The State of the Climate 2026 | Ep242: Zeke Hausfather

Bryony Worthington

The U.S. has gone back to coal.

Zeke Hausfather

The U.S. is keeping coal plants open. You're not allowed to close them anymore.

BW

And it’s burning more coal. Like 2025, the U.S. burnt more coal. 

ZH

That's largely because gas prices increased and there's a colder winter. There's a few other factors there that I don't think are going to last. But certainly this idea of if governments actively subsidise coal, prevent coal plants from being shut down, like we could end up in a higher emission scenario than we think we're heading to under current policy today because we can reverse current policy. Now, I tend to think the 21st century is long and climate isn't going away as an issue and we're much more likely to strengthen policy over time, but it's not outside of the realm of possibility. But this SSP-3 scenario is interesting because it's specifically described as a world of resurgent nationalism and isolationism with fraying international ties, less priority on technological research and development. And if that becomes the dominant framing of the 21st century, if we're sort of all isolationist, regional conflict focused, not investing in the future as much as we used to, we could end up in a much worse outcome.

BW

Hello, I'm Bryony Worthington and this is Cleaning Up. My guest this week is Zeke Hausfather, an American-based climate scientist who has been navigating the complexities of the climate for decades. He's become known for being able to cut through the complexities and communicate clearly what's happening to our climate and why. I was delighted to sit down with Zeke today as the 2025 data of just how warm last year was, was released. Please join me in welcoming Zeke Hausfather to Cleaning Up. 

BW

Zeke, thank you so much for joining me here in Oakland to discuss climate science. Before we get going, could I ask you to introduce yourself, please, in your own words?

ZH

Sure. So I'm Dr. Zeke Hausfather. I'm a climate scientist, a lead author of the upcoming IPCC Seventh Assessment Report and a previous author of the U.S. National Climate Assessment. But I'm a somewhat unusual scientist in that my role has primarily not been in academia. I came out of the cleantech startup world, founded a company out of grad school, did that for 10 years, eventually went back to get my PhD. And now my main job is actually as the climate research lead for Stripe, which is a big financial technology company where I'm helping deploy a billion dollars on permanent carbon removal technologies. I also work as a research scientist with Berkeley Earth, which produces one of the main global surface temperature records.

BW

Amazing. I don't really know how you fit all of that in, but we're going to talk about all of those elements. But I wanted to start because we actually happen to be meeting on the day in which, I mean, you've already had a busy morning when the 2025 climate science data has been issued. And there's been a lot of commentary already in the media about what 2025 was like relative to recent years in terms of warming trends. And I wanted to give you an opportunity to just update us on, well, what was 2025 like and how did that compare to sort of recent years?

ZH

So 2025 was quite warm. Despite being a La Nina year, which is a variation of ocean temperatures in the tropical Pacific that tends to lead to cooler global temperatures, we ended up more or less tying as the second warmest on record with 2023, across the eight different groups of international researchers that produce global temperature records. So it was quite warm and in many ways a bit warmer than many of us had anticipated at the start of the year.

BW

And that's, I mean, 2023 for me was quite, I mean, that was an anomalous year, wasn't it? So to tie with that year, it still indicates that things are running quite hot, right?

ZH

Yeah, we're quite close to 1.5 degrees this year, even though we're not over it like we were last year. I think the average of all the groups came in at about 1.44°C above pre-industrial levels this year. And, you know, with an El Nino growing in 2026, we expect that 2027 will probably be the next warmest year on record.

BW

As in it'll break records in 27?

ZH

Yeah, so this year we're starting in La Nina conditions again. But most of the models are expecting an El Nino to develop, which will push global temperatures up, you know, obviously on top of the long-term human-driven warming trend.

BW

Yeah, I think that's all quite concerning. And it's also in the context of, you know, already this year we've seen the US pull out of the IPCC, the International Panel on Climate Change, which is the body that you collaborate with, that you work with. For you personally, does the US pulling out affect you? Or are you sufficiently independent that you can just carry on doing what you're doing?

ZH

So the US not participating in the IPCC has already had a pretty big effect. Like in previous years, NOAA, the National Oceanic and Atmospheric Administration, was the organisation that had the most scientists participating in the IPCC of any in the world. This year there is one, or this time around there's one, and that person only could do it because they're retiring at the end of the year and didn't really care that it wasn't allowed. So there's pretty much no US government scientists participating in the IPCC already. And in fact, the person who is supposed to nominate US scientists, because governments are normally in charge of nominating scientists to participate in these reports, got fired the week before the nomination forms went out. So if you even tried to apply to be a nominee from the US government for the IPCC, the email bounced back.

BW

Wow. 

ZH

So it's a bit of a mess. That being said, there are still about 60 odd US scientists participating entirely from academic institutes and private sector organisations. And there was a pretty big push to get philanthropic money to support the US scientists participating. Now, we're all volunteers. No one gets paid, except for a couple of people in the technical support unit to write the IPCC reports. But because we're doing it in our free time to make it work, we need travel expenses, lodging expenses, things like that covered so we can meet together and work on these reports.

BW

And has philanthropy stepped in? 

ZH

Philanthropy has stepped in, yeah. 

BW

OK, that's good. And at a state level, I mean, California still has some of the best research institutions in the world on climate, Colorado. There are places where at a state level… How can the states do anything to kind of compensate for the federal pulling out? Or are they just not allowed?

ZH

I mean, states can certainly do a lot to support science. It's not quite at the level of the federal government in terms of budget, but certainly the university systems have been quite supportive of climate science more broadly, though it does vary a lot state to state, right? It's a lot harder these days to be a climate scientist in Texas where they're policing what's on the syllabus now to make sure it's not woke, which I guess is the new way of saying it's politically correct for the current administration.

BW

Or you could say it's asleep, right? 

ZH

Whereas in California, there's a lot more freedom. So even the private sector and universities are having some impact, but it's not as much.

BW

And then just sticking with the kind of change of, the really sudden change in the US policy, there's been this kind of active attempt by the White House, particularly to pull funding from US institutions. How serious is that? And how much is perhaps Congress still able to hold the line? And, you know, where money's been actually allocated to science is what's happening.

ZH

So it's a bit of a confusing mess at the moment. You know, historically, Congress and constitutionally, Congress is the one who has the power of the purse. They get to decide how federal money is spent. And they have been pretty reluctant to make big cuts to the sciences more broadly in the US, including not just climate, but healthcare and all these other areas. They have proposed budgets that have slight reductions, 5% or so for some sectors, but they're not huge. Whereas the Trump administration has proposed a 50% reduction in funding for the National Science Foundation, which is the biggest funder of science in the US.

How this all settles out is really going to depend. You know, historically, the way it's worked in the previous Trump administration, for example, is that they propose a budget. Everyone sort of pretends to plan for it. But in reality, Congress passes what they pass and not nothing much actually changes. This time around, they've sort of set a precedent of not spending money that Congress has allocated. And so the worry is that even if Congress gives full funding to the National Science Foundation, maybe the administration will just hold half of it back or refuse to fund things that they don't like, like work on vaccines or work on climate science or these other areas that for one reason or another have become politically charged.

BW

And what happens in that case where Congress issues a direct statement? This is where we want the money to go, White House… Well, sorry, this would be the departments, right? Just not cutting the checks. What recourse do you have to justice in those circumstances?

ZH

I mean, it's going to lead to lawsuits.

BW

Lawsuits. Yes, it always leads to lawsuits.

ZH

Yeah. And so far, it's kind of been a mixed bag. You know, the courts have let them do some cuts. They have not let them do other cuts. There's not been a Supreme Court ruling on this yet. So everything's a bit up in the air in terms of what is actually going to happen.

