Guest Post by Willis Eschenbach
Eric Worrell posted an interesting article wherein a climate “scientist” says that falsifiability is not an integral part of science … now that’s bizarre madness to me, but here’s what she says:
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.
For some time now, I’ve said that a computer model is merely a solid incarnation of the beliefs, theories, and misconceptions of the programmers. However, there is a lovely new paper called The Effect of Fossil Fuel Emissions on Sea Level Rise: An Exploratory Study in which I found a curious statement. The paper deserves reading on its own merits, but there was one sentence in it which struck me as a natural extension of what I have been saying, but one which I’d never considered.
The author, Jamal Munshi, who it turns out works at my alma mater about 45 minutes from where I live, first described the findings of other scientists regarding sea level acceleration. He then says:
This work is a critical evaluation of these findings. Three weaknesses in this line of empirical research are noted.
First, the use of climate models interferes with the validity of the empirical test because models are an expression of theory and their use compromises the independence of the empirical test of theory from the theory itself.
Secondly, correlations between cumulative SLR and cumulative emissions do not serve as empirical evidence because correlations between cumulative values of time series data are spurious (Munshi, 2017).
And third, the usually held belief that acceleration in SLR, in and of itself, serves as evidence of its anthropogenic cause is a form of circular reasoning because it assumes that acceleration is unnatural.
Now, each of these is indeed a devastating critique of the state of the science regarding sea level acceleration. However, I was particularly struck by the first one, viz:
… the use of climate models interferes with the validity of the empirical test because models are an expression of theory and their use compromises the independence of the empirical test of theory from the theory itself.
Indeed. The models are an expression of the theory that CO2 causes warming. As a result, they are less than useful in testing that same theory.
Now, the scientist quoted by Eric Worrell above says that scientists believe the models because they “accurately reproduce climate trends and processes”. However, I see very little evidence of that. In the event, they have wildly overestimated the changes in temperature since the start of this century. Yes, they can reproduce the historical record, if you squint at it in the dusk with the light behind it … but that’s because they’ve been evolutionarily trained to do that—the ones that couldn’t reproduce the past died on the cutting room floor. However, for anything else, like say rainfall and temperature at various locations, they perform very poorly.
Finally, I’ve shown that the modeled global temperature output can be emulated to a very high degree of accuracy by a simple lagging and rescaling of the inputs … despite their complexity, their output is a simple function of their input.
So … since:
• we can’t trust the models because their predictions suck, and
• we can emulate their temperature output with a simple function of their input forcing, and
• they are an expression of the CO2 theory so they are less than useful in testing that theory …
… then … just what is it that are they good for?
Yes, I’m aware that all models are wrong, but some models are useful … however, are climate models useful? And if so, just what are these models useful for?
I’ll leave it there for y’all to take forwards. I’m reluctant to say anything further, ’cause I know that every word I write increases the odds that some charming fellow like 1sky1 or Mosh will come along to tell me in very unpleasant terms that I’m doing it wrong because I’m so dumb, and then they will flat-out refuse to demonstrate how to do it right.
Most days that’s not a problem, but it’s after midnight here, the stars are out, and my blood pressure is just fine, so I’ll let someone else have that fun …
My regards to everyone, commenters and lurkers, even 1sky1 and Mosh, I wish you all only the best,
w.
My Usual Request: Misunderstandings start easily and can last forever. I politely request that commenters QUOTE THE EXACT WORDS YOU DISAGREE WITH, so we can all understand your objection.
My Second Request: Please do not stop after merely claiming I’m using the wrong dataset or the wrong method. I may well be wrong, but such observations are not meaningful until you add a link to the proper dataset or an explanation of the right method.
“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.”
By this statement the “scientist” has essentially admitted incompetence for herself and all like-minded colleagues. They might as well admit that they are working in another field that is not part of what we recognize as science. Perhaps “climastrology” fits the bill.
