The Blunt Truth about Global Warming Models
By Paul Homewood
h/t Philip Bratby
I may be one of the first scientists in the country to know that predicting long-term temperatures is not possible.
Almost 50 years ago, while in grad school, I had a contract from an Army research lab to use a state-of-the-art models to predict long-term temperatures. I quickly realized that the goal of the project, to forecast accurately the temperature long-term, was impossible because small errors in data inputs could result in huge forecasts errors. Equally important was that errors compounded so quickly that it caused the error ranges to explode. The results were junk. As an example, what good is a temperature forecast with an error range of plus or minus one hundred degrees?
I give university speeches to scientists and tell them: if you ever see some data or forecasts, your first question has to be "what’s the error range?" If you don’t know the error range, the data are almost useless. It’s not coincidental that the Climate Mafia don’t highlight this problem.
Full article here.
Comments are closed.
NOT A LOT OF PEOPLE KNOW THAT 
What a refreshing change! Somkeone putting the question mark in the right place. Excellent. I would however, like some more meat on the closing paragraph about MrNA vaccines?
About “the closing paragraph”
It seems to come out of the blue — I wonder what he knows and if he has made a more complete statement elsewhere. Why? Because Our-World-Data says: “70.5% of the world population has received at least one dose of a COVID-19 vaccine. 13.5 billion doses have been administered globally, and 14,863 are now administered each day.”
A small percentage of people die every day so if 5.5 Billion are vaccinated there will be deaths.
So yes, there is no nuance here.
The NWO mass extinction event perhaps?
Yes, it was not at all obvious to me why he put that sentence in.
I agree, a welcome common sense approach. I also do not know enough about the mRNA vaccine to question his belief.
“I would however, like some more meat on the closing paragraph about MrNA vaccines?”
How long have you got? Firstly, the Pfizer Biontech pre and post EUA trials have been so comprehensively exposed as monumentally fraudulent – from their own report as revealed in the US FDA releases by order of a US Judge.
Secondly, from what I know, the delivery system – LNP – was shown to be toxic to humans because of an evident inflammatory response, information in the public domain, since ~2009 AND mRNA so called (SARS COV2) vaccines (inducing the production of a protein inimicable to humans) which are experimental gene editing drugs ( evidenced by Japanese scientists who have found S1/S2 protein impairs the immune system despite efforts to trash it – “SARS-CoV-2 Spike Impairs DNA Damage Repair and Inhibits V(D)J Recombination In Vitro”) which have been grossly inadequately tested – on a very limited, healthy cohort……..it gets far far worse than this.
None of this is great for human beings……
I heartily agree. Spike protein found in cadaver liver, brain, heart and worst of all womb. The unexplained excess deaths all over the world that no-one in the MSN seems to care about. Myocarditis rates through the roof and spike protein found again. But this is the line I would like more information on….. “after that, selling the lie that even though mRNA vaccines have never worked in thirty years and generally killed all their test subjects ” some proof would be excellent in forming the argument!
To Chris Treise: this is off topic so sincere apologies to PH for the diversion; see this for some background: https://rwmalonemd.substack.com/p/why-is-usg-rewriting-the-history?utm_source=substack&utm_medium=email
The basic truth is that very, very few forecasts of anything interesting are ever right. Yet we are deluged by more and more forecasts every day, none of which have any chance of being right. As Mervyn King sets out in his recent book on economic forecasting, models only work when everything is static. But nothing interesting happens in such models. It’s only when things stop being static that interesting (good or bad) things happen and it is then that the models fail. The idea that we can model change is simply completely fallacious in virtually every system because we cannot model what we do not know. Alarmists however claim that not only are their climate models right but their models of climate without human CO2 are correct (which is of course also unknowable) AND their models of the global economy going out decades under various climate scenarios are correct. This is beyond laughable. No serious, intelligent person should consider such a thing possible.
“I believe that economists put decimal points in their forecasts to show they have a sense of humor.” – William Gilmore Simms, 1806-1870
The absolute killer paragraph is :
“From the 2001 IPCC Third Assessment Report: “The climate system is a coupled non-linear chaotic system, and therefore the long-term prediction of future climate states is not possible.”
Yet, we see the climate alarmism getting ever more desperate with the UN Sec General’s absurd “boiling planet” to the Met Office “heatwave summers are the future”!
Slightly off topic, but not only can they not get the forecasts right, but the Met Office don’t even get the daily temperatures right. According to the Daily Mail we have had a mini heat wave this week. I hope most of you have enjoyed it, but here in Cheshire the only heat I have felt has been when my heating has come on.
Mini heat wave my foot – humidity high, temperatures in the mid 20’s – FFHS how is that a heat wave?
Normal weather , sticky nights for August….FO Met Office!!