BW

Oh, goodness. Well, OK. So in that context, though, it is encouraging to know that there's at least 60 of you who still can participate in these international fora as individuals, as private institutions. And so let's talk a little bit about this process of how we're navigating, predicting what's going on with the climate, because it's a complex subject. I know it's something you've looked at in a lot of detail. And for our viewers, can you just give a snapshot of how we marry up observational data with the models? Because there's a huge number of models out there trying to predict the future, super hard. And then we've got this observational data that's just today been released for 2025. And we've got that record? How do those two things get kind of reconciled?

ZH

Yeah. So we've been doing climate modelling or what you consider modern climate modelling, at least since 1969, when Syukuro Manabe, who recently got the Nobel Prize for his work and Wetherald published the sort of first modern climate model. They've improved significantly over time from there, right? 

BW

The first one was a sort of back of the envelope? 

ZH

Not quite back of the envelope. It was reasonably complicated, but certainly not like today's models that run on the most powerful supercomputers in the world. But even those early models ended up doing a pretty good job. So back in 2020, I worked with Gavin Schmidt at NASA and a number of other scientists to go back and collect all the early climate model predictions that had been made, you know, starting in 1970 with the first one, and going through the mid 2000s when sort of the IPCC really took over a lot of the modelling coordination and work. And we found that they did quite well. 

Of the 18 models that we assessed, 11 of them were pretty much spot on with observations. indistinguishable from the uncertainty in the observations of what happened after. And the ones that weren't, you know, about half of them were a bit too warm and half were a bit too cold. And then what we also found is when you correct for the amount of CO2 and other forcings, as we call them, being emitted, then 14 of those 18 were pretty much spot on. Because you could have the best climate model in the world in 1970 but if you assumed — we didn't, but pretend you did — but if you assumed we doubled the amount of CO2 in the atmosphere by the year 2000, your model would be off. And knowing how much humans are going to emit is more of a question of economics and politics than a question of physics. 

But how we marry them up is, you know, models are physics based computations. We're not, this isn't economics. We're not fitting curves to the data. In fact, in most cases, we try to hold back the actual observations so they can be used to evaluate the models and not train the models directly or indirectly on the observations.

So once you have a model run, you spin it up and say the year 1850, you feed it the observed concentrations of CO2 and methane and nitrous oxide and the other greenhouse gases in the atmosphere, and then you run it up through to the present. And then you give it some future scenario of expected emissions. And then you run the model with that and see what happens in the future. And so you can then compare the model to the observations and say something about, does it reliably follow what actually happened in the real world, both in terms of global temperatures, but also a million other things, right? Is it getting cloud cover right? Is it getting snow cover right? Is it getting precipitation right? You know, there's a million complicated parts of the climate system that you can use to evaluate models against.

BW

Yeah, it's almost making my head hurt just trying to think of all the parameters that you've got to just hold into these equations. So in 2020, you did this reconciliation of how well have the models done in terms of predicting the future and seemed like they didn't do a bad job. But since 2020, we've had some pretty anomalous years, like 2023 — the second half of 2023 — being the classic one where everyone went, ‘oh, hang on, something weird's happening.’ And have you been able to, is there a plan to do that exercise again, just to see like with this last five years of data now, which of the models are performing best?

ZH

Yeah, so we've been looking at that already. The latest generation of models that accompanied the most recent IPCC report generally performed pretty well. If you look at the rate of warming over the last 15 years, for example, it's slightly above, but close to the middle of the range of model projections. But if you look at some of the earlier generations, the ones that came out in 2014 or 2008, those ones predicted a bit less warming on average than we've seen over the last 15 years.

BW

But what about the last five? I suppose the reason I'm interested in that is because it's the period since the last IPCC. It's the period where sea ice cover seems to have shrunk in an extreme way. We've had these extreme heat events and it did all seem as if things were tracking on quite a bad path. That all could level off or even go and we could start to see cooler times. But which of the models has been able, were there any of the models who saw this coming?

ZH

It's hard to use a period as short as five years to tell us all that much. What we can say is that models generally expect some acceleration in the rate of warming, even in a current policy type scenario. And the primary reason for that isn't that greenhouse gas emissions are increasing more than expected. In fact, if anything, our emissions of CO2, the main greenhouse gas, have plateaued over the last decade or 15 years. But rather what's happening is that we haven't reduced greenhouse gas emissions yet. And at the same time, we are very rapidly reducing emissions of planet cooling aerosols like sulphur dioxide. So sulphur dioxide is a byproduct of burning fossil fuels primarily. And it has been masking about 0.5°C of warming that would have otherwise occurred from our emissions of CO2 and other greenhouse gases.

BW

That feels a lot. 0.5°C. 

ZH

Half a degree is big. 

BW

So if we had no aerosols in the atmosphere right now, or at least no man-made aerosols, warming would be closer to 2°C today than 1.5°C. And this is something that I think perhaps our models, you tell me, but perhaps they thought it was a less dynamic parameter, perhaps. Because what seems to have happened is in the 1980s, 90s, we started stripping sulphur dioxide out of fossil fuels, especially out of coal and heavy fuel oil, because it was causing acid rain, right? So this was a, you know, multi-decade problem. We kind of brought in regulations, stripped that out. And then, oh dear, we suddenly realised, ‘ah, that may have made the climate problem worse.’ We may have solved acid rain, hopefully saved a lot of damage. But we've inadvertently now revealed that we were once cooling the planet accidentally. Now that mask, that sunscreen's been removed. We've kind of got a termination shock, have we? A little bit of a termination shock.

ZH

Now, there's a pretty big range across models of how big in a cooling effect we think these aerosols are having. You know, 0.5°C is the average across all the models. It really ranges from as little as 0.2°C to 1.2°C. You know, at the very high end, that's a huge termination shock. At the very low end, it's not that much. And that's this question of how much aerosol cooling there is is also very closely tied to this question of how sensitive the climate is to our emissions.

One of the challenges with climate modelling in general is that the Earth is really complicated. We know pretty well if you double the amount of CO2 in the atmosphere, how much additional heat that CO2 by itself will trap. And if you do that and you run that experiment, you get the world will warm about one degree centigrade if you double CO2. Not that much, actually. The problem is, when you double that CO2, you kick off a whole bunch of other processes. The big one is the water vapour feedback.

So every degree you raise global temperatures, you can hold about 7% more water vapour in the atmosphere. Water vapour is itself a very strong greenhouse gas. In fact, the strongest greenhouse gas, the biggest part of the overall greenhouse effect. But water vapour can't be a forcing because it only lasts for, you know, a couple of days before it rains out. Usually. But as you warm the atmosphere with something else, like long-lived CO2 concentrations, it increases the amount of water vapour as a feedback to that warming. And that adds another degree on top of that. And then on top of that, there's changes in clouds. You know, more low-lying clouds can reflect more light back to space and cool the planet.

More high clouds or less low clouds can lead to more warming. There's changes in snow and ice cover and sea ice. And when you put all of these together and try to figure out how much warming will actually get, you get a pretty big range. So in the most recent IPCC report, we said that if you double the amount of CO2 in the atmosphere from 280 parts per million pre-industrial to 560 parts per million, which is roughly where we're headed right now by the end of the century, you'll probably end up with somewhere between 2.5 degrees centigrade warming and 4 degrees centigrade warming. But that's only a likely range that says there's a two-thirds chance it's in that range. We also said there's a 90% chance, you know, 9 in 10, that it's somewhere between 2°C and 5°C warming. But even that is not fully constrained. It could well be more than 5°C warming, and that would be really catastrophic.