I can only re-iterate my earlier comment wrt models: Models are great for interpolating between observed and matched data; they’re results must be increasingly suspect as they extrapolate beyond the observed data…
“I can only re-iterate my earlier comment wrt models: Models are great for interpolating between observed and matched data; they’re results must be increasingly suspect as they extrapolate beyond the observed data…”
The same can be said for polynomial data fits and similar techniques. So, we’re saying that GCMs are a costly and complex way to generate dubious results that could be obtained with a lot less effort using alternate approaches?
Excellent summation Willis!
Many kudos to Jamal for framing climate models succinctly!
This issue goes to why I am always commenting that any graph of climate model results must show the date the model was run as a bright vertical line so that the reader can distinguish between hind-casting (which any model can do) and forecasting.
“However, there is a lovely new paper called The Effect of Fossil Fuel Emissions on Sea Level Rise: An Exploratory Study ”
It’s an interesting paper. Since only Brest amongst their 17 data sets shows sea level to be correlated with emissions, I reckon that science has proved that France is doomed. Should we tell the French or let them live in happy ignorance?
“For some time now, I’ve said that a computer model is merely a solid incarnation of the beliefs, theories, and misconceptions of the programmers.”
True enough. Plus which, unless the model is trivial, the code is likely to contain unintentionally flawed representations of those beliefs, theories, and misconceptions — i.e. “bugs” The bugs which cause whacky results will almost certainly be tracked down and fixed. Bugs with more subtle effects very likely will not. The closer the models plus bugs match the anticipated results, the less likely the bugs are to be identified and exorcised.
It has been claimed that “all non-trivial computer programs have bugs.” The extermination of the bugs becomes more problematic with increasing complexity of the program, and, as you point out, less likely to be even identified as the output more closely matches what is expected.
“all non-trivial computer programs have bugs.”
Except mine. My programs have no bugs.
And I don’t need to roll down the windows when I phart either. Ask anyone. My dog has a problem with that, but I don’t.
Afterthought (not that anyone will read it). I think there is probably something somewhat akin to confirmation bias in play here. (Minor) bugs that cause results to be what the modeler expects/hopes for are possibly somewhat less likely to be found and fixed than similar bugs that don’t support the expected result. Given the obvious warmist bias in climate “science” that might result in the models running a little hotter than they might in an ideologically neutral environment.
Don K,
I read it and I agree.
I’m always amused by the argument that a climate model is reliable because it fits past data. If a model that fit well with past data was a reliable predictor of the future behavior of the system, nobody would ever lose money in the stock market.
You guys so so bad…in good sense though……:)
cheers
You ask a good question: “then … just what is it that are they good for?”
The answer might be found in a TED talk given by Gavin A. Schmidt. He argued that the models were “artful”. As art they are mere decoration open to the interpretation of the viewer. Some people want to find deep truths in them, but that is only a reflection of that persons own thoughts.
Richmond,
Some have complained that they can’t define art, but they know it when they see it. I’m afraid that I fail to see the “art” in the extant GCMs.
OK, perhaps the only art is in the mind of Gavin Schmidt. On the other hand, as artful deception, the GCMs do seem to work on many people. This is like the art of all messaging, some advertising campaigns work better than others. Some zombie ideas resist all efforts to kill them off. Just look at the alleged consensus of scientists that never existed. You should be congratulated on failing to see wonderful clothing that are all imaginary and only visible to the true believer in the GCMs
Even that admission is disingenuous as the models are waved around as the explicit description of the future and the rationale for a war on modern civilization. The proponents of this assault may be largely unable to see that the economic commotion they are instigating would be disastrous, but that is because they are Leftist by inclination and the history of Leftist government is unforeseen economic disaster. They usually blame this on Capitalist enemies instead of examining the realities of human economic interaction.
The Socialist ideal is not achievable and not even desirable as a degree of contention and competition is essential for human civilization to progress. When Capitalism raises hundreds of million out of poverty the notion of throwing out the baby with the bathwater becomes the critical consideration. Eco-Socialism is power mad immorality!
And that is why Gavin Schmidt refuses to debate anyone.
Steven Mosher August 11, 2017 at 2:04 am Edit
Mosh, thanks for your reply. First, you say:
Sorry, but that is not true at all. You mistake a change in forcing for a change in temperature. That is a bridge way, way too far. What we do know is that GHGs cause increased forcing. Period.