I found his main point rather lazily made principally by conflating prediction with forecasting. In the sense used in the opening sentence of the second paragraph it was clearly about climate probability distributions, for example the relative frequency of various weather patterns. But he then proceeds to ridicule the idea by pretending the topic was about forecasting in the sense of saying what the weather is going to be like at some future moment contingent on its current state.
While I also deny the feasibility of predicting future climate (doubly so as a basis for policy) that is not because chaos renders projection of a current state of the atmosphere beyond a few days into the future indeterminate. If weather had one clear and totally dominant forcing agent whose sensitivity was known and whose future state could be surmised then predictions of the future makes sense. But that is patently not the case.
The sun can play the forcing agent role better than anything else.
You serious? Do you honestly think that knowing about “the sun” you will suddenly know all about, say, day-night, summer-winter, coastal-inland, lowland-upland, land-ocean, cloudy-clear, droughty-moist contrasts and variations; in other words these and all the other myriad tiny details of what we call and experience as climate.
Very true. Climate projections are wrong for two reasons: first we cannot know nor model starting conditions. Second, climate is as the IPCC said, a coupled, non-linear system. There is no dominant force, everything affects everything and there are theregore, in terms of modelling, no “causes”. Of course its also true that climate scientists don’t properly understand the physics of climate and almost certainly don’t know all the physical processes that affect climate. But rather than admit the problems, they demand more powerful computers to model in greater detail, which can only increase the error.
I am not sure that you have grasped the difference between uncertainty and a chaotic complex system.
In quantum physics uncertainty can be quantified precisely by the Heisenberg principle. The probability density function of an electron orbit can be determined through the Schrödinger equation. It turns out that the average radius of the orbit is equal to the classical radius for the simple hydrogen atom, justifying the use of the classical radius for some calculations. Statistical averaging works, at least within the limits on Heisenberg uncertainty. The amplitude that the glass on your table will spontaneously indulge in telekinesis is vanishingly small.
In a chaotic system it is possible for semi-stable states to persist for some time (meaning that parameters change only a little between iterations), but they may reach a bifurcation point at which a small difference can result in very different future paths: you can find examples of this within the Mandelbrot set for example. Even estimating the probability of such events becomes effectively impossible in a complex system – it’s hard enough/impossible in the system z(n+1)=z(n)^2 + c, z(0)=0, c a chosen complex number |c|<2 that defines the Mandelbrot set – certainly a handful of kludged models are inadequate to estimate it. The models are steered by the parameterisation black boxes that are used in absence of computing power and data or even understanding of the underlying physics (e.g. in clouds understanding remains rudimentary).
Not sure who this is addressed to as yours is the only mention of the word “uncertainty” above. In normal climate parlance the word would not generally be associated with such a high-falutin issue like quantum physics. It would most likely come up when discussing more mundane matters like instrument error, observational deficiencies, model inadequacy – things we could (or should) surround with error bars and propagate through to the final output.
My instinct is that climate processes sitting inside larger Earth system ones must be subject to strong attractors and not behave like the sort of predator-prey or Mandelbrot-type chaos generating trajectories that are the stuff of textbooks. So while those pretty phase diagrams you get out of Lorenz’s differential equations might apply locally in the atmosphere, they soon rub up against some other buffering subsystem elsewhere within the total system that nudges the subsystem back within bounds.
I realise this instinct is just a play-back of the anthropic principle – we wouldn’t be sitting here talking about it if things were as wild as those phase diagrams suggest. So is indeterminacy after a few days really the product of coupled non-linear systems or just what you can expect from poor process descriptions and inadequate data?
Quite right Eastdevonoldie! In all the years I have been following the CC nutters I have never once seen a counter argument to the ‘non-linear…..’ exposition.
The BBC are reporting a prediction of the penguin numbers in Antarctica in 2100 on their latest warning about carbon dioxide. Any biologist worth their salt knows that population numbers are totally unpredictable. I cannot see how they can report this with a straight face. They have no idea about fact checking. “Not even wrong” – Wolfgang Pauli.
This article reminds me of a similar topic from James Gleick’s 1987 book, Chaos. He gave the example of DNRs around the country estimating wildlife populations to set bag limits for the coming hunting season. He pointed out the equation used was not linear, and moved into chaos by the second iteration. I have been told our local DNR is staffed with the otherwise un-employable brother-in-laws- of the department heads. Pick a number–any number.
DNR? Do Not Resuscitate?
Department of Natural Resources?
DNR = Dept of Natural Resources, yes. That is what we call the fishing/hunting administrators in the U.S. Every state has one, and they are as merciless as the are ubiquitous.
I was turned off by the first sentence.
‘I may be one of the first scientists in the country to know that predicting long-term temperatures is not possible.’
Okay, yeah, you may have been the first scientist, but non scientists already knew it. It’s preposterous on its face.
‘to use a state-of-the-art models’
Can’t do singular/plural?
‘was impossible because small errors in data inputs could result in huge forecasts errors. Equally important was that errors compounded so quickly’
Well, this would be the GIGO that some on these pages love to cite.