BW

And just to put this in context, because, you know, for some people, 5°C might not sound that much. It's the difference between a sort of spring morning and a summer morning… Or like a 5°C range in weather terms isn't a lot. But isn't that the difference between us and the last ice age?

ZH

Roughly, yeah. Roughly. The last ice age was about 6°C cooler for the global average. And this brings up...

BW

So essentially we're saying, turn up the thermostat by the same amount that it was cool during the ice age, where we had mile high ice sheets across most of the Northern Hemisphere.

ZH

It's a big shift. It's a big shift. And to be clear, we're not going to see 5°C degrees warming by 2100, even if climate sensitivity is on the very high end under a current policy scenario. Now, if we decided...

BW

Well, you say that, but, you know, there's a 10% tail risk here.

ZH

Well, when we talk about climate sensitivity, we're talking about at equilibrium, which essentially means if you increase, or if you double the amount of CO2 in the atmosphere and you just hold it constant for 300-400 years, how much does the system warm up? Thankfully, the world has oceans and the oceans are really good at absorbing heat. And so the actual warming at the end of the century is only two thirds or so of the equilibrium warming. That being said, if we keep CO2 at that level in the atmosphere and wait a few more centuries, we're going to keep warming up.

BW

It's perhaps worth double clicking a little bit on that cloud point, the sort of feedback mechanism of that, because you've shown me a very compelling chart which shows in the last 15 years, cloud cover has really fallen, hasn't it? And it's a particular type of cloud, it's the kind of reflective clouds, those nice fluffy clouds that are near to the surface of the planet, which reflect the sunlight. There's been a, they've diminished quite noticeably.

ZH

We've seen an unprecedented decline in the reflectivity of low lying clouds over the oceans in the past 15 years. We only have records going back to 1970 or so when we first launched satellites that could measure cloud cover. But certainly in that period, there's nothing like what we're seeing today. And so there's a bit of a debate right now in the scientific community as to what exactly is driving that change. It does seem to be one of the main contributors to this more rapid warming we're seeing. And there's sort of three possible explanations that we've settled on. One, and I think this is definitely part of the equation, is the reduction in aerosol emissions. So the sulphur dioxide, little particulates like that, can both directly scatter light back to space because they're very reflective, but they can also serve as little bits around which clouds form. So for clouds to effectively form, you need a certain amount of haze in the air. If you have super, super clear skies and with no little particles, it's much more difficult for cloud formation to occur. And so if you get a bit more air pollution, you end up with a bit more clouds. And so reducing sulphur has reduced the cloud formation due to that. The other potential explanation, which is the more worrying one for me at least, is that what we're seeing is at least in part a response to warming. So in some climate models, the majority show some reduction in low-lying clouds associated with a warmer world. But the models that are the most sensitive to our CO2 emissions that show the most warming for doubling CO2 show the biggest reduction in low-lying clouds. So if some of what we're seeing is a stronger cloud feedback, that could imply a higher climate sensitivity. And then the third possibility is we only have a record back to 1970, right? Maybe there's other periods if we had a record back to 1900 where we'd see temporary declines in cloud cover for a decade or two. We don't think it's that. There's no clear physical mechanism that would explain why internal variability could do that, but it's hard to fully exclude it until we have more data.

BW

I can remember back reading James Lovelock's kind of hypothesis about there would be some kind of natural process which would help us cool the Earth down, you know, the kind of claw hypothesis, which was this idea that as temperatures rose, we'd see more microbial activity, we'd see more dimethyl sulphide created, and that would act as a kind of, that would create the cloud cover that would help to protect us. Is there an element in which there's a biospheric component to this cloud formation as well, which might be being affected by the heat?

ZH

I mean, to an extent, though, the changes in, say, microbes that produce dimethyl sulphide are much slower than the changes that we're seeing due to human activity. Like over geologic time, there's lots of Lovelockian-style feedback, right? My favourite one is actually the silicate weathering feedback. So, you know, if you have a big set of volcanic eruptions like the Siberian Traps going off, you can get significant global warming associated with that. You know, very short term, you have cooling because of aerosols, but longer term, volcanoes over geologic time commit a lot of CO2. That leads to warmer states for millions of years, but that also leads to more evaporation, more rainfall, and that rain then erodes more rocks. The same rocks that volcanoes spit out that they strip the CO2 out of to emit to the atmosphere can absorb CO2 as they get weathered. And so over geologic time, it's this balance of volcanism and silicate weathering that really governs the temperature of the climate. Again, there's lots of these examples of things like that that work over hundreds of thousands or millions of years. But what we're seeing right now, this dramatic increase in global temperatures, already close to 1.5°C in just the last 50 years. Well, 1°C in the last 50 years, close to 1.5°C in the last 150 years is really unprecedented in its rate from anything we've seen in the geologic period. And so the normal sorts of processes that might be thermostatic, that might hold temperatures back down, are much slower to respond.

BW

I suppose my question was really just to understand the known unknowns in this complexity that we're describing here, that you've got to factor in the biospheric response, the volcanoes erupting, the solar cycles, the forcing factor we're introducing with all of this man-made changes in both pollution, traditional pollution, and the CO2 and methane. And it's a lot.

ZH

It's a lot. And that's why we have really complicated models to try to figure it all out. And the nice thing is you can then use those models to say, ‘well, what if this didn't happen?’ What if we only had volcanoes and solar and no change in greenhouse gases? The cool thing about that experiment is you actually get a slight cooling over the last 50 years because solar activity has declined a little bit.

BW

Right. So then you can use that to sort of push back against some of the, well, there's a whole topic there about how people are interpreting their own version of science and coming up with theories that it's all just sunspot activity and you guys are all crazy and you're in some part of left-wing conspiracy theory that's just about keeping your job. And there's that whole environment that we're living in now where information has become so contested, right?

ZH

Well, you can just Google your own reality if you want to, right? There's no lack of information to confirm your preconceptions and most people have a hard time differentiating what sounds compelling from what is, you know, actually scientifically rigorous or accurate. 

BW

What can we do to kind of counteract that? I mean, you and Berkeley Earth and others are, you know, you're out there trying to interpret the data, trying to get people engaged. But is there more that we could be doing to try and..,?

ZH

I mean, it's challenging, especially because we live in a world where expertise is not granted the privilege it was, say 10, 15 years ago. Like it used to be, if people wanted to figure out where science stands, they'd ask the National Academy of Sciences to write a report. You know, it was created by Abraham Lincoln to specifically advise Congress on settling matters of science.

And you ask them to write a report about climate change and they'll come back with something very similar to the IPCC report that says it's real, it's caused by humans. It's going to be bad if we don't do something about it. Nowadays, people are a lot less likely to trust those sorts of things. And so I do think it's important for more scientists to be proactive communicators. Historically, communication was frowned upon in the sciences.

Carl Sagan very famously did not get inducted to the National Academy of Sciences. Jacques Cousteau was similarly shunned as an entertainer, even though he did a lot of important scientific work. I think that has changed a little bit now where there is a bit more in traditional academic settings, at least respect for scientists spending some of their time being communicators. At the same time, I don't think if we just explain the science better, it's going to solve this, right? There still is this very fragmented information ecosystem. There still is, you know, partisanship poisoning people's views. One thing, and maybe this is my Silicon Valley bias showing through that I am curious about is as we move toward a world where AI tools are increasingly used to answer questions that people have about information. In some ways, they are designed to serve as consensus machines.

They're giving a single best one-shot answer to whatever complicated question you're asking them. And at least right now, they're actually pretty good on climate science. Like, even if you ask Grok, which is problematic as far as AIs go, it'll give you an answer to most climate science questions that's right and pretty in line with the latest research. That's not to say that people can't put their thumb on the scales of those things as well. But I also wonder if companies who are trying to design the best AI tools to sell to Fortune 500 companies if they're going to want to, you know, manipulate them too much. Because I feel like if you design an AI to give bad, inaccurate answers on climate science, it's probably going to give bad and inaccurate answers on a lot of other things that people want true answers on, too.