However, given that the climate RESPONDS to changes in temperature; and given that tropical cloud albedo is highly correlated with temperature; and given that the increased forcing from a doubling of CO2 would be totally counteracted by a 1% change in albedo; and given that thunderstorms are also correlated with temperature and cool the surface in a host of ways; given all that and more, we have no reason to assume a priori that a change in GHG forcing will perforce change the temperature, and we have a heap of evidence to show that that may NOT change it.
On the most basic level, I know of no other complex natural system where the output is a lagged linear function of the input as climate models claim. Complex systems are … well, they’re not that simple, which is why they are called “complex”.
As one of many examples of such evidence that the prevailing theory is incorrect, if forcing change is linearly related to temperature change as the theory falsely claims, volcanoes would have a huge effect on the global temperature … but as I have shown repeatedly, if you don’t know when the volcanoes occurred it is NOT POSSIBLE to identify them in the global temperature datasets.
Next, consider: almost every model in the CMIP5 dataset used a different set of forcings as input … but all of them are able to reproduce the historical temperatures … how can that be true if a) the models are based on “physical first principles” as is often claimed and b) temperature changes are linearly related to forcing changes as the prevailing theory falsely claims?
Next, you say:
Yep. All you need to do is choose which aerosol dataset and which methane dataset to use, and whether or not to include HFCs, and then tune your model, and Dang … perhaps that impresses you. For me, that’s merely evidence that the “model of the planet” that you started with has serious problems, and that you are just messing with tunable parameters.
Why is it that no climate model is ever trained on half of the historical temperature record, and then tested to see if it works on the other half of the historical record? Every other scientific model is tested that way, but noooo, not climate models. The world wonders …
Moving on, you said of my words as follows:
OK, you want to get picky, fine. The theory is that CO2 and other forcings cause warming. Since “internal variability” is assumed to cancel out in all the climate models, I’m not sure why that is relevant, but sure, toss that in as well.
Happy now? It doesn’t change the underlying argument by one whit, but if it makes you feel better, fine.
However, all of the models have different external forcings and different “internal variability” (whatever that might mean) … but they all hindcast the past equally well. How is this not a huge problem in your world?
The problem is that because they ONLY contain external forcings and internal variability, the models all lack the thermoregulatory mechanisms that I have provided heaps of evidence for—cumulus clouds, thunderstorms, dust devils, the El Nino/La Nina pump, squall lines, and all the rest. Many of these are not included because they are “sub-gridscale”, smaller than even the smallest of the climate grids in the most detailed models.
So you are trying to model the climate while leaving out a large range of crucial phenomena because they are too small … and you see no problem with that.
Truly, model enthusiasts like you live on another planet. It’s called “ModelWorld”, and it has one bizarre characteristic. Unlike the real world, it is linear, with the claim that changes in temperature are a linear function of changes in temperature.
And that wouldn’t be a problem, but over and over you guys keep insisting that ModelWorld is enough like the real world to serve as a valid proxy for the real world in a variety of calculations.
Sadly … it’s not, and it doesn’t.
w.
Willis,
+1
Thanks, Willis. Spot-on, as usual.
At the risk of sounding pedantic, that’s a physically incorrect statement. Since GHGs don’t SUPPLY any energy to the planetary system, but merely increase the opacity of the atmosphere to LW radiation, the total system forcing (TSI) remains UNCHANGED. Because of the dominant role of other mechanisms (primarily evaporation) in heat transfer, only the atmosphere’s ABILITY to retain radiated terrestrial heat is increased–not necessarily its actual heat CONTENT at any given time-scale. The whole AGW meme arises from the misguided notion that radiative transfer in the atmosphere is dominant in setting the surface temperature.
“The whole AGW meme arises from the misguided notion that radiative transfer in the atmosphere is dominant in setting the surface temperature.”
When in fact we can very easily demonstrate the principal mechanism of energy transfer in Earth’s atmosphere is convection. It can be done in a lab and it’s measurable in vivo. The focus on radiative transfer models is misguided and also misdirects the attention of the general public.
