Claim: Climate Science Does Not Have to be Falsifiable

Guest essay by Eric Worrall

How do you falsify a climate model? Australian National University Climate scientist Sophie Lewis acknowledges that climate models are not falsifiable – yet claims we should trust them anyway.

Climate change has changed the way I think about science. Here’s why

Sophie Lewis

Research fellow, Australian National University

August 10, 2017 3.30pm AEST

I’ve wanted to be a scientist since I was five years old.

My idea of a scientist was someone in a lab, making hypotheses and testing theories. We often think of science only as a linear, objective process. This is also the way that science is presented in peer reviewed journal articles – a study begins with a research question or hypothesis, followed by methods, results and conclusions.

It turns out that my work now as a climate scientist doesn’t quite gel with the way we typically talk about science and how science works.

1. Methods aren’t always necessarily falsifiable

Falsifiability is the idea that an assertion can be shown to be false by an experiment or an observation, and is critical to distinctions between “true science” and “pseudoscience”.

Climate models are important and complex tools for understanding the climate system. Are climate models falsifiable? Are they science? A test of falsifiability requires a model test or climate observation that shows global warming caused by increased human-produced greenhouse gases is untrue. It is difficult to propose a test of climate models in advance that is falsifiable.

Science is complicated – and doesn’t always fit the simplified version we learn as children.

This difficulty doesn’t mean that climate models or climate science are invalid or untrustworthy. Climate models are carefully developed and evaluated based on their ability to accurately reproduce observed climate trends and processes. This is why climatologists have confidence in them as scientific tools, not because of ideas around falsifiability.

Read more:

The problem with Sophie’s position is that fitting a model to past observations is not a test of whether the model is right; all fitting the model tells you is that you have found a way to fit the model. What counts is the ability of the model to predict the future – to accommodate observations which were unknown at the time the model was created.

“Careful development” just means current prejudices are carefully applied. But there are many more ways to be wrong than right – especially about something as complex as the global climate.

Climate scientists are desperate for their educated guesses to be accepted as science; so desperate that at least some climate scientists openly challenge the very keystone of science, the requirement that scientific theories must provide a means by which they can be falsified.


newest oldest most voted
Notify of

Computer chip designs are complicated, yet we never put one at the helm of your aircraft unless it has been thoroughly modeled.

Rick C PE

Well, I would certainly hope that any computer chip at the helm of an aircraft I happen to be on be thoroughly TESTED.


For Bernie:

It turns out that my work now as a climate scientist doesn’t quite gel with the way we typically
talk about science and how science works.

From the above quote, I have replaced all “climate science” or “science” words with “Engineer” or Engineering” in the below quote

It turns out that my work now as a Engineer doesn’t quite gel with the way we typically talk about Engineering and how Engineering works.

Would you ride on my plane?

Not falsifiable =/= science.


Where exactly are computer chips “at the helm” of airplanes?
Did you mean to say “used in the helm controls”, or words to that effect?
There was just a survey done recently, in which the vast majority of people said they would not fly on an airplane without a person flying the plane. An onboard person.

Patrick MJD

“Menicholas August 10, 2017 at 7:27 pm
There was just a survey done recently, in which the vast majority of people said they would not fly on an airplane without a person flying the plane. An onboard person.”
Aircraft already do fly themselves for the most part. Aircraft have been able to take off and land since about 1965 (London Heathrow it was first tested IIRC). And automation is increasing so much so many pilots like to fly older craft BECAUSE they actually have to fly it.

Menicholas- here is some food for thought. Pilot Error versus computer.


I had a long response I was assembling that included links to various stories over the past year re the guy in Williston who drove his Tesla into a tractor trailer turning into his path, and went from there.
It began with my admission of having misstated the issue…people want a pilot in the cockpit.
You know…someone to turn on the autopilot and set it, someone to be on hand to, wel,l frinstince…land the plane…
It went away when wordpress bug caused my page to freeze up…i am too tired to redo it.


I can usually see why a comment was bumped into moderation…wondering what was in that last one to do so.


Oh, probably the first four letters of the word for the forward most section of an airplane.

Chips are intelligently (I hope) designed and thoroughly tested. They are based on good science and engineering. Complex systems, such as human behaviour and climates, are so complex that all-encompassing models are dangerous. It is better to do what science is very good at: breaking complex problems down into small steps that can be worked on using the tools of rational science. Most of the proponents of global warming I know have a geography background. Geography is an eclectic discipline, so is the antithesis of most science. This is not to say it isn’t valid – it just should not make unsound predictions.Stick to maps.


Bit of a side issue, but computers aren’t intelligent. They only think they are.

Menicholas — after several such mishaps, I have made it a habit to compose longer comments in a separate editor first and save them as a local file.

DC Cowboy

Computers don’t ‘think’ at all. It’s people that think they are intelligent when they aren’t.
Back in the later 70’s early 80’s I was selling homes for a builder and programmed a Sinclair Z80 (with a cassette tape ‘program’ and a very small, portable TV as the monitor) to handle loan qualifying for my prospects. I quickly noticed that people were far more willing to accept the computer results than they were the exact same calculations performed manually with paper, pencil and a calculator. I always found it striking because I realized I could ‘program’ the computer to give whatever results I wanted them to see and they would accept them over ‘human computing’ even though they couldn’t see or check the computer generated numbers.
As far as AI Pilots go, the problem set for flying a commercial aircraft is far less complicated than that for driving a car in traffic (my father turned down American Airlines in the 50’s because he said he didn’t want to be a ‘glorified bus driver’ no matter how much they paid him), yet people seem more accepting of that than they are of an AI Pilot. I think pilots of the future will end up being in a position that they don’t fly the plane at all, they’ll just sit there through the flight in case something goes ‘wrong’ and they will be paid far less than today. I’m pretty sure carriers like Fedex are exploring unmanned transport and the Navy has successfully ‘taught’ unmanned aircraft to make carrier landings, pretty much the most complex thing a plane has to do.

Samuel C Cogar

Where exactly are computer chips “at the helm” of airplanes?

All modern passenger carrying airplanes have …… computer chips “at the helm” which is commonly referred to as “flying by wire”.
Auto-pilots also depend on computer chips “at the helm”.

I don’t care about the chip I care about the pilot.


Menicholas, Michael Palmer, others, too,
I too have this irritating problem with this site – is it due to WordPress? – that sometimes freezes on me it I go to another tab [ yes, attention span of a gnat, I know] – say to check sport scores – and come back.
Even if I am not composing a gripe!
There we go.
Still come back most evenings.


What if the pilot’s name is Chip?


Auto: I had that problem until I installed ad-blocker.

Samuel C Cogar

Menicholas …… “It went away when wordpress bug caused my page to freeze up…i am too tired to redo it.
Auto …….. “ Menicholas, Michael Palmer, others, too,
I too have this irritating problem with this site – is it due to WordPress?

MarkW ……. “ Auto: I had that problem until I installed ad-blocker.
To all the above, ….. I’ve experienced all of the above with my PC, and it seems to be getting worse each day …… and it’s not just Worldpress …. but FOX News has become the most annoying web site that I regularly view/visit.
Try reading a news article on FOX …… and a window “pops” up and a video of that news article starts blaring in your eardrums. What sort of dumbass programmer or webmaster would think that “watching n’ listening” to a video of an article while reading the printed article is what all their viewers want?
Those sites are so screwed up that I have to keep switching back n’ forth between Internet Explorer and Microsoft Edge … and am constantly having to use the Task Manger to “End Task” when they “freeze up” in an internal “loop” and gab 100% of C PU time and 99% of available RAM.