BW

My experience is that if I tell my 15-year-old son something, he will instantly go check it on ChatGPT. And if ChatGPT agrees, he'll be, ‘OK, fine, you're right, mom.’ But, you know, there's a sort of trust being put into these things because it is the wisdom of the crowd, as you said.

ZH

And some of the companies, the big AI labs, have been more proactive in working with scientists to actually design things like evaluation tools to rate how well models get answers to scientific questions accurate. And so there is some work being done on making sure that the models are performing well here as well.

BW

We've talked a little bit here about all these complicated inputs. And one of the things we talk a lot about on this show is the CO2, you know, the profile of emissions over time. And regular listeners to this show will know that Michael has had a particular discussion that he's kickstarted around, I think he and others, that some of the scenarios that were being put into the models appeared to be somewhat out of the realms of possibility, like the representative concentrations pathway 8.5 or RCP 8.5 was a scenario that appeared to be reliant on more coal than is in the world being burnt to get to that kind of scenario. And I know you weighed into this conversation. And, you know, Michael's absolutely right in predicting that if you force models in unbelievable ways, you're almost discrediting the whole thing. And it's a big risk to do that. So from your perspective, you know, what was going on in the IPCC with these scenarios, you know, and how much damage? Well, why did it first? First of all, why did it occur that we had these slightly implausible profiles? And then, you know, has Michael been correct? Has it been misused?

ZH

So there's a bit of a long story for how we ended up with RCP 8.5 and what ended up happening with it.

BW

And we should just explain what these are. These are scenarios that try to give us a sense of what could happen in the future.

ZH

So to back up, when we're trying to project how much the world might warm in the future, you need to start with emissions. How much are humans actually going to put in the atmosphere? Then you go from that to concentrations, how much CO2 and other greenhouse gases remain in the atmosphere.

And that's modulated by carbon cycle feedbacks and various processes of land and the ocean over time. The sinks, which in turn, are affected by warming. So there are dynamic processes. And then you go from those concentrations and what we call forcings to the actual warming, which is where you get all these different feedbacks in the climate system, how snow and ice respond, how water vapour responds, etcetera, how clouds change. But none of that's going to work if the emissions are wrong. And so getting the emissions right is important, but it's also just a really difficult question to answer.

If you hop in a time machine back to the year 2000 and you say, what are the plausible emission trajectories for the next 100 years? It's going to be notably wider than it is today, right? There have been some credible cases that there are certain fundamental limits.

So Justin Ritchie, a researcher in Canada who did a lot of really good work initially on coal limitations, pointed this out in a paper in 2017. But even there, it still gets down to these questions of what is economically viable, right? And if you looked back in the year 2000 around what economically viable natural gas reserves would be, you'd also get a very wrong answer because we didn't have fracking back then. So purely in terms of the amount of fossil fuels on Earth, there's not particularly hard limits there. Where you get into limits is what will happen with technologies, how expensive or cheap will they become, both as we deplete low-hanging fruit in terms of fossil reservoirs or develop alternatives like solar or wind or nuclear. And that's, I think, where we've really seen the biggest evolution in the last 20-25 years.

So back in the year 2000, actually, there was a set of scenarios developed by the, as part of the IPCC process called the SRAS scenarios, which stands for the Special Report on Emissions Scenarios… Scenarios. We're terrible with acronyms. And that set of scenarios essentially looked at a bunch of different worlds that were all sort of plausible, no policy outcomes. And by that, I mean that like technology would still advance, other things would change, but there wasn't the assumption that governments around the world would institute a particular carbon price or a particular set of policies. And those ranged on the low end from a little bit below three degrees warming to the high end, close to four and a half degrees warming. Those scenarios had grown increasingly outdated by the late 2010s. Or sorry, the late 2000s. And there was a desire to create a set of new scenarios that would ultimately become the SSPs or the Shared Socioeconomic Pathways.

BW

Shared Socioeconomic Pathways. So that's the new phrase for what were Representative Concentration Pathways?

ZH

I'm skipping RCPs. I'll get back to RCPs in a second. Back in the late 2000s, there was an effort to develop what would become the SSPs.

But that effort was taking too long and the climate modellers needed a set of emission scenarios to use for the IPCC fourth assessment report. And so they ended up... Or sorry, fifth assessment report. Mixing up my reports. So they didn't have the SSPs done in time for the IPCC fifth assessment report. So they said, okay, let's just pick four scenarios from all of these we've developed that are distinct enough.

They'll give us a range of possible climate futures to model. Because you don't want to model two scenarios that are very similar because you're not really going to get much out of the climate models then and spend all this time and money running these supercomputer models. One of those, RCP 8.5, was chosen to represent sort of the highest end of possible outcomes. It was the 90th percentile of no policy outcomes in the literature when it was published. On the other hand, RCP 2.6, the low end, was a world where we kept warming below 2°C degrees with a two thirds chance or so. And the other two were in between.

The challenge is that that distinction, that the high-end scenario wasn't intended to be the most likely outcome. It was intended to be the worst case outcome. That got a little lost in translation, in part because all three of the other RCPs used some sort of carbon price in the model that generated them, even though two of the other ones, RCP 6.0 and RCP 4.5, could have been consistent with the middle to low and very low range of no policy baseline outcomes. So this is getting a little wonky. But the short version is that the fact that RCP 8.5 was not business as usual, but rather the worst case outcome in the late 2000s was lost in translation.

BW

But Michael's critique was more that there was just no physical way that you could burn that much carbon in the system at all. That it was like there was more. It was assumed coal burn was just more than the coal reserves.

ZH

More than the economically extractable coral reserves in 2010. Again, it's hard to know how technologies will develop to change the economics of these fields, as we see with fracking. I'm a little less convinced that was a strong constraint.

And I can also see if you look from the vantage point of 2007, 2008, you know, global emissions had been increasing at 3% a year for the last almost two decades. China was building a new coal plant every three days. The idea that the 20th century or 21st century could be dominated by coal didn't seem as far-fetched. Now, again, this was supposed to be the most extreme case of the fossil fuel dominated 21st century scenarios. But I'm less convinced it was completely out of the realm of possibility. Certainly since 2010, it has very much moved out of the realm of possibility. And that's largely due to the dramatic falling cost we've seen of clean energy. Like the idea that we're going to be burning three to five times more coal by 2100 wouldn't pass the laugh test today.

BW

Exactly. But different factors are now kind of coming into force. And I guess this kind of socioeconomic element of the IPCC work is super hard, right? Because you're trying to not only put the physical complexities of the planet, but then you've got the socioeconomic complexities. And if you just look at the world today, you've got two completely competing arguments going on where China is basically showing that you can electrify almost everything. Electric tech has this really rapid learning curve that brings costs down really rapidly. And so suddenly, you know, we've got a new energy vector arriving that's affordable and attractive and lots of countries are going to adopt it. And then, you've got the kind of U.S. going back to a petrostate, “drill, baby drill.” And so the world isn't even, you can't even characterise the world effectively now with one description. You've got to take two hemispheres doing two different things and then try and work out… I mean, and I guess I worry that the U.S. is kind of isolationism, or the America first, or America alone, actually, you could describe it as now. Is that a kind of precursor to what the world's going to be like when we get more and more stresses from climate change, as we get more unrest, more immigration, all these factors?

