I have a question Roger: have you come across data used for models that would cause the model to underestimate the rise of global temperature in the future?
There are many other variables in the models that are not in your field of expertise that influence it's outcome.. Despite many media reports on things being worse than we thought, it seems to me that things aren't as bad as we assumed and that when we continue to decarbonise at a rate a bit higher than we historically have, we need not worry too much about future impacts.
Great work. I think testing the robustness of the assumptions in the scenarios is critical to understanding the accuracy of the climate models.
I would love to see another robustness test for percentage of energy produced from coal. I have heard that RCP 8.5 assumes a massive increase in coal use. This implausible assumption must have a big impact on climate projections.
The UN population forecast failed to respond to their own data on things that effect fertility rate. So, shouldn't we also consider the many things that trend to reduce the sensitivity to CO2? Things like measured temperatures consistently below IPCC's models, new work on clouds, UHI, sea temperatures, the history of IPCC to report temperature forecast toward the higher end of the range, new work by on convection, feedback, spectral response, greening of the earth. etc.
Roger, this piece is truly profound. The iimplications need to be understood by all and sundry. Thanks a million for giving us this one. I'm going to read it over about four times to make sure I understand it.
One other important factor: as the world population declines after peaking perhaps around 2060 the average age will be higher. The population in many countries (for example China) will have more older people than younger. The amount of energy used per person will decrease as a result.
These projections assume that additional increases in CO2 will have the same effect as previous increases. Happer and others argue against this. My physics is not good enough to evaluate the arguments. (Long ago very high levels of CO2 at least intuitively support the claim.) Unless you have strong reason to dismiss that view it should be noted as an additional uncertainty in the projections.
Correcting the SSPs. RP considers the effects of new population projections on SSP2-4.5. It is now 2024. Each SSP has a projection (more than one, actually, depending on climate model used) for temperature in 2024. In general these projections are incorrect. So one must also correct for these errors. (Not a trivial task.)
Article 4.1 refers to "the second half of the present (19th) century" as the baseline period. This is 1850-1899, not 1850-1900. The latter is 51 years. Half of a century is 50 years. And ... the year 1900 is in the 20th century.
I noticed that no one anywhere is willing to discuss, women's empowerment in the form of birth control, religion, resource scarcity. Perhaps more constructive confrontation and less political correctness?
Agree. You need to add "better health metrics" to your list. But then don't forget that women's empowerment has come hand-in-hand with substantial technological innovation and energy use. As we continue to innovate, the empowerment of which you speak will continue to grow (I hope).
Roger, I was startled to read this: "Consider that SSP3 — which has global population implausibly approaching 13 billion and rising in 2100 — has been the subject of almost 1,500 papers so far this year." I would guess this is because the basic ideas for these models were developed in the early 1990s, and thus include assumptions that might have made sense back then, but don't now. And they've been kept because people want to keep the models consistent over time. Is that correct? Also, are the other SSPs as out to lunch on their population assumptions?
Exactly right. The SSPs were published more than a decade ago, and were developed over the decade prior, grounded in assumptions that go back into the decade before that.
The reason for the slow pace in updating them has IMO less to do with a desire to keep consistency, but rather, to align the scenario schedule with the needs of physical science modeling (e.g., under CMIP) which evolves on a decadal pace. The needs of policy and the needs of climate science are thus inconsistent. Because scientists dictate scenario schedules, you can guess which priorities win out.
So, we are in a race to end humanity. Which will get us to the "finished" (as a species) line more quickly: unrestrained global climate change, or falling birthrates? Hold on to your hats, gentlemen!
John Ibbitson, a feature writer for the Globe and Mail (Canada's foremost daily newspaper) and Darrell Bricker (CEO of Ipsos Public Affairs) wrote a book in 2019 titled "Empty Planet - The Shock of Global Population Decline".
Their prediction was peak population in 2050 which is consistent with Fernandez-Villaverde's 2054.
Fascinating post. Good exercise for the brain too following all your steps.
Question: Any idea what kind of reputation DNV has for their energy forecasting? I’d be curious to dig into their latest outlook and see their GDP and population assumptions. They’re projecting a central estimate of +2.2C for 2100 (always good to note that these central temperature projections are plucked from a distribution).
Another excellent article. The population declines seem consistent with and the next step of what Barry Commoner called demographic transitions over 50 years ago.
I have a question Roger: have you come across data used for models that would cause the model to underestimate the rise of global temperature in the future?
There are many other variables in the models that are not in your field of expertise that influence it's outcome.. Despite many media reports on things being worse than we thought, it seems to me that things aren't as bad as we assumed and that when we continue to decarbonise at a rate a bit higher than we historically have, we need not worry too much about future impacts.
Great work. I think testing the robustness of the assumptions in the scenarios is critical to understanding the accuracy of the climate models.