Computers don’t ‘think’ at all. It’s people that think they are intelligent when they aren’t.

Do you think?
Now careful. Traditional C is not “intelligent”, but neither is traditional DNA. Intelligence is tricky to define.

Patrick MJD

Funnily enough, they are tested, to destruction, well before they get to fly commercially.


So was the electrical componentry on the American Vanguard rockets that kept blowing up on the launchpad.

Bill F

Bernie that makes zero sense. I award you no points.


He’s trying to claim that since we trust models in one area that proves that all models are good.

D. J. Hawkins

@Mark W
Re-read Bernie’s post. He was saying the exact opposite.


Models are subject to validation and verification. It is impossible to validate and verify climate models because the climate is chaotic. That was discovered by Edward Lorenz, a pioneer numerical modeller and discoverer of chaos theory. Idiots have been ignoring him for the last half century and have wasted billions by promising that the next super computer will be able to predict the weather more than a week in advance.
What you are referring to about aircraft could be model based design. It’s an extremely rigorous design methodology. It’s the polar opposite of climate modelling.

bob – if you want to test a model, tell
us the co2 emissions for the
next 30 years, and the changes in
solar irradiance, and the major volcanic eruptions
over that time period.
then we can compare in 2047.


crackers345 August 10, 2017 at 6:32 pm
… thanks.

Just pick some plausible numbers out of mid air. The results will be equally valid.


Being limited to plausible numbers only would be overly constraining for a “climate scientist”, and anyone else who considers sticking to what can be objectively verified as being scientific.
Imagine a grown man or woman being so poorly educated as to think that childlike notions of objectivity applies to their own self, and the work that they do.
Then imagine the level of edutainment needed to reached the enlightened stage of cognition in which becomes obvious that when things get complicated, one has to invent one’s own definitions and standards if one is to become free of the fettering tethers of objectivity.
How childlike so many of us are…for thinking that the strictures of scientific methodology applies to those doing Gaia’s work!

David Blackall

From my recent paper, sorry I’m in a hurry, but the models can’t possibly get all the variables in there, so they are flawed:
Published in the eminent journal Nature (Ma, et. al., 2017), ‘Theory of chaotic orbital variations confirmed by Cretaceous geological evidence’, provides excellent stimulus material for student news writing. The paper discusses the severe wobbles in planetary orbits, and these affect climate. The wobbles are reflected in geological records and show that the theoretical climate models are not rigorously confirmed by these radioisotopically calibrated and anchored geological data sets. Yet popular discourse presents Earth as harmonious: temperatures, sea levels, and orbital patterns all naturally balanced until global warming affects them, a mythical construct. Instead, the reality is natural variability, the interactions of which are yet to be measured or discovered (Berger, 2013).
Science is a self-correcting process, and Karl Popper defined this as an empirical falsification (Popper, 1957). Scientists test, measure, observe and retest, and they must be able to verify and repeat results (Errington et al., 2014). Uncertainty is always present (van Der Sluijs, 2005), but when uncertainty is replaced by ‘consensus’ (post-normal science), a culture of gatekeeping ensues (Lindzen, 2009). Post-normal science is said to be appropriate when ‘traditional methodologies are ineffective. In those circumstances, the quality assurance of scientific inputs to the policy process requires an ‘extended peer community’, consisting of all those with a stake in the dialogue on the issue’ (Funtowicz & Ravetz, 1993). Then, and dangerously, dissenters are silenced so that chosen and ‘necessary’ discourses arrive in journals, conferences, and boardrooms.

“commieBob August 10, 2017 at 6:03 pm
Models are subject to validation and verification. It is impossible to validate and verify climate models because the climate is chaotic. That was discovered by Edward Lorenz, a pioneer numerical modeller and discoverer of chaos theory. Idiots have been ignoring him for the last half century and have wasted billions by promising that the next super computer will be able to predict the weather more than a week in advance.
What you are referring to about aircraft could be model based design. It’s an extremely rigorous design methodology. It’s the polar opposite of climate modelling.”

You are conflating “rigorous design” with “rigorous testing”. They are not the same.
There are engineer specialists, known as “test” engineers; for virtually any branch of engineering.
Based on requirements, test engineers design tests, including hardware, software, data, decision tree, human error, etc. to rigorously test equipment, software, processes, etc.
Even the process of testing typically generates more conditions and variables that get added to test decks and scenarios.
Only after rigorous and documented testing is a product validated and perhaps certified.
Models can simulate testing designs and they do help weed out problems before proceeding to a product’s next step. Models do not replace testing.
Which is one reason climastrologists are frequently reminded that they should submit their “climate models” for testing, validation, and certification.
Climate models should publicly acknowledge versions with a rigorous software change process to prevent insalubrious model adjustments.
As long as climastrologists avoid treating models as verifiable products, climate science is extremely unlikely to match observations for any length of time.
The thought does give some amusement. The thought of a test engineer reading the “Harry, read me” file before interviewing climastrologists for their “model” expectations.
Perhaps we should start nick naming climastrologists after various types of fudge?
Maple nut fudge?
Rocky road fudge?
Baldy mountain fudge?
Bubble gum beard fudge?
Treacle fudge?
Peanut butter fudge?
Tofu fudge?
Double dip chocolate fudge?
Sugar free fudge?
Seriously hot fudge?
Praline pecuniary fudge?
Tortured caramel fudge?
Pulled taffy fudge?
Double bias fudge?
Fracking free fudge?
Carbonized fudge?
Carbon free fudge? No carbon atoms whatsoever.
Callous arrogant flavor fudge?
Climate debate coward fudge?
Publish perish fudge?
Waffle words and fudge?
Seriously. Don’t some of those names remind of various climastrologists?


ATheoK August 10, 2017 at 8:49 pm
… You are conflating “rigorous design” with “rigorous testing” …

Model Based Design includes both in spades.

Joe Crawford

ATheoK, I had to laugh when you said: “…get added to test decks…” It brought back memories of my early involvement with computers. At that time test decks were actual decks of IBM punch cards. The test decks for a new hardware/software product might require anywhere from one drawer to many whole card filing cabinets depending on the complexity. We even carried blank cards around in our shirt pocket for taking notes.


As the climate models lack about 50+ major climate factors and ridiculously exaggerate the role of CO2, climate models are junk. To rely on them for anything but failure is to perpetrate a fraud. Computer models are also patently NOT SCIENCE as they are programmed to do what they want.
Real science can be done on a computer. For instance, the different Milankovitch cycles can be modeled as wave patterns and then overlaid and their interference patterns examined. The resulting interference pattern recreates quite nicely the records we have for sunspot numbers and solar cycles, showing a high correlation. It’s for sure that the Earth does not cause sunspots, but it is quite likely that solar factors related to sunspots affect Earth’s climate. Of course, the climate computer models pretend that the Sun does not vary enough to affect our climate.


Yeah, but climate models are ‘close enough for government work’ (i.e., redistribution of taxpayer money).

Anders Valland

Bernie, what did you mean by your statement? It really does not make any sense, as the statement is way too small for the issue at hand.