ZH

So one of the goals of developing these socioeconomic pathways was to explore what these different potential futures might look like, not just in terms of emissions. So that's something that comes out of some of these modelling assumptions, but also in terms of adaptive capacity.

You know, one of the neater things, I think, that came out of the SSP work is that the worst scenario in terms of climate damage isn't the one with the most warming. You know, there's this. 

BW

It's the one where we're poorest.

ZH

It's the one that has a lot of warming, but we're also very poor and very unequal. Some countries can adapt to it. While large swaths of the world are still living on potentially a dollar a day and can't. And that's the most damaging outcome. And so understanding the socioeconomics is important because it tells us how bad these climate impacts are going to be independent or at least at the same level of warming. You could have very different impacts because of different adaptive capacities. And there is one of the SSPs that I think is a little close to home these days. It's called SSP3. You know, its baseline scenario probably has too much coal. We can quibble about that.

BW

But the US has gone back to coal. The US is keeping coal plants open.

ZH

You're not allowed to close them anymore.

BW

And it’s burning more coal. Like 2025, the US burnt more coal.

ZH

That's largely because gas prices increased and there's a colder winter. There's a few other factors there that I don't think are going to last. But certainly this idea of if governments actively subsidise coal, prevent coal plants from being shut down. Like we could end up in a higher emission scenario than we think we're heading to under current policy today because we can reverse current policy. Now, I tend to think the 21st century is long and climate isn't going away as an issue. And we're much more likely to strengthen policy over time than reverse it. But it's not outside of the realm of possibility. But this SSP3 scenario is interesting because it's specifically described as a world of resurgent nationalism and isolationism with fraying international ties, less priority on technological research and development, like very much the type of thing that's been happening in the US, particularly over the last year, but even over the last decade. And if that becomes the dominant framing of the 21st century, if we're sort of all isolationist, regional conflict focused, not investing in the future as much as we used to, we could end up in a much worse outcome.

BW

Yeah. It's really hard at this point in time to know exactly the way it goes. What you have got, though, is the rest of the world, apart from the US, suddenly a little bit freaked out that this kind of consensus, the world order, the orderly way of doing things is breaking down. And that is causing people to think more about energy security or what can I do at home that doesn't expose me to all of this? And maybe that actually does push the rest of the world into a far more electricity based, clean energy based system, purely because actually this perturbance in the world order, it may spread, but it also may just cause everyone else to galvanise and do things differently.

ZH

Yeah. The great thing about a lot of clean energy technology is that they're domestic, right? You're producing your own electricity. You're not relying on imports of coal or oil or gas from other countries.

BW

Yeah.

ZH

You're still relying on imports of clean energy technologies from China these days, which is its own set of geopolitical challenges. But there definitely has been a bifurcation where the US has embraced its role as a petrostate, while China has embraced its role as an incubator of the technologies of what I think at least are going to be the 21st century, including wind, solar, advanced nuclear, electrification of vehicles and transport, all these other things.

Michael Liebreich

Cleaning Up is supported by its Leadership Circle. The members are Actis, Alcazar Energy, Arup, Cygnum Capital, Davidson Kempner, EcoPragma Capital, EDP, Eurelectric, the Gilardini Foundation, KKR, National Grid, Octopus Energy, Quadrature Climate Foundation, SDCL and Wärtsilä. For more information on the Leadership Circle, please visit cleaningup.live. If you're enjoying this episode, please hit like, leave a comment, and also recommend it to friends, family, colleagues and absolutely everyone. To browse the archive of over 200 past episodes, and also to subscribe to our free newsletter, visit cleaningup.live.

BW

So we've talked a lot here about your work to communicate climate change and you're working with your colleagues at Berkeley Earth and the work you're doing at IPCC and I've been following your work and you're excellent at helping make this very clear to understand and so thank you for that. But you also wrote an essay last year or an article actually in the New York Times where you started to talk about what could we have to do in terms of geoengineering to buy us some time? And I guess this stems from the fact that there are so many unknowns and risks that we could track in a negative way or in a worst case scenario. So tell us a little bit about what stimulated you to write that article and what the article talks about in terms of are there things we can do to buy us some time. It’s called geoengineering but they're essentially extreme adaptations, aren't they?

ZH

So David Keith, who's a longtime researcher working in this field and myself wrote this in the Times, where we weren't necessarily arguing that we should do geoengineering, you know, whether or not we take measures to temporarily cool the planet is ultimately a decision for policymakers given the costs and the benefits of doing so. Our argument was that if we were or if policymakers were to ever decide to do so, we need a clear upper bound for how much we deploy. And the reason is that at least most geoengineering techniques we're talking about, putting sulphur dioxide high in the atmosphere, mimicking volcanoes, don't actually solve the underlying problem, right?

If I emit a tonne of CO2 to the atmosphere today, that tonne of CO2 is going to keep warming the atmosphere at pretty much current rates for about a thousand years, maybe a little less by the end of that period, but it's remarkably long, the lifetime of warming from CO2. So if I'm putting sulphur dioxide in the stratosphere to counterbalance that, it's just like dropping a bigger and bigger ice cube in the ocean every year to solve climate change once and for all, as Futurama famously put it in that skit many years ago. It's a Band-Aid to treat the symptoms of the problem without curing the underlying disease.

There are times when you need a Band-Aid, but it only makes sense in the context of taking strong action to cure the disease. If you're just putting more and more Band-Aids on yourself over time, you're quickly going to run into bigger and bigger problems because all of these techniques that have been proposed into engineering have side effects. If you put too much sulphur in the upper atmosphere, you start harming the ozone layer in a pretty serious way, right?

You're going to change rainfall patterns, potentially affect monsoons, though there's a debate about that, right? The more you alter the climate in this sort of way, the bigger unintended side effects you're going to have. And so what we suggested in that piece is that if society ever decides to go down that path, we need to strictly limit it because the only way to ever stop doing it is to invest a huge amount of money in carbon removal to take all the carbon we've added back out of the atmosphere to cool things down to where they were before we started doing it in the first place.

And if we put too much of it up there, if we are relying too much on geoengineering, it's going to be too expensive to ever stop doing it. We're going to become addicted to it in a way that would cost trillions upon trillions of dollars to reduce temperatures enough to stop without a large termination shock. And so what we suggested is a logical way to set an upper bound on the amount you would ever deploy would be, let's just replace the cooling we get today from our accidental geoengineering.

The 75 million tonnes of sulphur we're putting into the lower atmosphere every year from burning fossil fuels is masking 0.5°C of warming today. That's already gone down. It's going to keep going down as we phase out of fossil fuels.

And so if we just replace that, slowly ramping things up, it could provide a reasonable way bound for a future deployment. Now, again, hopefully we won't ever need to, right? Hopefully the world will get its act together, start cutting emissions rapidly. We won't end up with warming so dangerous that we need to mask its symptoms by intervening in another way in the climate system. But I think that the world is not heading in the right direction on this right now. And I think there could emerge a point where the impacts of climate change get so bad that the risks would be more attractive to policymakers.

And certainly I've seen a lot of people who were previously sceptical about topics like geoengineering, who are very worried about climate impacts. Some folks on the more doomer side perhaps have become more open to this idea. Because if you think that climate change is going to cause the end of the world, you're probably willing to try anything that's going to avoid the end of the world.

I don't think climate change is going to cause the end of the world, but certainly you could imagine a world where climate sensitivity is really, really high, or there's tipping points in the climate system that we're not fully aware of today. And we realise we're very close to them. Might make sense to do some other intervention to put a temporary break on warming. While we figure things out.

