The information you provide here is recognizable from your other writings on the subject. One thing that strikes me when reading this is the numbers. Do the data really support figures such as 1 in 13.9 years? Does that mean uncertainties are in the second decimal?
One other thing is this notion about 100 year events, 1000 year events etc. You have written about this before, how the 100 year event really is an event with a 1% probability of happening every year. But is it not so that to actually know what a 100 year event is you need 300 years of observations, in a non-changing environment?
Attribution 'science' is so skewed! By definition no one is looking at the avoided/reduced extreme weather events. How can we make that more balanced? Is that possible at all?
It has such a strong component of 'confirmation bias' in it at the moment. A few years ago all the weather/climate experts in NL were explaining that the recent couple of very dry years (resulting in a continuous 'deficit' and low ground water levels) were a result of climate change and we should prepare for more of that.
Last night, while discussing an exceptionally wet year, the same experts were explaining that because of higher temperatures the air contains more water and more rain is inevitable in the Netherlands.
Both can be true of course, but it would be nice if they would have at least referred back to their warmings about increasing droughts a few years ago, and explained if and how it all fits together. And why this is not just a matter of natural variability of the weather.
I asked Copilot the following question: If the probability of an incident occurring in 2020 was 10% and the probability of that incident occurring in 2024 was 15% how much more probable was that incident occurring in 2024 than in 2020? Copilot gave the answer 50%, reasoning that the difference of the two probabilities was 5%, but the increase in the 2020 probability was 5%/10% or 50%. Looks like Copilot computes like the guys at Imperial College.
Maybe see what Copilot says about the difference between relative and absolute risk too. This is an issue people should be aware of when they read articles on health studies that try to generate clicks by highlighting the change in relative risk. It’s not that it’s wrong; it’s just not the whole picture.
For example:
“A new study published in the Lancet indicates that brain tumours in later life were **three times as common** among people who had two to three CT head scans in childhood, while the risk of leukaemia was **three times greater** in patients receiving five to 10 scans during childhood.
But the researchers who carried out the study emphasise that the overall absolute risk of people developing cancer after receiving CT scans remains small – amounting to two extra cases of cancer per 10,000 children who received CT scans.”
So looking back at your scenario at the top it makes me realize just how important it is to specify relative or absolute when one is talking about changes in risk.
Very good point on absolute vs relative risk. By the way, I asked Copilot the same question except changed my probabilities to 0% in 2020 and 1% in 2024 and got the answer that the probability increased by 1%. But then I ran it with 0.1% and 1% and got the answer that the probability increased by a factor of 10.
Right. Note the language in the Imperial College press release:
"A “Chido” type storm is about +40% more likely in the 2024 climate compared to a pre-industrial baseline"
They claim that the "storm" is 40% more likely. But that is objectively false. The "storm" is 2.8% more likely. The probability increased by 40%, not the storm. Too clever for me!
An example I came up with did not make the cut for this post -- Imagine that the chances of the Earth being swallowed by a black hole are 1 in four quadrillion. Then I decide that they are instead 2 in 4 quadrillion. In both cases the chances are essentially zero. Imagine the headline, "Expert Increases Chances of Earth's Demise by 100%"! ;-)
Attribution studies are also being done to eventually be used in supporting lawsuits against companies and other entities that emit large amounts of CO2 and have large amounts of money.
Analogies are, of course, never to be confused with proofs. But they can be illustrative and instructive.
Consider this analogy:
My broker calls me and says, remember that investment (extreme event) you made 10 years ago? At the time I told you the expected payback period was 14 years. But guess what, you’ve received payback of all of your initial investment this year. It arrived 40% (14/10) sooner than expected because of inflation (climate change).
Good question, although in this case it's not possible to show a statistically significant correlation yet it's baldly reported as causation and without reference to the overall actual record of hurricanes in the time frame
I do have to admit that the IPCC attribution standard is also hard to meet. An increase from 7,2% to 10,0% is real, no matter whether it takes 2.100 years to be more than 90% certain.
It's increasingly my believe that people are not rational, but they do like to rationalize. This process is everywhere in our society and we are all subject to it. Rationalisation is often used in the process of manipulation.
However, given the magnitude of internal variability in tropical cyclones the idea that we could ever detect a change this small seems completely unrealistic.
Fortunately, there is absolutely no decision or policy that hinges on this small difference.
Roger, I think you are being inconsistent here. You state in your original post: "Claims that certain events have increased in their likelihood of occurring should always be accompanied by a time series of those events to demonstrate that increase — increases such as 7,000% or even 40% over 40 years should be obvious, even without statistics." This seems reasonable to me, and I find the empirical frequency or ACE data shown over a long period to be very helpful. But if the internal variability is sufficiently large that we cannot reliably detect a difference between 7.2% and 10.0%, then we should not take much comfort from the lack of trend in frequency/ACE over 40 years. I am no statistician, but it seems to me that if it really were the case that frequency had increased from 7.2% to 10.0% we WOULD see some effect "visually" in the time series data.