So what? Aircraft are designed to have the same response to the same input every single time. We design the system to do that. And the number of possible inputs is limited. That means any model of that system is going to be pretty accurate.
So now design a chip to fly a plane when there are a thousand unique inputs and the response to each input varies essentially randomly each time.
Can you design that chip? Can you then model whether it works?


That chips are modeled before the first one is built is well known.
Chip fabrication is hideously expensive and modeling is relatively cheap. So they use models to verify as much of their design as they can before committing to silicon.
However, they still spend months testing the actual chip before any of them are sold to the public.


Does anyone remember the Intel Pentium? They discovered a bug in the math co-processor after a couple of million had been sold. Microsoft had to deliver a software patch to “fix” the problem.

The RePentium.

What counts is the ability of the model to predict the future.
No exactly. What counts is the ability to predict the unknown. So for example, one could design a double blind experiment in which the past was hidden from both the model and the model builder, and see if the model could correctly predict the past.
The problem is that the model builders are “peeking” at the model results, and using this to modify the model. So in the end the model is not making a prediction. It is the model builder making the prediction, by modifying (tuning) the model such the results of the model match what the model builder believes to be correct.
This is the fundamental problem with climate models. They are not modelling climate. They are modelling what the model builders believe climate will do in the future. Because otherwise, if the model results don’t match what the builders believe, even if the results are 100% accurate, the model builders will see this as an error in the model and adjust the model to deliver the a more acceptable answer.


Richard Bell

Chips are thoroughly modeled before they are fabricated and then they are tested to verify that they behave as modeled. If chip design had as many parameters as the climate, we would still be using analog computers and hybrid computers (digital portion handles all of the patch cording necessary to program the analog portion, as well as handling the I/O of the analog portion).
The looming problem of chip testing is that a larger word size increases the number of test states by a factor of 2^(# of new data pins + # of other pins). Even before Y2K (I took a course on chip design in 1999 and the professor very graciously gave me a final mark of 50 for my help in making her a more effective lecturer and the very astute observation that I would not design chips after graduating), much research was put into finding heuristics to prune the number of test cases that were needed to verify that the chip behaved as the model.


Bernie- for liability reasons aircraft manufacturers go to great lengths to design, test, and validate their equipment, including autopilots and the whole aircraft design and performance.
That means extensive modeling and testing of the airplane’s aerodynamic design both the physical structure(is it strong enough, durable enough, failure modes-when some little part fails does the structure fail catastrophically or does it bend without collapsing?). They also do extensive testing to show that any computer software behaves as it was designed to do. They validate the programs by line by line analysis and tracing every possible path it could follow. The software is also written in a language(ADA, among) others) that makes it very difficult to make “blue screen of death” failures and that the compilers used don’t make errors in the machine code translation.
NO ONE, including the FAA accepts models as anything but corroborating evidence. The aerodynamic models are used to help predict performance, fuel economy, flight characteristics, etc. That is then verified by flight testing. No aircraft for civilian or military use gets accepted without flight testing.
Test pilots use models to give themselves some level of trust that the plane won’t come apart or do something wonky on the first flight. Then they make extensive tests gradually pushing the boundaries to ensure that the plane does fly as expected.


You mean tested against simulated environmental conditions. The computer chips themselves are not modeled but tested. The thing modeled is the environment that interacts with the computer chip.

Kenneth Mitchell

A “climate model” is EASILY falsifiable. Make your prediction – and in a year or a decade, if the circumstances predicted by the model haven’t come to pass, then the model is FALSE. WRONG.
If it isn’t falsifiable – testable against some experimental result – then it isn’t “science”.

The models are easily falsifiable:comment image

Russ S

You don’t even need the observed temperature in the graph. The projected temperature outputs don’t even match each other. At the very least, that should disqualify all but one.
I’d hate to be analyzing an electronic circuit and one simulation program had a significantly different result from another one.
Climate models are like stock market models. Completely bogus!

Surely this is valid. Yet why does the statement persist?
The models, possibly even all, are shown to be in error, that is, they are false.
Of course the observations themselves would need to scientifically valid themselves.
Not sure where that takes us?


Quite right Eric. All of those models cannot match the observed temperatures (but one is close). Wait a minute – lets adjust the temperatures to fit the models! Grant funding saved.

DaveR I can see you’re making an honest point, but what the graph shows is true, despite differences in the scales. The graph makes it clear that CO2 emissions have gone up dramatically over the last two decades, but temperatures have been flat. The scales need to be different so that fact is totally clear.

DavidR, sometimes I misclick in this maze of comments. As I said below … in the wrong spot … my comment above was meant for daved46. Sorry. (Hopefully this comment goes in right spot.)

I should add .. they posit that CO2 causes climate warming, but the historical record does NOT show that.
Start with the last two decades. CO2 has been emitted at record rates, but … we got nothing as far a temperature reaction:comment image
And look at the entire 20th century. The rate of mild warming in the early 20th century was virtually identical to the rate in the late 20th century: so, again we see no signal from CO2. And, go back hundreds of thousands of years and we see from ice core data, again, zero evidence that CO2 has been affecting climate temperatures. We know (but the people don’t know) that the evidence is that temperature changes is causing changes in CO2, but not the reverse:comment image
And, further still, go back hundreds of millions of years and the lack of a causal correlation between CO2 and temperatures is even more stark. Their theory is a joke. And so the leftist plan to downsize all the economies of the world (just the Western economies actually…) to curb non-existent warming is even worse than a joke: it’s insidious.

Dr Bob

Can we have an attribution for the graph please? I’d like to be able to use it with confidence.

@Dr. Bob
The image is from the show The Great Global Warming Swindle.
That very image is used is this key 4 minute excerpt from that show:


Eric, I do wish we skeptics would stop using graphs like the first one you have. When you have two separate sorts of data in a graph each with a basically linear trend, you can always make them either match or look totally different. In this case by making the temperature scale from -.5 to +.5 the two would about the same. Not that it would mean anything, but it would at least be less confusing to many readers.


That upper graph is a little contrived and misleading — the vertical scales are not commensurate. Scaling one or the other gives a very different picture, and will show high correlation.


And they seem so puzzled and taken aback when some of us do not feel so keen about betting the house on them!


Hmm, this should have posted in reply to this from Russ:
“Climate models are like stock market models.”
And they seem so puzzled and taken aback when some of us do not feel so keen about betting the house on them!

Sorry, DavidR, that reply was meant for daved46 below!


Climate models are easily falsifiable. All you need to do is compare their predictions observable reality.

Curious George

Climate models don’t make predictions any more. They make projections. Projections are not falsifiable, an important legal point.

Mother Nature does not do Man’s laws . . . either.

John M. Ware

That sounds like a weasel-word substitution to me. Perhaps a projection can’t be legally assailed, but it can surely be checked: Five years ago model X projected that the temperature in Valley Falls* would increase by an average of .2 degrees C per decade. I shall check with the Valley Falls weather office and find out whether this year’s temperature is .1 degree warmer than it was when the projection was made. If so, we have one datum of support for the model; if not, we don’t. Obviously, on a long time-scale like a century, five years isn’t a reliable gauge; and yet, even after five years we can see whether the data conform to the projection.
*Valley Falls is a fictional town inhabited by Chip Hilton and his buddies, written about by Clair Bee.