BW

And I'm definitely one of the people in that camp who's kind of become more and more open to it, as I've, I suppose, felt, well, two things really kind of helped me to think we should certainly research this, right, that we should definitely not rule this out. And one is that ultimately, those sorts of interventions are relatively quick to spin up to scale, like the sort of physical amount of materials you need to move to kind of get a kind of sunshade effect from these, the spraying of these particles to slightly increase the reflectivity is feasible, we hope. We still have a lot of unknowns there, but, secondly, that it is reversible.

Because as you described, you know, the CO2 problem is wicked because it stays around for a thousand years, right? That is the fundamental problem we face is that once it's up there, it's super hard to get it back out again. Whereas the spraying or, you know, this kind of making the world a bit more glittery is reversible, because you stop doing it, it stops. So, you know, that feels like an experiment where the regrets are less than the current unwitting experiment, where we're just throwing CO2 into the atmosphere and hoping, right?

ZH

Yeah, though, I think the ease of doing it is a double-edged sword, right? You know, if people realise that they can effectively stop climate change temporarily for, you know, less than $10 billion a year, it's going to be a lot harder to convince the world to spend, you know, $2 trillion a year to actually reduce emissions.

BW

Yeah, I've heard that argument a lot, but two things, A, I don't think it'll cost 10 billion, I think it's cost orders of magnitude probably more than that, you know, it's just, you're going to need lots of R&D to kind of work out how to do this correctly. I don't think you can just take a 747 and start spraying this, so there's that. But also, if we are in a world where clean energy is the cheapest form of energy, that moral hazard of, ‘oh, but you'll stop, it'll stop all this wonderful work we're doing in policy terms of driving people to do mitigation efforts.’ If the mitigation efforts are now kind of almost business as usual, because it's the cheapest way of generating electrons, and electrons are great at displacing fossil fuels, you know, that moral hazard is decreasing.

ZH

I mean, to an extent, I think we need to be careful about that, because we are somewhat picking the low-hanging fruits of decarbonisation today, right? It is going to become more expensive as we move from electrification of the light vehicle sector, or, you know, conversion of electricity generation to renewables into industrial heat and aviation and to agriculture and to all these other sectors, the economy buildings, where we don't have solutions that are as mature today, or where the costs of prematurely retiring capital is much more expensive. I also think at the end of the day, if we don't include the cost of climate change in the cost of fossil fuels, or the cost of clean energy through subsidies or taxes, we're still going to end up with a lot less mitigation than we need, right? At the end of the day, there is a real economic cost of climate change. And just hoping that technology is going to get cheap enough to solve the problem without ever internalising that cost is a dangerous bet. You know, there is also a world where we have another fracking style revolution and fossil fuels get equally cheap, right? Technology cuts both ways. It doesn't just benefit clean stuff.

BW

Yeah, agreed. But I just think if you look at the real economy of fossil fuels, especially oil and gas, there's a heck of a lot of things that have to happen in a chain, all of which are capital intensive, huge waste in the system, massive thermodynamic inefficiencies.

ZH

I'm bullish on clean energy. I just think that it's a lot easier for me to create a scenario where clean energy bullishness without strong policy support gets us to 2.5°C, but a lot harder to see a world where we get to below 2°C, not to say anything about 1.5°C.

BW

Yeah. So, you know, we're talking about the moral hazard of different options. And some people would say that, compared to carbon removal, actually, the moral hazard of geoengineering is tiny because it's not really competing with mitigation because it's like adaptation plus, whereas carbon dioxide removal is a direct competitor with mitigation because we've made a fungible market of credits, right? You can emit your CO2 that will last for a thousand years, and then you can not cut down a tree. And somehow that's equivalent. 

ZH

Well not cutting down a tree is avoided emissions, it’s not a removal. 

BW

So that's an unfair thing. But, you know, OK, plant a trillion trees. That's the cheap way of doing CDR. 

ZH

Yeah. So if there were a world where CDR were functionally equivalent to mitigation. So, like, instead of preventing a tonne of coal from being burnt, you take enough CO2 to the atmosphere and reform a tonne of coal. In that world, it doesn't matter if you do CDR, if you do mitigation, the atmosphere doesn't care. It has the exact same climate effect, right? And that's sort of the world we're trying to build with my work with Stripe that we can talk about in a minute on. You know, durable, high quality CDR. But the other challenge is that stuff is expensive, right? CDR that takes CO2 out of the atmosphere for as long as burning fossil fuels puts CO2 in the atmosphere costs upwards of $300 a tonne of carbon today. And at $300 a tonne of carbon, 90% or so of all mitigation is below that cost. So, yeah, maybe there's 10% of the economy that it's cheaper to use CDR to compensate for than to actually reduce emissions. But that's only going to be 10% or so. Like, if companies are spending $300 a tonne on CDR instead of cutting their own emissions for less than $100 a tonne, they're lighting money on fire. So, I'm not as worried about that as a functional moral hazard. Where I do think there's challenges, though, is the lower quality side of the CDR market, if you will. You know, there was a period of time where you'd board a Delta flight and they'd announce over the intercom before you took off that this flight is carbon neutral because they had paid someone in Borneo not to cut down a forest. Turns out that if you pay people not to cut down forests, suddenly everyone wants to cut down their forest so they can be paid not to.

BW

Yes, the moral hazard there is obvious.

ZH

If there's a carbon market where there are bullshit offsets that cost $4 a tonne, that's a real moral hazard because it's always going to be cheaper to write a check for $30 million to buy a bunch of $4 a tonne offsets than to actually reduce your own emissions. Whereas, you know, doing direct air capture or ocean alkalinity enhancement or enhanced rock weathering or bioenergy with carbon capture and storage, that's expensive stuff.

BW

You just rattled through a whole bunch of technologies there. Can you just say that slowly? Because these are the things, you know, spending your time trying to sift through to find the ones that they can scale or that could come down in cost, right? So, what are the big groupings that you're looking at?

ZH

Well, to back up a little bit, so the work I'm doing now with Stripe is: we are partnering with a slew of other companies, folks like Google, folks like Shopify, JP Morgan, Salesforce, and a number of others. Microsoft is doing their own thing, but we are working closely with them on, you know, chatting about strategy and things like that. But we have this pool of a billion dollars that these companies collectively contributed, and we're trying to figure out the best way to spend that billion dollars this decade to figure out which of these technologies work and which can scale in the decades to come. Because we know we're going to need a lot of them in any world where we get to net zero emissions, there is going to be that last 10% of emissions that are fiendishly difficult to fully eliminate. And if we overshoot our climate targets, or we do geoengineering, the only way to bring temperatures back down or ever stop geoengineering is to deploy a huge amount of carbon removal.

BW

And there's also the weakening of the sinks, right? Because it's not just like we fail to reduce emissions, man-made emissions, there's also, oh, you know, we lose the Amazon. Yeah, or if climate sensitivity ends up being higher, right?

ZH

You're not going to know that you're heading toward a worse outcome until it's too late to shift it without something like carbon removal to draw emissions back down. So it is a bit of a hedge against that.

BW

But talk me through why companies felt, given what you said, that it's $300 a tonne today, and possibly always going to be quite expensive. What was going on in the minds of these companies when they thought, ‘this is the most important thing I need to do with my billion dollars?’ Because a billion dollars could buy you quite a lot of mitigation in other places.

ZH

Well, to be clear, most of these companies are spending much more money on mitigation, right? The biggest buyers of clean energy in the US now are companies like Google.

BW

Well, but they're missing all their targets. We did an episode about this. And unfortunately, this big AI boom is-

ZH

Yeah, it turns out if you're building data centres willy-nilly, you're going to be missing your targets. I mean, that's a separate debate.

BW

Yeah.

ZH

I think the idea was that it is an area where a relatively modest amount of money can make an outsized impact, because solar is a mature technology, right? If I spend a billion dollars buying solar panels, that's great. I'm going to cut my emissions.