But it’s a lie. IF THERE WERE NO ERRORS IN THE INPUT, you still couldn’t predict future weather. Therefore, it is creepy to cite it.
Yes, this point annoys me too. Lots of people claim Covid models are the same as climate models and that’s why they were wrong, but that is false. Viruses are linear and the infection curve is a very well known shape. To model any virus accurately you simply need good inputs – transmissibillity, IFR, current infections. But we didn’t have those for Covid so Imperial built a model that went into vast detail to pretend that such detail could produce unknown data. By contrast we simply cannot model climate, even if we knew all the necessary inputs (which we can’t e.g. starting conditions).
But of course as soon as you introduce human response into the picture that pretty much poxes up those nice forward projections of the course of the disease at the population level. In this regard modelling epidemics and climate part company in that human response is marginal – maybe even smaller than marginal – in its influence on the course of the basic process.
Yes and no. If the difference is measurable, you simply reduce transmissibility and you can put that new number in to your model. If you know that people will react in that way and that reduces infections, it’s forecastable at the outset. But new viruses inevitably become endemic and inevitably mutate to transmit more easily, because all a virus cares about is infecting someone new.
My future “climate” forecast is to take, say a ten year period, such as 1990 to 2000, or 2005 to 2015, or 2010 to 2020. Describe it in about 250 words and fewer than 11 numbers.
Next, explain that any 10-year period out 70 or 80 or 100 years from now will be remarkably similar to the 10-year current period.
Amongst the apocalyptic and increasingly implausible forecasts for warming, we very rarely seem to hear of their 95% confidence intervals …
The high gain in errors, may explain why there is so much 1/f^n noise. However 1/f^n noise, is predictably unpredicatable.
It is more than time that soothsayers paid for by Government using your tax dollars or pounds, should have some encouragement to be accurate.
Before any of their prognoses are given any attention, let alone used to justify policy proposals, they should be made to predict what parameters will actually measure in the year they retire.
When that date arrives, any variance from their predictions greater than 10% must imply a reduction in their pension by a similar amount. Twenty percent or more wrong? Thanks, you just lost your pension.
Such a scheme might help to concentrate minds.
To be fair though, governments demand forecasts of things that we cannot forecast. The OBR in the UK has got its forecasts on government revenue and spending and thus borrowing wrong again and it’s forecasting just a few months out. But it can’t forecast such things accurately. It probably knows that but government still demands a forecast. The story from WW2 is that a statistician looked at the accuracy of long range weather forecasts being used and determined they were worse than random guessed. He showed this to Eisenhower who nodded but then said “but I must have something to help me make a decision.”
I agree. But, it is obviously desirable, even necessary to have an idea how many extra school places or new homes will be needed in ten years time. And being even 50% out isn’t the end of the world because the projected number will likely be “right” after another ten years.
I note that this type of forecast is either NOT produced today, or is based on absurd assumptions and data that is manifestly wrong.
Witness the housing problems, provision of medical services and school places etc.
But the only need for modelled estimates of “global temperature” or “ice extent in Antarctica” or all the myriad other things, 50 or 100 years hence, is for agit-prop purposes to destroy what we have now (and have paid for, and what works) and to promote GangGreenery. (And nice brown envelopes for the alarmist “Scientists”.)
And note their precision. We might be looking at an extra 2°C. Or three. Or some “experts” say 5°C. Curiously those who suggest zero or minus 2°C are ignored.
They then suggest that all this malarky is “consistent” with their lunatic theories. Ignoring, as I wrote in the past, that it is also “consistent” with the old speculations concerning naughty witches.
The soothsaying must be controlled and the media’s and politician’s use of it reduced.
Thatcher was told that the population of the UK would decline over the next 20 years by experts. Businesses were told that they could take pension holidays because they were over-funded. They weren’t wrong by 50% but wrong in entirely the opposite direction. Forecasts are too often wrong like that, and thus we do the exact wrong things. And then there’s opportunity cost when we are wrong but in the right direction.
With the OBR and BoE you have to assume that they aren’t interested in accuracy, only in pursuing a political agenda. I believe that the recruits for the OBR have come preponderantly from a rather narrow range of think tanks. The BoE had a blood letting exercise under Carney where they got rid of the talent that thought in different ways outside the box, challenging the new “conventional” views and making for a proper debate. It’s why hearing from Mervyn King or Charles Bean is likely to give a more perceptive insight.
” …[penalty] scheme…”
Ancient Athens had such a scheme for the architects of its public works.
However, the UK has such a cosy, revolving-door, scheme for its top managers that – as we all know – the penalty for failure is a move sideways, with a doubling of pay.
The hilarious truth about climate models, is that the alarmists often fail to guesstimate tomorrows weather, never mind the climate in 50 years time – no better than Mystic Meg reading the tea leaves
Soviets were predicting capitalist sharknados back in 1961!