Prediction, projection, forecast are synonyms.

M Seward

In similar vein,
According to Dr Lewis
“I’ve wanted to be a scientist since I was five years old.”
as quite distinct from ‘I’ve been fascinated by science since I was five years old’.
One is about the ego and one is about the curiosity of the mind.
It would seem that Dr Lewis is the epitome of a ‘climate scientist’ ( i.e. a ‘climate egotist’

Sweet Old Bob

A projection is the aftermath of a power puke …
…. and has a lot in common with “climate science “

seward, do you understand the
difference between a
“prediction” and a
If so, tell us the

the difference between a
“prediction” and a
a projection has no predictive power. It is simply a line drawn from known to unknown data.
a prediction on the other hand, may or may not have predictive power, and thus can be falsified.
climate projections by definition have no ability to predict the future, so there should be no expectation that future temperatures will match model projections. IPCC 2 recognized that climate prediction was impossible given current mathematics, and thus “prediction” was changed to “projection” for IPCC 3.
At that point, climate science became a pseudo science, because everyone involved still pretends that model projections have predictive power. But they do not. Otherwise they would be falsifiable and they would not be projections, they would be predictions.


SMC – As long as the ‘observations’ of reality are trustworthy. IMO the global scale of the observations (compiled from thousands of individual sources) needed to do a valid comparison with the output of any given GCM, and the degree of tampering that has apparently occurred with these sources by numerous parties, makes it unlikely that any ‘falsification’ would be widely endorsed. Is there an easier way?

what “tampering”?
you mean correcting for


Okay, just this once…but from now on you best be learning about looking stuff up.
Golly…the world must be a baffling place for you werdz challenged snowflakes.
tampering (present participle)
1.interfere with (something) in order to cause damage or make unauthorized alterations:
“someone tampered with the brakes on my car”
synonyms: interfere with · monkey around with · meddle with · tinker with
2.exert a secret or corrupt influence upon (someone).
synonyms: influence · get at · rig · manipulate · bribe · corrupt · bias · fix
Hey, now I am wondering, is a Cracker 345 more like a Triscuit, or a Wheat Thin?


crackers, don’t you mean inserting bias?


“Falsifiability is the idea that an assertion can be shown to be false by…… an observation,”
Too late, been there already……got the tshirt and mug


Cracker 345, presumably you are using a computer or smartphone.
Do you know how to use it to look stuff up?
Prove it by looking stuff up for yourself, such as how to use a thesaurus or a dictionary.

Menicholas: again, what
im asking seriously


*rolls the eyes*

meni – i didn’t think
you could do the difficult
work of proving your assertion.


And which assertion might that be?


In any case, everyone is tired of trolls who try to hijack threads by demanding that others provide links and verification of things that have been discussed for years and that anyone can look up.
Do your own homework.
If you think a single sentence from you can settle the question of the wildly fraudulent alterations to the historical databases…guess again.
How about instead you explain in detail the precise rational for every so-called correction, and the proof that each of these was exactly and precisely done to “correct” for unambiguous and verifiable biases.


crackers355 check in Australia where ongoing tampering has just been busted.

M Seward

@Crackers 345, I was stating my impressions of Dr Lewis’s written words. Why are you playing word games with terms neither she nor I used?
Reading some of you other posts you seem to have a bee in your bonnet about slights against Dr Lewis. Dr Lewis has, imo, either made a fool of herself despite being basically competent or simply exposed herself as a naive idiot with a childish understanding od science.
I am an engineer by profession and use models all the time for substantive purposes in engineering analysis. The idea that ‘models’ are not falsifiable’ is so ridiculous in my profession as to amount to criminal negligence. I suppose for scientists without the responsiblity of constructive use of their work and certifying such constructive use, it is mere negligence or simple incompetence.

SMC – yes. It seems that this person confuses falsifiability with actual falsification. For any scientific hypothesis or “model”, it should be straightforward to spell out a set of experimental observations that would falsify it. Of course, if the hypothesis is novel at all, one could not be certain of such an experimental outcome.


Wasn’t all that long ago that they said a 15 year period without rising temperatures would falsify the models. Until we got a 15 year period without rising temperatures.

Karl to the rescue.


Didn’t the IPCC itself declare something to the effect that the climate is a coupled, non-linear chaotic system that cannot be modelled? What am I missing here?

Pat Lane

Anything can be modelled. I have a model that says the temperature will double each day starting today. It’s 22 degrees Celsius today. To be really scientific, I’m using degrees Kelvin to make sure I’m ABSOLUTELY correct.
My Gosh! In just 10 days the temperature is going to be so high not even Excel can calculate it!
Quick, give me heaps of funding and complete political and economic control of the Earth. Before it’s too late! It’s worse than we previously thought! Don’t you care about the planet? Don’t you care about your children?

Pat Lane

Day 7 temperature: 1200000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 +- 0.000000000000001 degree Kelvin. To get back to Celsius, either add or subtract 273, I forget which.

Mike McMillan

Doggone zeroes ran off my monitor and onto the floor, all 705 of them. You might want to recheck your model’s coding to see if it truly reflects your underlying hypothesis.

Pat Lane

Should have been 158 zeros.
My scientific principle: “The model IS the hypothesis!”


To get back to Celsius, either add or subtract 273, I forget which.

At this point, it doesn’t really make much difference.


Long strings of characters without any spaces breaks the WordPress text formatting. Maybe it is dumb that it does not hyphenate, but there it is.


Pat, same as climate models, your model is broken.
22 C = 295 K
10 doublings gets 302,080. (yes, I checked)

My scientific principle: “The model IS the hypothesis!”

Very well put. The model is the product of all the assumptions and assertions that went into it.

22°C = 295.15K
× 1024 = 302,233.6 (temperature after ten doublings)
Pat, did you do ten squarings instead of ten doublings?
I found an online arbitrary precision calculator and calculated 295^1024 = a 2,530-digit number, approximately 1.2524E2529


“Doggone zeroes ran off my monitor and onto the floor, all 705 of them. ”
I adjusted the zoom to 33, and scrolled left…then you can see them all.

non-linear chaotic system that cannot be modelled
such a system can be modeled, however the results cannot be trusted to be reliable.
Mathematically, climate modelling is equivalent to the n-body problem in celestial mechanics. We can solve for 2 object in orbit around each other. However once the number of objects in orbit is 3 or more, our mathematics falls down, except for the simplest case where all the objects are in the same plane.
The basic problem is that the error term does not converge to 0, as it does for simple systems. The plus and minus errors do not average out to zero in the long term. Rather the error term grows in size in the long term, such that it quickly overwhelms the correct answer. Only at infinity does the error term average out to zero, which means you must wait infinite time for a meaningful result.comment image


“Science is complicated – and doesn’t always fit the simplified version we learn as children.” It did before politics took it over.


Climate scientists are desperate for their educated guesses to be accepted as science“…
Maybe but I think the situation a little more basic than the desire that these climate scientists with their dodgy models be accepted as science, its the fear of losing their place at the government financed trough where they feed…
Very High Probability Of Fraud By Government Climate Scientists

richard verney

His recent series of bite sized videos is very good.

“climate models are not falsifiable”
Maybe not but false statistics is noticeable

“Climate scientist Sophie Lewis acknowledges that climate models are not falsifiable” NOT?