The bigger problem here, it seems to me, is the selection bias on the part of those reporting. They naturally want to take whatever model appears to give the most eye-popping effect. The 7.2% going to 10.0% seems to me highly inconsistent with the findings of Knutson et al, back in 2019 for the AMS.
Hi Moray- For TCs we would be unlikely to see the effects of a change in probability of 7.2% to 10% over centuries of data, it is just too noisy. 40% is a different story. Detection of change under the IPCC framework for D&A is a high bar, and that is because climate is noisy and chaotic. It is easy to see patterns in randomness so the IPCC sets a high bar.
Until recently we were faced with an impossible communication problem in getting to the truth with institutions such as the UN and governments and media avidly supporting the science fiction you have mentioned here.
The good news is that we now have new governments and voters that just wont listen to this narrative anymore.
I believe many climate advocacy groups that sucked up all this stuff will disband in 2025 as they will see a population less interested in listening.
We will see a firm move toward climate realism and a dismantling of the NetZero policy set in most western nations.
The bad news is that it will take significant time to unwind the damage done in funding direction and those trapped in the false universe of truth, and who have hitched their livelihoods to this imploding star.…
Our job will be to assist in the re-education as the climate reality takes hold.
None of them will disband until the money stops flowing. They're not there to get listened to. They are there to get paid.
$4B per year in anti-industrial money to NGOs, laundered through "charitable" foundations and ultimately sourced from billionaires and foundations that hate humanity.
Happy new year to all at THB!
The information you provide here is recognizable from your other writings on the subject. One thing that strikes me when reading this is the numbers. Do the data really support figures such as 1 in 13.9 years? Does that mean uncertainties are in the second decimal?
One other thing is this notion about 100 year events, 1000 year events etc. You have written about this before, how the 100 year event really is an event with a 1% probability of happening every year. But is it not so that to actually know what a 100 year event is you need 300 years of observations, in a non-changing environment?
Decision based evidence making.
Roger, thanks a million for the reference to the Tu paper. Excellent information, very much on point.
And Happy New Year. All best wishes for a prosperous, productive year.
Finally, someone who understands what the phrase “begging the question” actually means.
Attribution 'science' is so skewed! By definition no one is looking at the avoided/reduced extreme weather events. How can we make that more balanced? Is that possible at all?
It has such a strong component of 'confirmation bias' in it at the moment. A few years ago all the weather/climate experts in NL were explaining that the recent couple of very dry years (resulting in a continuous 'deficit' and low ground water levels) were a result of climate change and we should prepare for more of that.
Last night, while discussing an exceptionally wet year, the same experts were explaining that because of higher temperatures the air contains more water and more rain is inevitable in the Netherlands.
Both can be true of course, but it would be nice if they would have at least referred back to their warmings about increasing droughts a few years ago, and explained if and how it all fits together. And why this is not just a matter of natural variability of the weather.
I asked Copilot the following question: If the probability of an incident occurring in 2020 was 10% and the probability of that incident occurring in 2024 was 15% how much more probable was that incident occurring in 2024 than in 2020? Copilot gave the answer 50%, reasoning that the difference of the two probabilities was 5%, but the increase in the 2020 probability was 5%/10% or 50%. Looks like Copilot computes like the guys at Imperial College.
Maybe see what Copilot says about the difference between relative and absolute risk too. This is an issue people should be aware of when they read articles on health studies that try to generate clicks by highlighting the change in relative risk. It’s not that it’s wrong; it’s just not the whole picture.
For example:
“A new study published in the Lancet indicates that brain tumours in later life were **three times as common** among people who had two to three CT head scans in childhood, while the risk of leukaemia was **three times greater** in patients receiving five to 10 scans during childhood.
But the researchers who carried out the study emphasise that the overall absolute risk of people developing cancer after receiving CT scans remains small – amounting to two extra cases of cancer per 10,000 children who received CT scans.”
https://news.cancerresearchuk.org/2013/03/15/absolute-versus-relative-risk-making-sense-of-media-stories/
So looking back at your scenario at the top it makes me realize just how important it is to specify relative or absolute when one is talking about changes in risk.
Very good point on absolute vs relative risk. By the way, I asked Copilot the same question except changed my probabilities to 0% in 2020 and 1% in 2024 and got the answer that the probability increased by 1%. But then I ran it with 0.1% and 1% and got the answer that the probability increased by a factor of 10.
