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
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: http://theconversation.com/climate-change-has-changed-the-way-i-think-about-science-heres-why-82314
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.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Who ever modeled Dr. Sophie certainly didn’t get the required result they wanted.
https://researchers.anu.edu.au/researchers/lewis-sc/image
“Sophie Lewis
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.”
So, now she’s a climate scientist, the science and the scientific method does not apply to her work any more. Thanks for confirming what seems obvious to the rest of us not on that gravy train!
She blinded me with Science!
Forrest,
So she has failed to become what she wanted to be because her idea of what a scientist is and does was different when she was 5 years old. She is, by her own admission, not the scientist she always wanted to be.
These allegations in Lewis’s article are those of someone who has lost “the debate.”
At least all the nut cases and bad singers are in Climate Science and are leaving the other scientists alone to practice real science.
In her defense, climatology is an observational rather than an experimental science, so it can’t be quickly and decisively falsified. that’s what she should have asserted.
Roger,
Actually, models are experiments because they embody everything that is assumed to be important to run an experiment. It is just that the modelers, and the alarmists that use the results to get the public to believe it will warm intolerably, have a problem accepting their own invalidated results.
In fact, the Australian says that climate models don’t have to be falsifiable. Of course not: they’re simply complex methods to calculate the results of a specific hypothesis, given available real world data and reasonable assumptions. The hypothesis that has to be falsifiable is the hypothesis that CO2 is the chief driver of climate: assuming that the climate models represent accurately results if that hypothesis were true, a comparison between model results and real world observation falsifies the hypothesis that CO2 is the chief driver of climate.
Peter Hannan commented”
“In fact, the Australian says that climate models don’t have to be falsifiable.”
why are you
taking your science
from a
newspaper?
Sophie implores
Yes it does, Sophie. What you do as a climate scientist isn’t science. You may want it to be and it may be too complex to do any other way…but that still doesn’t make it science. C’est la vie.
Climate models posit that increased atmospheric CO2 causes increased global temperature. Antarctic and Greenland ice core studies show that increased temperature precedes increased atmospheric CO2. The models are shown to be false. Stick the fork in it, this one is done. Next hypothesis, please.
A model is a mathematical construct based on known first principles, perhaps using coefficients derived from physical experimants and deliberate and precise locations within the model and perhaps with a ‘fudge factor’ to make it fit to a longer data set.
What Dr Lewis is talking about are not models but simply statistical fits of some formula that for parametric reasons appears to follow the data behaviour. The fit is all in the fudge and is useless for the future.
Dr Lewis’s understanding of this area does not seem to have improved much since she was 5 it seems to me.
She works in a ‘science’ where heads you lose tails I win , is considered the standard approach. Where you start with the results you need and then ‘find’ the data to support them, and one where even lying is considered normal and is rewarded. Has a little girl she wanted to be a scientist, well if judge by the normal standards she is still some way short of that.
She is pathetic, just pathetic.
.
That is because it is NOT science, it is political activism. That is why the rules change. If you always wanted to be a scientist, you should have chosen a science subject, not political activism.
It’s not the method that you are supposed to be testing, it is the hypothesis. You really weren’t listening were you?
Climate scientists declare to have confidence because they have tuned their models to give the results they expect which fit their agenda and biases. It is true that this has little to do with falsifiability.
No they don’t. There are no climate models which come even close to reproducing the early 20th c. warming.. Climate models do not have the resolution to even attempt key “processes” like storm formation and precipitation which are simply fudged by tweaking parameters to produce a desired result.
This inconvenient fact is glossed over by activist scientist like Sophie Lewis, who are quite willing to redefine what science is or what it means and have little understanding of it’s basic tenants .
Then it’s a shame you did not chose to work in science when you grew up and chose to be an Australian National University Student instead.
Wasn’t there a study which said higher levels of CO2 affect our ability to reason. They seem to have a point.
“evaluated based on their ability to accurately reproduce observed climate trends and processes” I can build thousands of different models that accurately reproduce “observed climate trends and processes”. Way more accurately than their models. In fact, I could build models that would reproduce the data exactly. All models very wrong and very incapable of prediction.
