Probability Estimates of a 21st Century AMOC Collapse
arxiv.org
arxiv.org
The collapse time is estimated between 2037-2064 (10-90% CI) with a mean of 2050 and the probability of an AMOC collapse before the year 2050 is estimated to be 59 +/- 17%.
We'll miss you, Gulf Stream.
TLDR: the Gulf Stream originates from a very different phenomenon than the AMOC and is not at risk of collapse. A small part of the heat transported by the Gulf Stream does originate from the AMOC though, but its collapse would be much less severe than a full collapse of the Gulf Stream.
[0]: https://bigthink.com/strange-maps/gulf-stream-collapse-amoc/
The Wikipedia page on the AMOC (https://en.wikipedia.org/wiki/Atlantic_meridional_overturnin...) links to a page summarizing some scientists’ reactions:
https://www.sciencemediacentre.org/expert-reaction-to-paper-...
However, I have no idea if this is representative or not.
https://www.wired.com/story/amoc-collapse-atlantic-ocean/
HN discussion: https://news.ycombinator.com/item?id=41087977
All of those models show the AMOC to be meta-stable (meaning it rapidly flips from on to off, as opposed to long gradual slowing).
All of those models show the AMOC is on track to shut off "soon."
All of those models show an AMOC shutoff event to be very serious, particularly for Northern Europe, but also for the world as a whole.
Prior to this paper, the estimates on what "soon" meant were not very good and generally took the form of "it could be this century, or it could be in the next couple centuries."
This paper has not been through peer review yet, but the authors are significant participants in the field.
The paper is trying to answer one of the most important question in science right now, which is how long until the AMOC shuts down and what signals should one watch for to improve that prediction.
Presumably this means the AMOC is on track to hit its tipping point "soon". How long after hitting the tipping point would it take to "stop"? 100 years? Less?
(Unless this leads to a war, which it might).
Bear in mind that Europe trying to compensate for an AMOC collapse with more greenhouse gases would involve enough greenhouse gases to make India uninhabitable and that India has nukes, and that such a temperature change would significantly alter the agricultural productivity of the USA and the USA has nukes, and that half of China is a desert and China has nukes.
I'm not sure if that kind of change would require global changes sufficient to melt the Greenland and Antarctic ice sheets, but if it did then that would also result in basically every costal nation also considering it an existential threat and thus warranting military action to prevent, including the UK.
The scenario of "post AMOC collapse Europe tries to compensate with more global climate change" is an existential threat to at least some other nations, how many depends on the scale of the change caused by the AMOC collapse and thus the scale of the change from the super-greenhouse-gasses produced in this scenario to compensate for it; even if the UK/France nuclear arsenals were significantly larger, this would not be sufficient to deter India from using nukes in this scenario if conventional methods failed.
Also, again, deliberately trying to push through an AMOC with even more greenhouse gasses is something most people don't model, so all parties would have to seriously consider if this would be at the level to melt global ice-sheets causing a (58m Antarctica + 7m Greenland =) 65m sea level rise rather than the much smaller one we'd expect in a business-as-usual scenario.
Set this map to 65m: https://www.floodmap.net
Then consider each scenario. If the AMOC fails, it's an open question how severe that even is — I've seen estimates between 2°C and 30°C of local cooling, and I expect the lower range of that hence my previous statement "Europe north of about Paris will look like southern Canada already does".
At lower levels of that range, Europe just goes back to where we were a century ago, at higher levels it could only be "solved" by greenhouse gas emissions that would cause the Antarctic and Greenland ice sheets to fully melt. Looking at the global map, my guess is that Spain and France end up like the USA Pacific coast, which isn't too bad, while the UK and Ireland have severe issues.
But even a "mere" extra 2°C globally is enough to be an existential threat to certain populations including in parts of India, who would do whatever it takes to prevent that.
If it's really a 30°C collapse — and it would have to be on that scale to seriously consider global geo-engineering rather than retrofitting buildings and building greenhouses over the farmland — then that risks the aforementioned ice sheet collapse, which neither the UK nor Ireland can risk because that would turn the problem from "it's cold and we need to import food" to "and now 50% of the country is underwater", likewise half of Belgium, all of the Netherlands, much of Denmark, and a significant part of the most densely populated regions across the continent including Italy.
France, the only other nuclear armed state in the EU besides the UK, is only affected a little bit by the sea level rise, and thus the only one which might consider that course of action. But even they would have a serious engineering project to stop the Seine from flooding back to Paris. And even with that, silver bullets (your nukes) can't stop golden guns (other people's nukes).
-
No, if this particular disaster happened we might try to mass-migrate to Africa (to a chorus of "oh no not again" from the destination), but I don't see our governments trying to solve it in the "old lady who swallowed a fly" approach.
Anyway, given that Europe is in some scenarios heating up a lot while in others (like here) it freezes - not sure any deserve much effort on speculation. Too much uncertainty.