Right. Note the language in the Imperial College press release:
"A “Chido” type storm is about +40% more likely in the 2024 climate compared to a pre-industrial baseline"
They claim that the "storm" is 40% more likely. But that is objectively false. The "storm" is 2.8% more likely. The probability increased by 40%, not the storm. Too clever for me!
An example I came up with did not make the cut for this post -- Imagine that the chances of the Earth being swallowed by a black hole are 1 in four quadrillion. Then I decide that they are instead 2 in 4 quadrillion. In both cases the chances are essentially zero. Imagine the headline, "Expert Increases Chances of Earth's Demise by 100%"! ;-)
Simple. Black holes are also caused by climate change.
Attribution studies are also being done to eventually be used in supporting lawsuits against companies and other entities that emit large amounts of CO2 and have large amounts of money.
Analogies are, of course, never to be confused with proofs. But they can be illustrative and instructive.
Consider this analogy:
My broker calls me and says, remember that investment (extreme event) you made 10 years ago? At the time I told you the expected payback period was 14 years. But guess what, you’ve received payback of all of your initial investment this year. It arrived 40% (14/10) sooner than expected because of inflation (climate change).
Would you accuse him with ‘gaslighting’?
Thanks,
Gene
Good question, although in this case it's not possible to show a statistically significant correlation yet it's baldly reported as causation and without reference to the overall actual record of hurricanes in the time frame
That should have read: Surely, if the likelihood of something happening increases so should the frequency of it actually happening .
If I were Mr. Mroz I would object to being called Shirley.
Surely, if they don’t mind likelihood of someth8ng happening increases so should the frequency of it actually happening?
I do have to admit that the IPCC attribution standard is also hard to meet. An increase from 7,2% to 10,0% is real, no matter whether it takes 2.100 years to be more than 90% certain.
It's increasingly my believe that people are not rational, but they do like to rationalize. This process is everywhere in our society and we are all subject to it. Rationalisation is often used in the process of manipulation.
Codification of rationalisation is the end play.
I agree with this
However, given the magnitude of internal variability in tropical cyclones the idea that we could ever detect a change this small seems completely unrealistic.
Fortunately, there is absolutely no decision or policy that hinges on this small difference.
Roger, I think you are being inconsistent here. You state in your original post: "Claims that certain events have increased in their likelihood of occurring should always be accompanied by a time series of those events to demonstrate that increase — increases such as 7,000% or even 40% over 40 years should be obvious, even without statistics." This seems reasonable to me, and I find the empirical frequency or ACE data shown over a long period to be very helpful. But if the internal variability is sufficiently large that we cannot reliably detect a difference between 7.2% and 10.0%, then we should not take much comfort from the lack of trend in frequency/ACE over 40 years. I am no statistician, but it seems to me that if it really were the case that frequency had increased from 7.2% to 10.0% we WOULD see some effect "visually" in the time series data.
The bigger problem here, it seems to me, is the selection bias on the part of those reporting. They naturally want to take whatever model appears to give the most eye-popping effect. The 7.2% going to 10.0% seems to me highly inconsistent with the findings of Knutson et al, back in 2019 for the AMS.
Hi Moray- For TCs we would be unlikely to see the effects of a change in probability of 7.2% to 10% over centuries of data, it is just too noisy. 40% is a different story. Detection of change under the IPCC framework for D&A is a high bar, and that is because climate is noisy and chaotic. It is easy to see patterns in randomness so the IPCC sets a high bar.
Attributing motives to those trying [badly] to attribute every weather event to climate change requires its own alchemy.
Fortunately the "attribution science" did not need to explain why none of Michael Mann's 33 named storms achieved CAT6. ;-)
YEP... Right on!
Until recently we were faced with an impossible communication problem in getting to the truth with institutions such as the UN and governments and media avidly supporting the science fiction you have mentioned here.
The good news is that we now have new governments and voters that just wont listen to this narrative anymore.
I believe many climate advocacy groups that sucked up all this stuff will disband in 2025 as they will see a population less interested in listening.
We will see a firm move toward climate realism and a dismantling of the NetZero policy set in most western nations.
The bad news is that it will take significant time to unwind the damage done in funding direction and those trapped in the false universe of truth, and who have hitched their livelihoods to this imploding star.…
Our job will be to assist in the re-education as the climate reality takes hold.
More on the situation at…
https://nigelsouthway.substack.com/p/climate-realism-is-on-the-way
None of them will disband until the money stops flowing. They're not there to get listened to. They are there to get paid.
$4B per year in anti-industrial money to NGOs, laundered through "charitable" foundations and ultimately sourced from billionaires and foundations that hate humanity.
Agreed... it will take time and will be a damage period.
Me thinks “A less charitable explanation …” is the correct one.
Really interesting and thought-provoking insight!