I would love to see another robustness test for percentage of energy produced from coal. I have heard that RCP 8.5 assumes a massive increase in coal use. This implausible assumption must have a big impact on climate projections.
True!
See our paper on this
https://iopscience.iop.org/article/10.1088/1748-9326/ac4ebf
Roger, when we see something like a 2.2C change over preindustrial, how much of that has already occurred?
Just about half
The UN population forecast failed to respond to their own data on things that effect fertility rate. So, shouldn't we also consider the many things that trend to reduce the sensitivity to CO2? Things like measured temperatures consistently below IPCC's models, new work on clouds, UHI, sea temperatures, the history of IPCC to report temperature forecast toward the higher end of the range, new work by on convection, feedback, spectral response, greening of the earth. etc.
Where in the calculations of temperature rise vs CO2 is the correction for Saturation, ala Happer?
Roger, this piece is truly profound. The iimplications need to be understood by all and sundry. Thanks a million for giving us this one. I'm going to read it over about four times to make sure I understand it.
One other important factor: as the world population declines after peaking perhaps around 2060 the average age will be higher. The population in many countries (for example China) will have more older people than younger. The amount of energy used per person will decrease as a result.
These projections assume that additional increases in CO2 will have the same effect as previous increases. Happer and others argue against this. My physics is not good enough to evaluate the arguments. (Long ago very high levels of CO2 at least intuitively support the claim.) Unless you have strong reason to dismiss that view it should be noted as an additional uncertainty in the projections.
Correcting the SSPs. RP considers the effects of new population projections on SSP2-4.5. It is now 2024. Each SSP has a projection (more than one, actually, depending on climate model used) for temperature in 2024. In general these projections are incorrect. So one must also correct for these errors. (Not a trivial task.)
Agreed
Article 4.1 refers to "the second half of the present (19th) century" as the baseline period. This is 1850-1899, not 1850-1900. The latter is 51 years. Half of a century is 50 years. And ... the year 1900 is in the 20th century.
Ha! I read the comment and knew it was you before I saw your name;-)
You are correct.
I noticed that no one anywhere is willing to discuss, women's empowerment in the form of birth control, religion, resource scarcity. Perhaps more constructive confrontation and less political correctness?
Not directed at you Roger at all. :-D
Agree. You need to add "better health metrics" to your list. But then don't forget that women's empowerment has come hand-in-hand with substantial technological innovation and energy use. As we continue to innovate, the empowerment of which you speak will continue to grow (I hope).
I agree with you. That would entail uncomfortable conversations that many people aren't willing to have.
Roger, I was startled to read this: "Consider that SSP3 — which has global population implausibly approaching 13 billion and rising in 2100 — has been the subject of almost 1,500 papers so far this year." I would guess this is because the basic ideas for these models were developed in the early 1990s, and thus include assumptions that might have made sense back then, but don't now. And they've been kept because people want to keep the models consistent over time. Is that correct? Also, are the other SSPs as out to lunch on their population assumptions?
Exactly right. The SSPs were published more than a decade ago, and were developed over the decade prior, grounded in assumptions that go back into the decade before that.
The reason for the slow pace in updating them has IMO less to do with a desire to keep consistency, but rather, to align the scenario schedule with the needs of physical science modeling (e.g., under CMIP) which evolves on a decadal pace. The needs of policy and the needs of climate science are thus inconsistent. Because scientists dictate scenario schedules, you can guess which priorities win out.
So, we are in a race to end humanity. Which will get us to the "finished" (as a species) line more quickly: unrestrained global climate change, or falling birthrates? Hold on to your hats, gentlemen!
Why not throw in AI while you're at it? Engineered pandemics, meteor strikes, etc.
John Ibbitson, a feature writer for the Globe and Mail (Canada's foremost daily newspaper) and Darrell Bricker (CEO of Ipsos Public Affairs) wrote a book in 2019 titled "Empty Planet - The Shock of Global Population Decline".
Their prediction was peak population in 2050 which is consistent with Fernandez-Villaverde's 2054.
Even earlier was The Empty Cradle (2004) by Phillip Longman.
Fascinating post. Good exercise for the brain too following all your steps.
Question: Any idea what kind of reputation DNV has for their energy forecasting? I’d be curious to dig into their latest outlook and see their GDP and population assumptions. They’re projecting a central estimate of +2.2C for 2100 (always good to note that these central temperature projections are plucked from a distribution).
https://www.dnv.com/energy-transition-outlook/
PS I think Matt Burgess’s work deserves wider propagation so good to see a new paper cited here.
PSS The first Burgess link seems to be broken for me. This one: “extraordinaire Matthew Burgess, of the University of Wyoming, and colleagues, who”
Thanks
I am impressed by the work of DNV, they deserve more attention
Agreed on Matt, and I fixed the link, thanks for noticing that.
Another excellent article. The population declines seem consistent with and the next step of what Barry Commoner called demographic transitions over 50 years ago.