In Ms. Lewis’ defense, sometimes science ISN’T falsifiable, at least not right away. It took years before Einstein’s theory of relativity could be tested, and many parts of quantum and string theory still can’t. The bleeding edge of science is often found just past the limits of our data collection and measuring capacity capacity.
Which isn’t to say that Climate Models ARN’T falsifiable. Just that we don’t yet have enough data to say for sure if they are. Even the Climate Faithful admit that Global Average Temperature is a difficult and noisy thing to try to measure. Even today we don’t really have a sufficient system (hence the alleged need for constant adjustments).
BUT. Every year that goes by without a clear increase in GAT does indeed add to the falsification of the models. Many already believe they have been falsified. and eventually even the most die-hard Climate ‘Scientist’ will have to accept that the models were wrong, though for some it might take 50 more years.

H. D. Hoese

This has bothered me for a long time and I cannot offhand find the paper which dates from 1977. It involved attempts to separate causes from effects in often difficult studies of population density “There is no need to test its (theory) validity against observations… .” This is a direct quote, but out of context and it has been a long time since I read the paper, which was maybe reasonably attempting to develop a methodology about a subject which still gives ecologists fits. Perhaps this has been developing somewhat along the line just suggested with difficult but elegant and reasonable sounding concepts that have pressures to produce solutions, such as in population models in fisheries. For that matter, maybe in a lot of subjects where we get the idea that the “best available science” is sufficient.


Just to be clear…string theory is no theory…it is made up ad hoc hypothesizing about something that will in all probability never be in any way testable.
And quantum mechanics has the problem of being impossible to understand in terms of cause and effect macroscopic logic…but it makes predictions and they are testable and repeatable and have been verified in many ways and many times.

Leo Smith

Some confusion.
A scientific theory must BE falsifiable and attempts to falsify it must have so far always failed.
In the case where it modify an existing theory, as in relativity, you have to find a case where a different result happens according to each theory. One theory will give a better answer than the other.
Climate change IS falsifiable and has been falsified completely.

Anders Valland

schitztree, the idea of a theory being falsifiable has nothing to do with the ability to test the theory, or the time it takes to test it. Einsteins hypotheses have always been falsifiable, and there are numerous ways to do it. Several predictions have been made based on the hypotheses, and these have been testet succesively as we have gained more knowledge and been able to devise methods for testing. But the fact that it has taken a long time to test some of these hyptheses has nothing to do with the falsifyability of the hyptheses themselves.
Einsteins predictions have been found to be true, thus strenghtening the hypotheses to the point of regarding them as theories. But there are still some possibilities to show he might have been wrong.
If you look at ideas posited about the singularity thought to be the initiator of our universe, any of it is beyond science since it is beyond the observable universe. Any ideas about what was before the Big Bang, and the Big Bang itself, is beyond science as it is beyond what is observable. Such ideas always suppose that ‘logic’ can be applied to determine what might have been, but there is no support for the notion that ‘logic’ existed at or before the Big Bang. There can never be.

In Ms. Lewis’ defense, sometimes science ISN’T falsifiable

According to Popper, is primarily logical rather than technical. A hypothesis is falsifiable if we can think of an experimental observation that would contradict this hypothesis. If the experiment in question is not currently feasible, that does not mean that the hypothesis is not falsifiable.
Consider the prediction that the sun will turn into a Red Giant in a couple of billion years. That prediction is logically falsifiable now, even though the corresponding observation will be feasible only in the far future.

Correction: According to Popper, the concept of falsifiability is logical rather than technical.

sometimes science ISN’T falsifiable
then it has no value, because there is no way to know if it true or false. and if you don’t know if something is true or false, you cannot use it in any meaningful way except to avoid it.
thus if climate models are not falsifiable, the only logical response is to not use them as the basis for any decision making, except for the decision not to use them.
as history shows, expert opinion is of no value in determining if something that is currently unknown is true of false. if anything, the evidence shows that expert opinion is in fact worse than the opinion of the average person in the street when it comes to judging true and false.

Curious George

Regarding NCAR Common Atmosphere Model, here is my experience:

Mark Cooper

Love the reference to the Graphic of Avg temp V pirate numbers in this article. Not only is the correlation spurious, but the raw data has been adjusted down… The fact is there are many thousands more pirates out there now than there were in the 17th Century…


Except we call them Wall Street Tycoons.


Re the pirate to temperature correlation, there is another pirate relationship that is a positive one: the number of Pastafarians correlates nicely with the number of pirates. I believe it is a one-to-one relationship starting with the pirate Mosley.


Aah, to be touched by His noodle appendage and thus be brought into the light!

Tom Halla

Huuh? I was taught that the nature of science is making descriptions of nature that are testable, and if one cannot come up with a way to test an assertion, it is not science yet. Sheri Lewis seem to be into politics and post-modernism, as there are ways of testing “climate science”. Her problem is that her favored models fail those tests.


A Blast From The Past
That one will fly right past so many here.
Therefor, for your enjoyment: Lambchop.

Lambchop is such a brat.

Gunga Din

Climate models are not “falsifiable”.
I played a PC game (Pacific General) where Japan defeated the USA.
Prove my computer-based claim is false.
You observed I’m not typing in Japanese?
Trust me. I should be.


Since it is known which climate model is correct, why is an ensemble of known incorrect models used?
Just pick the right one and stick with it. Running all those expensive wrong simulations on huge supercomputers requires enormous amounts of energy and generates tons of CO2.


August 10, 2017 at 4:33 pm
“Just pick the right one and stick with it.”
Actually it can be claimed that the “right” GCMs are picked up and selected, but for the wrong reason thus far.
The “realistic” ones are used for policy making, where the priority of these models, the priority of the experiment so to say, has shifted to policy making and politics….. WITH NOT MUCH REGARD FOR SCIENCE.
The “realistic” ones are the ones who do a realistic enough ppm CO2 trend in comparison with reality….the ones that give or result with an ~3C warming by the end of the century…..
From the outset, these GCMs do nullify the AGW-ACC hypothesis where the experiment shows the required condition to be observed in the case of a significant RF warming of atmosphere, which is what the anthropogenic climate change supposes to consist basically with…….the famous Tropical hot spot…..
And also as far as these GCMs concerned, the anthropogenic hypothesis stands falsified in account that the RF warming of atmosphere due to increase of CO2 concentration in a given anthropogenic effect should consist as a continuation of the correlation between the CO2 concentrations and temps,,,, with no departure or decoupling of one from the other…
From some angle, both conditions can be considered as nullifying and falsifying in the same time the AGW-ACC .

CAGW faith minority seems to be betting on even the blind chicken little finding the kernel. The bet outcome is already settled and the winner will be too after a few decades and Petadollars.


I am not sure, but in case that it may help……..falsification of any model in principle is the easiest process of all……..if it can not replicate is a false model…..any thing else apart from this as far as the results of models concern is the validation…..that is how I do address from my point of view the model falsification.
May not be right but so far it works for me fine……but you see, I am not a scientist…..can’t neither quantum or string things, or whatever in that realm.

Tim Neilson

“The problem with Sophie’s position is that fitting a model to past observations is not a test of whether the model is right; all fitting the model tells you is that you have found a way to fit the model.”
Correct. Ptolemaic astronomers were very good at adapting their model each time new data didn’t fit the existing version.