I could believe anything from half of us being dead from an accidentally mis-engineered industrial GM algae, to von Neumann probes having been let loose on the moon for mass production of aluminium sheets to be placed in Earth-Sun L1 as a solar shade, to the escalation of current conflicts into WW3, to a random terrorist getting one nuke and missile and using that to trigger a continental-scale EMP that shuts down industrial civilisation in a densely populated area.
(Only things I won't believe involve large-scale cooperation, so we're not all going to speak Esperanto by then :P)
What is [Country X] currently doing about it, then? I'd assume Country X is currently banning/disincentivizing cars, cement, AI clusters, plane travel, and all kinds of much easier measures than literal nuclear attacks on other nuclear-equipped countries.
The world is complex. There are very few (maybe zero?) cases where any given nation state is willing to do "whatever it takes" to do anything.
> What is [Country X] currently doing about it, then? I'd assume Country X is currently banning/disincentivizing cars, cement, AI clusters, plane travel, and all kinds of much easier measures than literal nuclear attacks on other nuclear-equipped countries.
Country X in this case being India, here's what they're currently up to: https://www.weforum.org/agenda/2024/02/climate-policy-india/
I was wondering if I'd emphasised strongly enough "in the scenario where AMOC has collapsed and Europe tries to resolve this by overcoming it with extra greenhouse gas emissions". This scenario is trying to correct a local 2°C reduction with enough extra greenhouse gas for a global 2°C *more than today*. Current temperatures are about 1.25°C over the pre-industrial average, so that brings us to a minimum lower bound of 3.25°C over the pre-industrial average. This is in the region of "about half the world population experiences deadly humid heatwaves every year, the Himalayan glaciers stop supplying fresh water to half of the Indian population, marine ecosystems collapse, nearly all crops are negatively affected and yields fall rapidly, 18% loss of global GDP": https://unclimatesummit.org/comparing-climate-impacts-at-1-5....
We're not currently on a course for that in the first place, the current policies reach 2.6°C relative to the pre-industrial average (1.35°C relative to today) by 2100, so what India (and everyone else) has to do to prevent that is… basically what everyone is already currently doing. Though we can do better, and we should, because 2.6°C over pre-industrial (1.35°C over today) is still pretty bad.
But if it came to that scenario? That hypothetical where the European continent (including the UK) gives the world a two fingered salute? The negative consequences of being on the receiving end of the UK and France nuclear arsenals, would still be less bad for India and its people even if they acted unilaterally than allowing the extra impact of that much additional greenhouse gasses. Never mind that all those problems with an additional 2°C beyond current would cause to everyone else which means that the USA won't be on Europe's side if Europe was hubristic enough to try to solve this with geo-engineering rather than something less drastic.
And I did also only say that as a follow-on if "Not selling them stuff to burn" was shown to be insufficient. There are of course many stops between "don't sell them stuff" and "nuke them", I didn't feel it was a good use of time to enumerate them.
So, my bet is that in case of AMOC collapse Europeans simply elect authoritarian governments that will make some unfortunate developing countries colonies once again, and use them for farming. Given bad enough existential threat, things like rules-based order, human rights and democracy cease to matter, and the strong take what they want from the weaker parties. That's just how humans work.
I've seen estimates between 2°C and 30°C of local cooling.
I expect the lower range of that hence my previous statement "Europe north of about Paris will look like southern Canada already does", and that's fine for northern Europe but may still bite Spain (the Florida of Europe) as it (or at least its Med coast) is destined for being too hot more than being too cold.
Agriculture is the main thing, and I don't feel confident about predicting that, might be OK or not.
30°C of local cooling (I have no idea how that might even be possible but I've seen it as the worst case estimate) is way outside what most of the buildings are designed for.
And geoengineering to spew more CO2 into the atmosphere won't really work and is idiotic enough, it won't happen. I guess I didn't understand that line of argument because of how ridiculous it is.
Much more likely that Europe just switches to energy-intensive hydroponic farming and to imports for whatever that doesn't work for.
On this we both agree :)
> it won't happen
Sadly this doesn't follow from the premise; governments regularly do stupid things that don't work.
Looking back, the greenhouse effect of CO2 was known in the 1800s, but even at the end of that century people thought it would take a millennium and be good.
Looking forward, it's technically possible (yes I have done the maths) to build a global power grid today such that the Scaninavian arctic midwinter can be powered by solar panels in New Zealand, but only China makes enough aluminium and only China has demonstrated any interest in attempting massive infrastructure projects of this kind for soft power projection.
When do we get von Neumann replicators? Do we make GM algae that turn atmospheric CO2 directly into crude oil to substitute fossil oil in the petrochemical industry? Will fundamentalists Hindus create a synthetic disease that induces an allergy to beef thus causing an end to cattle farming? Will the Russian invasion of Ukraine escalate into WW3?
I can't tell which specific big things will radically change the world, but I am sure there's going to be at least one every 25 years — one between today and 2050, at least two more by 2100, and at least four between 2100 and 2200.
IIRC Charles Babbage wrote in his survey of British Industry for the royal society words to the effect of "it is not presently known how exactly the coal exhaust is recycled back into the ground".
