospheric evolution via the biological carbon cycle would say that adding CO2 to the air is good. Global ecosystems would be far healthier with 1000 to 2000 ppm CO2 in the air.
“””””…..KevinK says:
January 24, 2013 at 5:02 pm
Steve Mosher wrote;
“yup. that is why it is weird that people think it is safe to geo engineer the planet by dumping C02 in the air.”
Yup, that is ALSO why it is weird that people think it is safe to geo engineer the planet by dumping;
Mercury (CFL bulbs)
Silicon Tetraoxide (Si PV Cells)……””””””
So why don’t you lay on us that Silicon Tetra-oxide waste material from PV cells again.
I’m familiar with Silicon di-oxide; aka quartz, and even SiO which has virtuall no place in Si PV cells, but so how exactly does Silicon Tetraoxide go together, and where is it used in the Si PV cell ?
Well there is Silane of course, but one would likely call that Silicon Tetrahydride; not Tetra-Oxide.
This is some new device Physics I’ve never learnt.
KevinK says:
January 24, 2013 at 5:02 pm
=======================================================
I wonder. Does GE make the equipment to recover mercury from landfills?
(Mods, I tried the “blockquote” thing fro the first time time. Hope I got it right.)
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.
————————————–
Steven,
Nothing more irritating or valuable than the guy who keeps you honest, although
it can be bitter at times. You got me at least; if all things were equal (if economic devastation wasn’t in the equation) I’d just as soon NOT have a boatload of CO2 pumped into the atmosphere, certainly not just with geo engineering as a justification, and yes, I’ve comforted myself with the rationalization that the CO2 might be a good thing, without having any rigorous scientific basis for doing so.
All things aren’t equal though. I definitely believe the cost of avoiding the risk vastly
outweighs the cost of taking it to such an extreme degree that it’s still a no-brainer.
But still, your point is taken, and thanks for posting it.
Regards.
I would like to know how they plan on predicting the weather/climate for the future when they are using the data for the last 30 years of of a system that cycles over a 60 year period? We have about 180years of fairly good records, 3 full cycles! I have seen 1 full cycle and I know that there are good records for twice that length of time. 30 years of data only guarantees that the the next year prediction will be close and after that the error bars will expand.
My own prediction is that in 60 years weather/climate will be about the same as it is now, or slightly cooler, after cooling and then warming, Don’t need fancy computer programs to figure it out, JUST PAY ATTENTION when you read up on recorded facts written by the people of that time. Believe it! humans of a thousand years ago were every bit as intelligent as those of the present, and they had no reason to lie about the facts to promote an agenda. pg
WillR says:
January 24, 2013 at 2:17 pm
If I was looking at an electronic circuit I would say that the model is missing a damping factor (negative feedback) fwiw.
Any suggestions?
Yeah screw Maxwell and his equations…..
This just might be the reason for all our problems. Well worth a post on WUWT IMHO.
yeah, that 98% error bar has me wondering if they have a clue what they’re actually doing…
Mosher quips: “yup. that is why it is weird that people think it is safe to geo engineer the planet by dumping C02 in the air.”
So now you got something against 7B people breathing and trying to subsist in our world? What is with you greens? Your ilk is responsible for thinking up plans on how to geo engineer solutions or to waste money in worthless investment for a CO2 problem that has not even remotely proven to exist. You won’t be happy till everyone else is feeling as miserable as you. And if that ain’t tough enough for you, wait till Willis gets here.
So, If I followed what they do, they feed a number of possible input variables into an adaptive network, as well as the output values they want to model, using the inputs and outputs, it constructs an equation that’s based on some of the inputs (at least in this case), it’s this equation with in this case the output generated temperature feedback, that generates a new output value. This would work if the output generated is based on the the outputs past time series and the external input (Sunspots). So basically they are saying that future temps are based on past temps, and the Sun. They have a model for Sunspots, based on the Sunspot time series of the past. with these two equations, and the time series required (in their model the sunspot count, and global temperatures as collect over the last 33 years) their adaptive model (will calculate a new value time series).
here’s a picture of the climate model
http://climateprediction.eu/cc/Main/Entries/2011/9/13_What_Drives_Global_Warming_files/droppedImage_1.png
It doesn’t ascribe a physical process to their output, just an equation that’s based on the data used for input that will predict a matching output (temperature series). This isn’t a lot different than fractals, and other self referential chaotic systems. Dr Glassman has a simple 2 part equation that did a good job of calculating temps that’s similar in nature.
It is a good thing that it doesn’t reference co2, the network did not find the co2 signal in the output. This is what we’ve been saying.
Where it would have problems is for dependent inputs longer than 33 years. So for instance the AMO/PDO time series if it’s longer than 33 years, and represented in the temp time series, which I would expect it to be, will only contain it’s effects from the 33 year period. Depending on the network delay parameters it might predict the whole cycle, but it is under sampled and probably won’t. But since I think both have switched during the last 33 year, it might do okay.
This is like Newton coming up with the equation for gravity based on a couple of inputs, but it’s a much simpler equation with fewer dependencies. That’s why no human has been able to do the same with climate.
It’s the same as using a fft to decompose some complex wave form into a collection of sine waves, that when summed recreate the complex waveform, but instead of a bunch of sine waves, this is an equation.
Easy peasy.