Peter Morris

Wow. She’s openly admitting that wishcasting and imagineering are what she thinks of as actual science. I’d be shocked if we hadn’t seen this a hundred times already.


She’s celebrating being a part of a movement that brings scientists into the age of the all-healing virtue-signal:
“After a few talks, I had to remove the photo from my PowerPoint presentation because each time I turned around to discuss it, it would make me teary. I felt so strongly that the year we were living was a chilling taste of our world to come.
Just outside of Sydney, tinderbox conditions occurred in early spring of 2013, following a dry, warm winter. Bushfires raged far too early in the season. I was frightened of a world 1°C hotter than now (regardless of what the equilibrium climate sensitivity turns out to be).
At public lectures and community events, people want to know that I am frightened about bushfires. They want to know that I am concerned about the vulnerability of our elderly to increasing summer heat stress. People want to know that, among everything else, I remain optimistic about our collective resilience and desire to care for each other. “


I think it fits better into the “I want to be your friend” behavior where you compromise your principles to belong with the “in” crowd.


CACA adherents are corrupting all science, with their postmodern philosophy of science, based upon consensus rather than the time-honored process of falsification.
It stands the scientific method on its head. Instead of “a single fact can show me wrong” and “the belief in the ignorance of experts”, it’s “trust us, we’re experts and 97% of us can’t be wrong”.

Forrest, true, science as a process exists independent of the would-be ‘scientist’ who either follows it or pretends to follow it.
The damage is in the redefinition of the process such that what was verboten, is now accepted by the community of ‘scientists’ but still described as Science(!). A revolution with in the form of the law. Keep the same edifice but completely re plumb the internals and sell it as the same thing. As long as the grants keep rolling in, from politicians looking for a veneer to cover the latest power grab. Squash anyone who dares question it as a knuckle dragging ignorant denier. Doubleplussungood!
Leftism as it’s finest.


Another climate scientist who doesn’t know what science is, and can’t stand having his ideas challenged.


“A test of falsifiability requires a model test or climate observation that shows global warming caused by increased human-produced greenhouse gases is untrue.”
She even gets this statement wrong. What you have to do is devise a test capable of falsifying the model projections of how much of warming is natural and how much is from CO2. Now, in theory, if someone came up with a computer model that went on to correctly peg global and regional average temperature swings, precipitation swings, etc. over several decades following the model making the projections, then this could be accepted as indirect proof that the model is capable of correctly replicating both natural climate change and CO2-driven climate change. So there is indeed a possible way of falsifying the models, and to the best of my knowledge the models have in actuality been falsified under this test since the models have never been able to reliably perform as an actual predictive tool.
Also problematical is the inference of some “model test” showing CO2-driven global warming to be false. The first issue here is analytical – how could any purely mathematical exercise qualify as a “test” at all? The practical problem with this is that no one ever uses models to try to falsify CO2-driven change – they build their models to implement that change. What would happen if climate scientists devoted an equivalent amount of time and effort into developing computer models that simulated a world in which rising CO2 concentrations had little impact, either through strong negative feedback or dramatically reduced climate sensitivity, and then compared the two series of models to see which had more predictive capability.
We’ll never know because most climate scientists would be afraid of that comparison and no one would be willing to fund it.
She also gets this statement wrong: “Climate models are important and complex tools for understanding the climate system.” Climate models are, at best, tools that flesh out a hypothesis of how the climate might behave. A climate model only does what its programmed to do; a person has to give the computer its understanding of the climate system. The computer does not, and cannot add to your understanding of the real world. That can only happen the old fashioned way.

Chuck Dolci

The quote by Sophie Lewis brought to mind two thoughts. 1. The recent stink raised by the Google engineer’s memo in which he pointed out the “differences” between males and females. Not saying women scientists are inferior (Judith Curry, for one, shows us otherwise), but perhaps more influenced by emotion. The quote just reeks of emotion ruling over reason. 2. Sophie Lewis is relatively young. I wonder if her attitudes about science are influenced by the recent tendency of young people to believe their are “subjective facts”, i.e. It is true because I believe it to be true. A man thinks he is a woman, therefore he is a women and we must accept the fact that he is a woman.
Just saying, those ideas came to mind reading the quote.

Russell Johnson

The statements by Sophie Lewis are not remotely related to “science”, there are zero “subjective facts”.
The desire to be relevant certainly clouds the judgement of many young minds. If Sophie’s attitude is typical, all humanity is at risk when critical decisions must be made. Totally Sad!!!!

Roger Knights

A test of falsifiability requires a model test or climate observation that shows global warming caused by increased human-produced greenhouse gases is untrue. It is difficult to propose a test of climate models in advance that is falsifiable.

How about the discovery of a basic math error in all those models, huh?
/ Monckton in a month or two

richard verney

Even assuming that there are no significant errors in his paper, you are being overly optimistic regarding the peer review process.
What are the odds of his paper passing before the deadline for AR6?
You cannot seriously think that the IPCC would wish to address problems with models and their projections.

Roger Knights

They may not wish to address it, but it will be awkward for them if they don’t.
As for beating the deadline, Monckton has indicated that it will be published (probably by a Chinese journal) within a few months.


After reading her paper, I disagree that this is cognitive dissonance. She’s attacking the scientific method as being outdated, and advancing a politically correct perversion of the scientific method.



There was once a model – one that exactly matched all previous observations, and exactly matched all subsequent observations for well-nigh 1,500 years. This fantastically accurate model was made by a Greek named Ptolemy.

Paul Penrose

“All models are wrong; some are useful”. And even if a model is useful, that is only true for limited purposes. For example, the global climate models are useful for exploring some of the many processes at work in the climate system. That does not make the suitable for predicting/projecting the global temperature 100 years from now.

Clyde Spencer

I think the proper role of models is to provide insight on the behavior of complex dynamic systems. Those insights should then be used to propose experiments to verify or reject the hypothesis on how the system behaves when perturbed by exogenous forcings. Instead, today’s modelers want us to accept, unquestioningly, what they claim the behavior of the system is, which is supported only by their belief that they understand everything there is to know and all the variables and feedback loops have been characterized correctly. That is asking a lot!

D. J. Hawkins

I think given the current scale of the models, our understanding of the underlying physics, and the computational limits of our current “big iron”, that your conclusion is still a bridge too far. IIRC, rgb[at]duke estimated that we are about 6 orders of magnitude away from having models fine enough to capture things like thunderstorms.


eric, models cannot “predict” the future — they can
only project it, based on assumptions about
future CO2 emissions, solar changes, volcanic
eruptions, etc.
this is pretty
obvious and everyone should well
it by now.

I Came I Saw I Left


Clyde Spencer

Based on what are largely unsupportable assumptions and unstated assumptions. Anticipating that you will ask me to support that claim, I offer up how energy exchanges in clouds are handled in GCMs, with parameterizations instead of actual modeling. The parameterizations are assumptions about what happens in clouds at a spatial scale different from what everything else is modeled.