I'm not going to move the needle much at all in the cost of solar panels globally. We've already seen so many doublings of that technology that to really drive down costs further through a single purchase is going to require hundreds of billions of dollars. Whereas when we started doing this work in carbon removal, the industry didn't exist. There was a shipping container in Zurich, and that was the sum total of durable carbon removal that existed at the time. And now we have a whole bunch of different approaches, hundreds of companies, a whole ecosystem out there that's trying these things in the real world and seeing what works. And so the biggest impact of that money is not the tonnes reduced, and all these companies are spending a lot more money producing their own emissions in cheaper ways. But rather, it's investing in this innovation, in driving down the cost of technologies. And again, using this decade to learn what works and what can scale. And that can be more impactful in the future.

BW

Yeah. And from a political perspective, we did an episode with Anand Gopal, a fellow Berkeley resident, who really said something that made me stop and pause. The reason why a politician might reach for the geoengineering rather than the carbon dioxide removal is because of the speed, right? That the CDR is just going to take a long time to kind of scale up.

ZH

Yeah, but CDR actually solves the problem, right? If you reach for geoengineering, you're still going to need the CDR, unless you want to continue doing geoengineering for a millennium, in which case it adds up to more than the cost of doing CDR in the first place. Not to mention the challenge of creating political structures so that those planes are dumping sulphur for a thousand years unchanged, given all the wars and conflicts and everything else that humanity has had over a millennium. You're sending off a time bomb, essentially, for humanity if you get too addicted to geoengineering, because you can never stop. Whereas CDR actually treats the underlying problem in the same way as mitigation.

BW

But isn't there a scenario where, because the biosphere is responding, and you have got natural processes of absorption into the oceans, into the land-based habitats, that a natural form of CDR is coming in that we... ‘Nature-based solutions’ is a phrase I hate, but there is a theory where you buy yourself time for nature to respond to this. So, it's not like we have to run the planes for a thousand years.

ZH

So, the problem there is the oceans. As I mentioned earlier, the reason why we're not potentially going to see 5°C warming by 2100 is that the oceans buffer the rate of warming of the surface. The surface is warmed about 1.5°C or 1.4°C. The deep oceans have only warmed less than a tenth of a degree. And 90 plus percent of all the heat being trapped by greenhouse gases is going into those oceans. So, the oceans are buffering the amount of warming the world experiences. Once we get to a world that's no longer warming, the oceans move from being our friend to our enemy, because they're still heating up. And as they heat up, the surface gets hotter. And so, yes, if you stop emitting CO2, the amount of CO2 in the atmosphere will fall, but the oceans will keep warming. And if you look at climate models, those two things more or less cancel each other out. And so, even once we get to net zero emissions, the planet doesn't cool down for a millennia.

BW

And is there also a risk that because the oceans have not just absorbed the heat, they've absorbed the CO2, that if you start pulling out your CO2 with CDR, the oceans just outgas some of the CO2, because they all get back to an equilibrium, won't they?

ZH

Well, that's already built into the math of this all. So, if I emit a tonne of CO2 today to the atmosphere, about half of that gets absorbed by the land and the ocean. If I take a tonne of CO2 out of the atmosphere, about half of that gets counterbalanced by outgassing the ocean and the biosphere to an extent. Otherwise, CDR would be twice as effective as reducing emissions, which would be great. I'd love that, but it isn't the case. So, it's balanced, right? And so, when we talk about CDR, we're really talking about undoing a tonne of emissions, and that sync behaviour applies to both equally. Okay. So, both are discounted by 50%, okay.

BW

All right. Well, and of those, so you've been looking at with this billion dollars and this kind of war chest of looking at these options, but what are your top two that you think or top three that we could realistically expect to scale in a timeframe that's necessary?

ZH

You're asking me to pick my favourite child.

BW

I know, I know. But, you know, well, just what are the features you're looking for?

ZH

So, I think there's two different areas of focus when I think about this. One is like, what is the most promising in the long-term? And I think those are the ones that could get cheap and don't have big limitations on how much they could scale. And I think there's three technologies there long-term that could really be big. You know, if we needed to actually undo a significant amount of warming, they could do it. Those are direct air capture, which today is the most expensive way to do it, but in a world where- Well, also, it's just collapsing, right?

BW

You know, we're seeing, famously, companies going bust.

ZH

It's hard to do. I mean, there's more effort happening, but it's still one of the technologically challenging options. But if you had unlimited free clean energy, if we developed fusion and it worked really well by like 2080, who knows? There's no lack of geologic storage to put CO2. So, the technology can scale to billions, tens of billions of tonnes per year. The other is ocean alkalinity enhancement. So, as the ocean has gotten more acidic, it's gotten worse at absorbing CO2, and it's weakened that sink. Essentially, at the end of the day, all acid is CO2 emissions. But on the flip side, all base or all alkalinity, as we say, is removal of CO2. So, if you make the ocean less acidic, it absorbs more CO2 from the atmosphere. It's sort of a convenient way, like every mole of alkalinity you add is about a mole of CO2 removed or avoided outgassing in freshwater systems. It's a little lower in the oceans because of salinity, but you can say one-to-one for rough approximations.

BW

Taking a substance that's mildly alkaline, adding it to the oceans to reduce the acidity.

ZH

Yeah. Well, as I said earlier, over geologic time, the biggest way that all the CO2 from volcanoes gets sucked back down is alkaline materials, mostly emitted by volcanoes originally, getting weathered and turning back into bicarbonate that then ends up in the ocean, eventually splits off into carbonate sinks, becomes limestone, gets subducted, it comes back out of the volcano. It's the circle of carbon.

BW

That's why we have the White Cliffs of Dover.

ZH

Exactly. Yeah. This is a huge process, but there's ways to speed that up. You can take alkaline materials like olivine or limestone. In the case of limestone, you have to calcine it to remove the fossil CO2, but then you get this nice calcium oxide material that's super reactive. You put it in the ocean, it absorbs CO2 really quickly, and forms more bicarbonate. And the nice thing about bicarbonate is that there's an enormous amount of it in the ocean. There's something like a thousand times or at least a couple hundred times more carbon in bicarbonate form in the ocean than there is in the atmosphere. And so you could put a huge amount of atmospheric carbon in bicarbonate form in the ocean and only affect the ocean bicarbonate buffer by a fraction of a percent.

BW

Right.

ZH

So you're not really going to change much in the ocean by putting more bicarbonate in there. The challenge with that, of course, is you still need this alkaline rock, feedstock, to start out. You need basalts, you need olivines, you need limestones, you need serpentinites, you need something that is good at absorbing CO2.

BW

Those rocks often bear copper. I think there's a lot of mining activity that produces a kind of waste product, a loess, like a fine mining tailing product that comes from some of these rocks.

ZH

Yeah. So some of the mining tailings are pretty alkaline, and there's companies working on figuring out ways to make those. 

BW

And isn't the country, isn't the whole of Oman, like basically one big serpentine rock, it's on the surface, right? 

ZH

Yeah. So there's efforts there to both store carbon in that and potentially figure out ways, like if you could grind it up. So actually, the third way, besides ocean alkaline enhancement, I was going to mention, that's potentially promising in the long run is this, is what we call surficial mineralisation.

BW

Say that again slowly.

ZH

Surficial mineralisation. So essentially, if you have really reactive...

BW

Enhanced weathering. Is that another word for it?