Indeed, moist convection–strikingly evident in the formation of cumulonimbus–is the principal means of heating the upper troposphere. At lower levels, oceanographers established in the 1970s (IIRC) that heat transfer by surface evaporation exceeds all other mechanisms combined.
Willis,
This is a “Mommie, the Emperor isn’t wearing any clothes” respite from the gobbledegook I’ve been wading through lately and finally decided isn’t worth my time deciphering, no matter how ignorant I may be of sufficient maths skills to read papers deeply and therefore the constant need for educating and querying.
Coincidentally, I’ve been studying sea level data and acceleration claims for the past three weeks–they’re in the air–so this intrigued me off the top.
I concur completely with Munshi. You can’t use a pencil to describe a pencil (his pt #1).
Anyway, thanks. Like having one of grandma’s lemonades at 5 pm on a hot August afternoon, and discovering she slipped in some Jamaican rum.
Mosh,
Then why can’t they model clouds, wind patterns, and ENSO (which Jim Hansen said he chose to ignore in his models because ‘it was too complicated’)?
I called up Dr. Bill Grey once, and he reeled off a list of the things that models absolutely can’t “model,” because meteorologists, forecasters, and scientists don’t know this information–it ain’t available–at more than an outside extreme of six months (for certain wind flow data). Can’t find the list right now.
And he was just talking about natural variability.
Willis writes: “I’ll leave it there for y’all to take forwards.
I’m sure you know there was a recent article on this site discussing this very subject; the role of epistemology in scientific discourse. In it, the author concisely stated the case that any person conversant with the scientific method was qualified to find fault in its application. That would be the exemplified by the arguments made by Dr. Sophie, who apparently believes the falsification (“falsifyability”?) of a scientific hypothesis is no longer relevant in a post-science world of five year old females who aspire to become scientists. It isn’t required because she says that’s just not the way she does it; cogito ergo sum.
You’re just doing it wrong because you’re so dumb. I flat-out refuse to demonstrate how to do it right.
Hope that’s clear now. Toddle off and let the adults speak.
Very good point W. Using models to test the theory is the same as using Muons to test relativity, that is, a logical fallacy, A lending to B so B can prove A correct
I have to spell it out for you?
OK, Here it is:
Getting research grants form the Federal Government, and getting tenure.
In every time, and at every place it is always about the Benjamins. That is all.
What Is It Good For
Willis should remember this one.
Climate models should predict baseball scores. Because “butterfly flapping its wings”. If they can’t predict baseball scores they are useless.
Still simpler: a model for the result of coin flips. There are very few variables: the side that is up before flipping, the upward acceleration of the flip, the angular momentum imparted to the coin, and the surface it lands on (other potential variables can be eliminated by doing the flips in a controlled environment). Should be simple, no?
Here is my rule: you cannot create a predictive model of a system that has multiple independent variables.
Mr. Simon:
The allusion to models predicting baseball scores is inappropriate as modern climate models do not predict. Instead, they “project.” When a journalist, left leaning politician or statistical neophyte sees the word “project” his or her brain automatically converts it to “predict” as he or she is unaware of the fact that there is a difference in meaning. Unless they are incompetents climatologist know there is a difference but they never in my experience correct the error. Why not? An attractive hypothesis is that to correct the error would be bad for business.
I have a computer model that suggests the planet will be attacked by wave after wave of aliens and it’s hopeless because every time we destroy one incoming wave the next is stronger and faster, the consequences are inevitable – we will lose and we will all be killed – the model proves it. Oh wait – I’ve just realised I was playing space invaders. Panic over.
Willis
This caused me to remember a post you did some years back that indicated the top of atmosphere radiation input output measurements had been adjusted to agree with the models.
Denis
Your post isn’t entirely on topic as Mr. Eschenbach’s article addresses falsifiability but adjustments in the global temperature data are unrelated.