What everyone understand very well by now is the vacuous opinionating of warmistas, and their belief that no one can see through their utterly transparent logical fallacies, squishy accountability for a endless litany of failed predictions, and a nearly incomprehensible ignorance of topics and areas of knowledge that have always ben considered within the purview of their chosen field of inquiry.
As far as anyone can discern, warmistas know more about Saul Alinsky-style character assignation than about Earth history, have precisely zero understanding of what constitutes actual science, are accountable to only those apportioning their next grant, and think believing something is the same as knowing it.

richard verney

Not overlooking sensitivity to forcings, feedbacks etc. etc.
Nothing is sufficiently known or understood to be of value. It is assumptions piled upon assumptions piled upon assumptions etc tuned to dodgy data (the adjusted temperature data etc)

Paul Penrose

To the average person, the difference between a prediction and a projection are about nil. This includes our politicians who have been asked to reorder our entire civilization based largely on the outputs of these models.

John Endicott

“It turns out that my work now as a climate scientist doesn’t quite gel with the way we typically talk about science and how science works”
In that case, with all due respect, what you are doing doesn’t qualify as science.


When you realize that what you’re doing doesn’t actually count as Science, but still want to be called a Scientist.
I mean really, even the psychologists TRY to run experiments to test their theories.

Sophie – Yes they are falsifiable in the real world.
However in your dream world where you mix with other other distorted individuals, that may be the case.
Sophie has just given us a rare glimpse of how these folks think, or is it don’t think. They consider themselves to be above anyone else, the special chosen ones who have been tasked with saving the planet. Therefore they can make their own rules. And they do.
All kneel, bow your head and pay homage to the chosen ones. And think yourselves lucky that we have them guiding us. You will recognize them among us, they have arrogant blank looks on their faces, and speak in a soft monotone.


What ever it takes to rake in grant money, that is what they’re prepared to do. Redefining what science is though, is not even clever let alone tolerable. Her motto should be “Science is hard – let’s do it wrong.”

the title is misleading
The title says “climate science does not have to be falsifiable”
Climate science is falsifiable.
Sophie’s comments are about climate MODELS. climate models are not climate science.
They are not even climate theory.
They are tools and methods used to understand and evaluate
“How do you falsify a climate model? Australian National University Climate scientist Sophie Lewis acknowledges that climate models are not falsifiable – yet claims we should trust them anyway.”
This misquotes her.
The simple fact is no model and no theory is strictly speaking falsifiable. Even Popper knew this.
The really funny thing is skeptics argue two things.
1. the models HAVE BEEN falsified
2. its impossible to falsify them
no joke. its too funny.


The really funny thing is alarmists fail to address two things.
1. the models HAVE never BEEN validated.
2. its impossible to validate them
no joke. its too funny.
(There . . fixed it for you.)

Clyde Spencer

Sophie is not being misquoted. She said, ” It is difficult to propose a test of climate models in advance that is falsifiable…Climate models are carefully developed and evaluated based on their ability to accurately reproduce observed climate trends and processes. This is why climatologists have confidence in them as scientific tools, NOT because of ideas around falsifiability.” It seems to me that she is arguing that climate models don’t need to be falsifiable. So, whether they are “difficult” or impossible to falsify, she is turning her back on the requirement.
On the other hand, as David so amply demonstrates below, most skeptics actually think that the extant models have been falsified. You confound another complaint of skeptics with this. That is, forecasts, predictions, scenarios, whatever, are so far in the future that few if any of us will be around to see if they come true. In that sense, they are unfalsifiable. Despite near-term inability of the models to track reality, alarmist modelers adhere to their 100-year unfalsifiable predictions.


Well, no one can accuse them of not eventually having learned to make their doomsday no sooner than the longest time they can hope to be alive.
And no one can accuse them of being teachable either.
It takes a very dense and particularly impenetrable cranium to be always wrong and yet supremely confident.


So if climate models are not science, why do you believe we should use their output to inform political decisions about our economy?
No joke. It’s not funny. You are admitting that we should not use models as they are not science.


Once again Steve you blame skeptics in general for what a true believer specifically states.
How about holding true believers to even a fraction of the ball busting you hold skeptics to?


I’ve got this revolutionary weight loss herbal tea I’m selling.
It’s true that you may have seen a slow down in weight loss over the last ~ 20 years but the underlying benefits are still there (basic biology) and as soon as other confounding factors abate, your weight loss will continue.

Falsified…comment image
Falsified…comment image
Falsified….comment image
Falsified…comment image
Falsified…comment image
Falsified…comment image
Falsified…comment image
Falsified…comment image
Falsified…comment image
Ms. Lewis is correct. The climate models can’t be falsified… because they already have been falsified.

Now David don’t be a naughty boy.
If you don’t behave we will have to send you to one of Al Gores special educational boot camps to correct your way of thinking.
Its all a matter of perspective.
They just need to make their lines much thicker on the charts, then they will be in touch with reality.
See how easy it can be

Clay Sanborn

I’m sure everyone has noticed that alarmists always try to make their public case with emotional appeals. For me, show me the proof; any proof. Alarmists are like those wackos that want us to believe intelligent extraterrestrial life forms are visiting planet Earth. Their “evidence” is always only fuzzy pictures and emotion fed conspiracy theories. The fact that there are multiple squadrons of credible scientists and institutions that say CAGW ain’t happening (proofs data not there) should raise at least one alarmist’s eyebrows. I mean, come on! BTW – as I understand it, out gov’t has spent over $100 Billion on globul warming climate studies; i.e. alarmist scientist. Look at the list of bad things that are blamed on Globul Warming. It’s a racket!

Javert Chip

Some physicists (Stanford U) said string theory was jus t”so beautiful, it didn’t need to be tested” until CERN blew SuperSymmetry out of the water.
Now were gonna go back & start doing real (ie falsifiable) physics, the way Newton, Einstein & .
Feynman intended,
It’s about time, we’ve wasted 30 years on an goof-ball intentionaly-designed-not-to-be-testable theory (we’re repeating this with “climate science”).


You’re wrong to be down on string theory.
For one – it is not being used to justify a global Marxist revolution like CAGW is.
The original discovery by Veneziano that topological approaches derived from Euler improved the solutions to strong force data, is still valid.
Just because the Plank length is small and related energies are very high, does not mean it can’t be true.
It is an epistemological failure to imagine that it does.

ptolemy2 wrote, “You’re wrong to be down on string theory.
For one – it is not being used to justify a global Marxist revolution like CAGW is.”
and how is that?
if your electricity is generated
from renewable
sources, will
you not plug your
tv into the same


Is that inane observance intended as a serious retort?
You outdo yourself, Wheat Thin.

javert: SS has
certainly not been
where did you ever
get that notion?


comment image?oh=374e1fddc53ddd660fbdedc1b38013df&oe=5A2E3EFF


Climate science has a falsifiability problem?
Solution – let’s insult Karl Popper by calling him, and the idea that science needs to be falsifiable, simple and childish.
There – that was easy!


And blame anyone who points out problems for the problems.

Pamela Gray

By fitting the models to past known weather pattern variations (which were set to a warming cycle by mother nature) and their outcome, climate scientists assume those same weather pattern variations will be at work in the future. That is an assumption clearly not supported by the archives. Weather patterns are a bit of a random walk at the peak of an interstadial, which is why our climate has been relatively stable. And at the peak of an interstadial, these weather patterns are keeping us generally warm (to be expected). The hysteria is over a tiny bump up falsely thought to be related to CO2. In the 70’s the scare was about a tiny bump down, again falsely thought to be related to CO2. In my opinion, a confounding factor monster bit them in the arse and now we have models that don’t fit the current weather pattern variation.

Curious George

Science?? They are laughing all the way to the bank.