Maybe some people correctly guessed that it wasn't recycled at all and in fact is an expendable (albeit quite large) resource, but their guess was as good as any other and far from something that was widely known and believed.
Arrhenius published an update in 1908 (not sure when he realised the mistake) accounting for the increased rate of CO2 emissions, stating that would be "beneficial"[0] in his own lifetime — he ended up dying in 1927, I'll let others decide if it was a net good or bad by that point.
[0] Wikipedia's citation for this is "Arrhenius, S., Worlds in the Making: The Evolution of the Universe. New York, Harper & Row, 1908," — https://dn790004.ca.archive.org/0/items/worldsinmakingev00ar...
I can find the "beneficial" part, but not (yet) the "in [his] lifetime" part. Quoting from the linked archive copy:
"By the influence of the increasing percentage of carbonic acid in the atmosphere, we may hope to enjoy ages with more equable and better climates, especially as regards the colder regions of the earth, ages when the earth will bring forth much more abundant crops than at present, for the benefit of rapidly propagating mankind." - page 63.
Aren't we still due for one between now and 2025 then?
The first draft of the Human Genome Project was only published in 2003, but that covered 92% of the genome; it only got up to 99.7% in 2022. Personal genome tests are now regularly on special offer for a few hundred bucks rather than the $2.7 billion (non-inflation adjusted) of the original Human Genome Project.
Genetic engineering was possible in 1999, but it wasn't until 2005 that CRISPR was discovered, 2006 that it was theorised how it might function, 2007 that this was demonstrated, 2012 until it was patented, and 2013 until it became commercially viable.
Mobiles and the web technically existed in 1999, but the former were mostly black and white screens with number-pad keyboards and the latter was mostly dialup; now it's good enough that the global pandemic we just went through had a lot of people switch to remote work — and ask to keep doing that after the pandemic restrictions ended. The tech that enabled the vaccines to be created so fast, or even their genomes to be sequenced so fast, didn't exist in 1999.
PV existed in 1999, but there was a total installed nameplate capacity of 0.8 GW; in 2022 that was 1061.7 GW, with an estimated 341-402 GW in 2023 and 600-660 GW this year, likely taking the nameplate capacity over the current total global electricity demand. (The capacity factor was 14% in 2022, so there's room for the nameplate capacity to become 700% of current global electrical demand).
3D printers existed in 1999, but cheap domestic models only happened around the time of RepRap in 2005. Brief period where it seemed everyone was worried about domestic printers making guns, now industrial printers make spaceship engines.
AI has gone from "squiggly line under a mis-spelled word", Norns with tiny neural nets, and the state machines of video game NPCs, to what is essentially the character of the main computer on TNG era Star Trek, complete with universal translator, creating interactive NPCs and environments on the fly, and letting us talk to an always-listening microphone when we want to change which music is playing or what mood we want the lights to provide.
China has gone from "LOL" to "superpower". North Korea got nukes (and if they can do it, anywhere bigger than a tiny island or a city-state can do it). Pakistan, which has nukes, had a military coup in 1999; 9/11 was a pretty big event that shifted American international behaviour, but the choices the US government made in Iraq and Afghanistan were sufficiently tone-deaf to the locals it likely played a significant role in the rise of ISIS.
There's an argument that agriculture and predominantly rice cultivation in China already did this. And that the Anthropocene started ~10,000 years ago and prevented CO2 in the atmosphere from falling to 150 ppm CO2 or so, and prevented the formation of continental shield glaciers by now. Since the end of the last ice age and Earth's axial tilt reaching its maximum 10,000 years ago, the Earth's axial tilt has been decreasing and that should have started a cooling trend towards the onset of a glacial period. Instead, we've had pretty consistent temperatures for the past 10,000 years.
[0] https://en.wikipedia.org/wiki/Stefan_Rahmstorf
[1] https://www.youtube.com/watch?v=ZHNNW8c_FaA "One of the most ominous risks for Europe is that of a major change in Atlantic ocean currents. Recent science suggests it has been greatly underestimated in the past -including by me, having worked on it for over 30 years. Here Prof. Rahmstorf presents his keynote in Vilnius in May 2024."
Their primary example is the AMOC.
" “All the negative side effects of anthropogenic climate change, they will still continue to go on, like more heat waves, more droughts, more flooding,” he told CNN. “Then if you also have on top of that an AMOC collapse … the climate will become even more distorted.”
...
The impacts of an AMOC collapse would leave parts of the world unrecognizable.
In the decades after a collapse, Arctic ice would start creeping south, and after 100 years, would extend all the way down to the southern coast of England. Europe’s average temperature would plunge, as would North America’s – including parts of the US. The Amazon rainforest would see a complete reversal in its seasons; the current dry season would become the rainy months, and vice versa. "
[0] https://lite.cnn.com/2024/08/02/climate/atlantic-circulation...
So apparently the “mean” is different from the 50% point. Would that be the median? And wouldn’t that be the more interesting point?