It is quite common to create models in some fields like process engineering. (In 1974 we had a connection from central Queensland to Chicago using fencing wire to manual exchanges to satellites to model a pilot plant process using chlorine gas at 1040 deg C to strip unwanted iron out of the beach sand mineral ilmenite.
Ten tons a day of chlorine in the middle of a town is serious for accidents, so we were thorough).
So what is a model for? Mainly, it is a test to see if you can create a mimic of what you know of a process using parameters that you think are enough for a complete description. If you run your model and it does not give the expected outcome, you have several options –
. junk the model
.refine the model
.overlook the error
It seems to get forgotten that the evaluation of the model is usually how close it comes to expectations. There are a few occasions when the model tells you new or unexpected information, which you then have to work on further if you want to get your sunk money back. If the model does not confirm your expectations, it can be because your expectations are wrong or because you have not calibrated well enough with the right parameters.
Creators are loath to junk models. When they give not quite the expected response, some creators try to push through saying this is state of the art, cost a bundle of $$$, is the best anyone can do, uses a $100 million supercomputer, etc etc. Such marketing exercises are not really excusable. They are vanity exercises. A coverup is a coverup.
It would help if model creators used proper, formal error analysis all the way through.
(I am trying to support the main Willis contention here using words and examples from my past.) Geoff
Geoff
How close the model comes to expectations is the question of accuracy and differs from the question of the truth or falsity of the conclusion of an argument. Modern day global warming models reach no conclusions thus not being falsifiable.
Hi, new guy here so I want to apologise if this idea was covered already…..
There is something missing in the Climate Debate….. Religion.
There are many, many people who look at the issue of Climate Change and claim some variation of “the science is settled.” These people have switched from a scientific point of view to a religious point of view. Lemme ‘splain…..
People who have faith believe in their God(s). It doesn’t matter how much evidence you give against their beliefs, they will keep believing in their God(s).
Climate Change has become a new God, well a new faith to be accurate. The belief that Climate Change is real and serious is immune to any scientific method or research or inquiry. The believers will not be swayed, the heretics who don’t believe will be cast out of the temples.
If you want to crack the shell of the Climate Change group you have to attack their beliefs, not their science.
+1 ^^
“If you want to crack the shell of the Climate Change group you have to attack their beliefs, not their science.”
And of course we have Maslow to tell us this is very, very hard. As a result I personally have little faith in the idea the rationalists involved in this debate are winning; they aren’t winning. At the very best they’ve sent their opponents into the sound proof booth for a little while.
It is their philosophy of science that must be attacked. It confuses pseudoscience with science.
Unbelievable. Look. If the model is not falsifiable IN THEORY then it is pseudoscience. Period.
Sure, just because your model’s predictions cannot be tested in the present does not render your model pseudoscience. Nevertheless, it should be plausible that your model’s predictions can be tested sometime in the foreseeable future.
In other words, the IN THEORY bit must have a pragmatic asepct to it.
RW, it’s important to recall there really was not just popular support for a flat earth, but also authoritarian support. Equally, the heliocentric theory wasn’t generally accepted by the masses or the reigning intellectual authority.
It has happened in the past, it will happen again.
A good thought provoking article. Need to be proved with examples.
Mr Eshenbach’s suspicion is correct. The claims of the climate models are not falsifiable thus global warming climatology is not really scientific. IPCC AR4 explains that falsifiability is an outmoded concept that in the modern era has been replaced by peer review. This, however,is an outrageous lie!
Your error is equating climate models with climatology. The study of climate is more than mere models. In fact models cannot be “falsified”…because a model is either useful, or it is not useful. A good analogy to illustrate your error is equating a hammer with carpentry. Carpentry is much more than a hammer, it includes rulers, saws and other TOOLS. Models are the tools of climatology, so your supposition that a model can be falsified makes no sense.
Mark
Usefulness is unrelated to falsifiability. Falsifiability is a property of a statement that has a property that is called its “truth value.” The truth value takes on the values of “true” and .”false.”
Your conclusion: “global warming climatology is not really scientific” does not follow from the inability to falsify climate model output.
How so?
Climatology is independent of the models. The study of the climate existed before the models were created, so there is no logical dependency.