ZH

Enhanced weathering is a slightly different thing. I'll get to that in a second. But the terms are a little confusing. Surficial mineralisation is essentially taking a very reactive rock that either just exposed the air or in a low-tech reactor, exposed the air and moisture, is going to form carbonate. So the carbon will bind with the mineral and form effectively limestone, in a sense. That happens naturally, very slowly, but you can speed it up in various ways. And the nice thing about that is you can directly measure how much carbon is in that rock over time. And so you can track where it's going. Enhanced weathering is slightly different, where you're applying usually less reactive, because you don't want to damage systems, feedstocks to agricultural fields, where they weather. And in that case, they dissolve, they form bicarbonate, and that ends up in the ocean. Enhanced weathering is ocean alkaline enhancement in a muddy trench coat, as my friend David Ho is fond of saying. 

BW

Okay. So those are the things that you're currently exploring.

ZH

Well, those are the long-term ones. I'll go through the short-term ones very quickly. So the ones that are promising near-term but have more limits to how big they can grow: one is bioenergy with carbon capture and storage, or generally biomass-based carbon removal. Nature's really good at taking up carbon by itself through photosynthesis. There's a lot of waste and residues like sewage sludge or some agricultural residues or some forestry residues that are pile burned today or otherwise decay and release methane. And if you can instead take those and inject them underground or burn them for energy and capture the carbon coming out of them or turn them into biochar, there's a whole bunch of different things you can do with biomass. The challenge with biomass, though, is there's a limited amount that's really waste and residues. And once you start taking away farmland from food production to grow energy crops to do carbon removal, that becomes a whole nightmare from a life-cycle kind of approach. So we think there's a couple gigatons globally of real good waste and residue biomass. You run out after a point. The other is enhanced rock weathering in agricultural fields. This is taking mostly basalt, some cases olivine or elastinite, these reactive alkaline minerals, spreading them in farmland where they're actually good for the soil because they make it less acidic, can increase yields, things like that. When that rock dissolves, it eventually ends up in the ocean as effectively a form of ocean alkalinity enhancement. But the challenge there is there's only so much farmland that's near rock quarries. You max out, again, probably a billion tonnes a year globally. And a billion tonnes is a big number.

BW

It is a big number, but it's not when you consider how much we emit on a global basis, on a yearly basis.

ZH

Yeah. If you think that 10% of our emissions today, we're not going to be able to fully mitigate for one reason or another, that's on a greenhouse gas basis, like 6 billion tonnes of carbon a year you need to do CDR for, which is what most models have that the IPCC looks at.

BW

Just to round us off then, here we are in January 2026. The year ahead of us. We talked a little bit about it probably not being a record-breaking year. It's a bit unfair of me to ask you to forecast this, but best guess, what does this year look like?

ZH

I've already published my forecast. I actually find it fun to forecast a year and two years ahead because you don't have to wait long to know you're wrong. And sometimes it's more interesting when you're wrong because you learned something. But there's four different groups at this point that have produced forecasts for 2026, so Gavin Schmidt at NASA, myself, the UK Met Office, and my colleague Robert Rohde at Berkeley Earth have all published estimates for where 2026 is going to be. We'll use different approaches, but we end up agreeing actually pretty well this time around. We all expect it to be somewhere between the second and fourth warmest year with the best estimate of more or less on par with 2025.

And that's because we're starting the year with the La Nina event, which tends to be associated with cooler global temperatures. There does seem to be an increased likelihood of an El Nino developing in the latter part of 2026, but that's primarily going to affect 2027 temperatures. Historically, at least, there's been a bit of a lag between when El Nino develops and when the effect on surface temperatures is, of about three months. So something happening in later 2026 is mostly going to affect 2027. And for that reason, I do think 2027 has a pretty good chance of setting a new record. James Hansen also had a piece on this last month where he had a very high estimate for 2027. I think this is probably a little too high, but my best estimate for 2027 would probably be setting a new record as the warmest year. But we'll see. If the El Nino doesn't develop then — and it's notoriously hard to forecast that in advance — if we don't set a record in 2027, it's not going to be long before we do. Global warming is marching continuously upward.

BW

Yeah. And then my last question, I suppose, is knowing what you know about the models, knowing what you know about the observations, where are the areas where, if you had a magic wand, or if you had the budget of NVIDIA, where would you instrument to get some of these uncertainties reduced? Because it always struck me that we have all of this amazing technology now. A lot of it can be deployed to help us narrow down some of these error bars because the error bars are still pretty huge. Where are the bits that we just really don't know? We haven't got the data that these models are really lacking.

ZH

Narrowing down the error bars is important, but I don't think they'd fundamentally change what we know and what we need to do. The impetus to reduce emissions doesn't depend on climate sensitivity being 5°C for doubling CO2. It's bad enough where we're headed. That being said, I do think that this very large range of climate sensitivity is a challenge for the community and means the tail risks are unacceptably high today. And that comes down to a lot of different factors, but two of the biggest ones are aerosols and clouds, and the two are obviously related, as we talked about earlier. And so having better measurements of both and better models of both, I think, would be a really important way, particularly on the aerosol side, where we should be able to better measure aerosol cooling effects, better satellite instruments do that, more advanced things, I think, would be quite helpful on that front. But I'm not a cloud retrieval or aerosol person, so I can't get too far into the technical details.

BW

And that's more important than, say, perhaps knowing what's happening with carbon that's locked up in the permafrost of the Arctic Circle. Because I've always worried that there's some of these feedback, these nonlinear effects that could start to really change the way the sinks are behaving. Some of that is just this loss of the ice that we're locked into now, resulting in potentially rapid releases of CO2 or methane.

ZH

I mean, it's something we definitely need to keep our eyes on. And we have satellite instruments now that are pretty good at detecting regional methane emission patterns. We don't see a huge flux yet from permafrost, thankfully. 

BW

It's more coming from the tropics, right? Weirdly.

ZH

It's more coming from the wetlands in the tropics, in particular, and fossil fuel systems, of course. There's also an interesting dynamic where, as the Arctic warms and snow and ice retreats, you get more vegetation, 

BW

Which lowers reflectivity?

ZH

Which somewhat counterbalances…

BW

Oh.

ZH

Well, it's dark and absorbs more heat, but it's also storing more carbon. And so there's sort of this balance of effects of absorbing more sunlight, sequestering more carbon, methane and carbon coming out of thawing permafrost that's really hard to assess. Most of our models today suggest that permafrost loss is pretty nonlinear, that if we can limit warming to 2°C, it'll be pretty modest, at least for the next few centuries. If we end up in a world of 3°C or above, it starts getting much bigger. And so it's yet another reason to try to limit how much we warm.

BW

Yeah. Great. Well, thank you so much, Zeke, for spending the time with me today on what I know is a very busy day with all of 2025 data coming out. And I look forward to bumping into you in Berkeley in the future.

ZH

Yeah, definitely. Thank you so much.

BW

So that was Zeke Housefather, climate scientist and expert navigator of the complexities of our global climate. My thanks to Oscar Boyd, our producer, to Kendall Smith, our manager, to Jamie Oliver, our editor, and the Cleaning Up gang who make these podcasts possible, including our wonderful Leadership Circle members. Please join us at the same time next week for another episode of Cleaning Up.

Michael Liebreich

Cleaning Up is supported by its Leadership Circle. The members are Actis, Alcazar Energy, Arup, Cygnum Capital, Davidson Kempner, EcoPragma Capital, EDP, Eurelectric, the Gilardini Foundation, KKR, National Grid, Octopus Energy, Quadrature Climate Foundation, SDCL and Wärtsilä. For more information on the Leadership Circle, please visit cleaningup.live. If you're enjoying this episode, please hit like, leave a comment, and also recommend it to friends, family, colleagues and absolutely everyone. To browse the archive of over 200 past episodes, and also to subscribe to our free newsletter, visit cleaningup.live.