Pamela Gray commented
“In the 70’s the scare was about a tiny bump down, again falsely thought to be related to CO2”
So wrong.
It was due to SO2 pollution following WW2. Not CO2.
Who attributed it to CO2? Name one person.

Patrick MJD

“crackers345 August 10, 2017 at 6:3 pm
Who attributed it to CO2? Name one person.”
There are some here…

richard verney

I agree that the cooling was not being attributed to CO2.
However, the accepted position, and this was based upon known principles of physics which are said to underpin the AGW theory, was that Climate Sensitivity to CO2 was low.
So for example NASA/GISS in the 1970s assessed the Climate Sensitivity to CO2 was such that an 8 fold increase in CO2 would lead to an increase in temperature of less than 2 degC. See the Schneider et al paper published in Science.

Pamela Gray commented:
“In the 70’s the scare was about a tiny bump down, again falsely thought to be related to CO2”
Crackers wrote:
“So wrong. It was due to SO2 pollution following WW2. Not CO2.”
Allan wrote:
The global cooling that occurred from ~1940-1975 was FALSELY ATTRIBUTED TO INCREASING SO2. Aerosol data was fabricated to force-hindcast the climate models to fit this cooling.
Here is supporting correspondence with Douglas Hoyt dating back to 2006:
In fact, the ~35-year global cooling period that commenced in ~1940 adequately falsifies the hypothesis that increasing atmospheric CO2 is a significant driver of global warming. The CAGW hypo is further falsified by the current ~20-year “Pause” in global temperatures.
That is why the warmists have more recently been falsifying the temperature data records to minimize the ~35-year cooling period and increase their alleged warming during the Pause.comment image
Since 1940 there has been ~22 years of positive correlation of temperature with CO2, and ~55 years of negative or ~zero correlation. The global warming hypo is contradicted by a full-Earth-scale test since 1940. CO2 is NOT a significant driver of global warming.
Regards, Allan

Lars P.

Thank you Allan for posting this.
These changing historical values are the reason why I do not look at ‘climate scientists‘ as ‘scientists’. This is no science.

Global warming alarmism is unscientific nonsense, supported by scoundrels and imbeciles.
George Carlin said it better. [Warning – language] 🙂

Old CS Prof

There is a general problem with creating a model in the first place. Predicting conditions in the next century based on the previous century is similar to predicting the conditions tomorrow based on the last 0.00912 seconds. There simply isn’t enough data to form a model or statically falsify it.

Juan Slayton

Entertained by, and wondering about the source of, your graph. Good parody, and easily falsifiable just by reference to the current ’17.’ Lot more than that just off Somalia. Calling Willis…. : > )


Why are my posts with innocuous language and no links going into moderation?
Why am I on the naughty step?


OK I get it – I mis-spelled my email address – my bad!
Reply: yup and I just noticed all this during cleanup~ctm

Pat Frank

Sophie Lewis, “It is difficult to propose a test of climate models in advance that is falsifiable.
No it’s not. It is instead difficult to find climate modelers competent to evaluate their own models.

The problem with the models is that they assume that “greenhouse ” gasses play a role in climate change. They don’t!
For an alternative model, Google “Climate Change Deciphered”
This model has been falsified (that is, empirically tested) multiple times, and has been validated each time.


Using Sophie Lewis’s logic, I can develop a climate model based on nothing more than a random number generator. I then ask others to trust the output.

Gary Pearse

Sheri wants us to trust climate models. Do we trust all of them? Do we choose our favorite color? Is Monckton et al’s ‘Bode’ based equation the one- it certainly is closer at following observations. Guy Callendar in the 1930s designed one that fits pretty well and having done it 80yrs ago makes his the strongest of them all. Will Callendar’s survive the downturn in temperatures, no it won’t, but we certainly would have had one that worked for 80 yrs eventhough it isn’t correct if CO2 keeps going up and temperature declines.
Ya see Sheri you were very badly taught and you clearly didn’t argue about it. You accepted it like Moses accepted the tablets with the 10 commandments. I’m probably wrong but I have an inkling there are 10 commandments in climate science, too.
A big thing that separates climate studies from science is when its delivered in an indestructible whole, like the Torah, or Koran or Bible… There can be no learning moments. When weather departs from holy climate, this should be a teachable moment. Don’t bend the pause back up to fit. It’s a missing factor that is being signaled. This is more sacred than being guided by your climatechism.

If you don’t get to test the underlying assumptions and programming of a climate model, climate science in general, then,
… You are simply being asked to believe.
Which is all this movement is based on. Believing, Not questioning. Politically correct shaming.
The exact opposite of reaching the correct answer. If you like not knowing whether your answer is right or wrong, not testing is the way to go. Climate science doesn’t care if their theory is wrong,
Is the global warming theory right or wrong. Just keep asking that question everywhere you can because eventually people are forced into a corner of actually thinking about it themselves.



Randy in Ridgecrest

Eating your cake, and somehow having it too


Me thinks Ms. Lewis is rather naive on the politics of science.

John of Cloverdale, WA, Australia

Models used by Australian Climate Scientists can’t even predict the weather for a month.
“The World’s Best Practice climate models predicted Australia would be hotter than normal in September, instead the maximum temperature anomaly was 1 to 5 degrees below average across most of Australia.” image

climate models
don’t predict weather, they
“project” climate.
how well does your
climate model so?

Patrick MJD

And climate is the average of 30 years of weather, models based on assumption as you make clear in another post. So, at least you admit models are, largely, based on assumption. I will ignore that projection if it cannot be based on observation.


There are billions of reasons to “believe” in GCMs.
Heck, the US federal government alone gave us over 29 billion reasons every year for many years running.


Pure sophistry.

“Climate models don’;t predict weather, they “project” climate: How well does your climate model do?”
To within .02 deg. C, or less, through years 1975 – 2011

Rick C PE

I would seriously like to see someone rerun the CMIP5 model ensemble with the ECS set to zero. Would they all project global temperature to flat-line? Maybe try ECS at 0.5 and 1.0 too. Is there an ECS value that would correspond to the satellite and balloon observations over the last 20 years? Would the models still ‘project’ warming if ECS is set to zero? Has this already been done and if so does anyone know where to find the results? If it hasn’t, why not?

Rick C PE commented
“I would seriously like to see someone rerun the CMIP5 model ensemble with the ECS set to zero. Would they all project global temperature to flat-line?”
“What’s Really Warming the World,” Bloomberg Business, 6/24/15


Is it UHI? Don’t know, they didn’t plot that one. Why not ask NASA
Is it due to over the top adjustments? Don’t know, they didn’t plot that one either. Why not ask Tony Heller.
Is it due to the PDO? Don’t know, they didn’t plot that one either.
Pretty pathetic on Bloombergs part and tells me they are just proselytes of a new found religion. Not to be trusted.

Rick C PE

Interesting, but the article says the graphs are from a 2012 study and show results up to 2005. Was the model used identical to the version used in the IPCC projections? Might it have been revised a few times more recently – maybe updated parameterizations to better tune to the historical record? If you start with the creation of a model based on the premise that CO2 is the primary driver of temperature you will end up with model outputs that show a primary response to CO2. I am neither surprised nor convinced.


They do provide very clear evidence of one phenomenon…circular reason that begins with the desired conclusion will never give you a result that makes you think twice.