From her profile page:
Strange the way she omits to put on her profile WHAT SUBJECT she got the PhD and BSc in ! I wonder why that is. It is rather important when you are looking for job or trying to demonstrate your credentials.
Let me guess that it was not a hard science subject.
Climate models should be trusted once they meet the following simple criteria. The data they are compared and successfully matched against data that has both the original data and any modifications , adjustments, corrections etc clearly documented with the reasons for them as well as the uncertainty factor these changes create. This successful prediction should be for a minimum of ten years.
All results should be verified by a quality assurance team not peer review as peer review is fine for pure science with no external financial implications but where money is involved it needs outside certification.
I never thought climate “science” was real science anyway. The words confabulation and fake Nobel prizes comes to mind when it is discussed.
Like chemistry, botany or cognitive sciences, climate science in itself is not falsifiable, but only those elements of it that attempt to constitute a theory, as for exemple the Equilibrium Climate Sensitivity to radiative forcing.
In no way are models experiments! Neither are they scientific theorie!.
They are just attempts to represent some aspects of a portion of [a complex] reality. Quite useful in engineering.
Models must be validated by confronting their outcomes with the osbervable reality.
As no climate laboratory can be set up, only hindcasts may provide some partial elements of validation, with a lot of uncrtainties; their invalidity is easier to demonstrate (see comment and graphs of David Middleton here above).
Even if a model seems to track well past observations, this success is no demonstration of its forecasting abilities.
Correct. The demonstration of predictive skill, or lack thereof, is the test.
I do wonder if she knows what she in saying in basic reason terms, let alone about science. She says models are not falsifiable then says that’s no reason not to trust them.
But that is nonsense in any context. Unless you can show something is trustworthy, you should not trust it. And you certainly shouldn’t trust it to make really large and significant decisions about say our economy.
She is simply describing faith, in any context or field.
And the whole point of science is to do away with the need for faith.
I can well understand the annoyance of people on this topic when someone claims that they no longer have to bother with the issue of whether their model claims have any basis in the workings of the real world. The point that annoys me is that these same “unfalsifiable” models are being used to cause actual harm by diverting billions of dollars etc from helping to combat real problems around the world to line the pockets of chancers and speculators who are peddling a crisis that never remotely sounded plausible and has singularly failed to materialise.
Nor have other sciences been unaffected by this warmest nonsense. In Europe we even had an Austrian professor of Music saying climate deniers (whatever that is) should be hanged! Can you imagine that! An Austrian! Do these people have no knowledge of the past?
By comparison with warmistas I am beginning to have far more sympathy with the fruitcakes who think we are being constantly invaded by aliens etc. At least they are relatively harmless ,often sincere, people who provide us with entertainment and like other forms of science fiction seem to get more of their science right than the people who advocate we are dangerously warming our climate through a modest increase of a beneficial gas that has never been implicated in being a primary driver of the Earth’s heat balance.
Isn’t being not falsifiable at the very heart of the climate change ethos? Everything is confirmation however contradictory – more cold less cold, more heat less heat, more rain less rain etc. Every previously natural weather event or pattern is now held up as ‘proof’.
“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.”
Sophie is a very confused person. A climate model is nothing more or less than a computer program that implements someone’s theory. Theories can only be tested by comparison of their predictions to reality. This isn’t rocket science, nor is it an example of “the complexity of science.”
Here she makes the absurd claim that a test of a model can be carried out by either observation OR A MODEL TEST.” She also makes the stupid claim that calling a model false requires proving that man made GHGs don’t cause global warming. While it may be true, depending upon a definition of falsibility that means “no ability to explain any warming, no matter how slight,”
that hardly is proof that the model has any practical value. All scientific theories are probably wrong , but that says nothing about their importance – Newton’s Laws we know are not absolute, but that didn’t stop NASA from using them to calculate space flights. You don’t have to prove a model false in order to validly claim it is useless for any practical purpose. It is astounding that this “scientist” would claim that climate models have shown excellent results. I guess she never got Michael Mann’s email in which he characterized all climate models as “junk.”
First they wanted to change the definition of peer review. Now they want to change the definition of science itself.
“I’ve wanted to be a scientist since I was five years old..”
Ah, so that explains your becoming a fake one instead, wanting to change the rules. Sorry, buttercup. Doesn’t work that way.