- [bright upbeat music] - Hi, I'm John Timmer. I'm the science editor at Ars Technica. And I'd like to thank Zeke Hausfather for joining us today for this live chat for Ars. So Zeke is the climate lead at Stripe, the financial services company. And he's also a scientist at Berkeley Earth, an organization that helps track the global temperatures. So, Zeke, thanks a lot for joining us today. - Thanks, John. It's a pleasure to be on. - So, a bit about yourself. How did you get into climate science initially? - So I had a bit of an unusual pathway to becoming a climate scientist. You know, I sort of got into climate issues as an undergraduate in college, you know, around 2001. During my freshman year, I read a book that came out 10 years earlier by Bill McKibben called "The End of Nature," which was one of the first big books about climate change. And it really got me interested in the issue. And back then, I got pretty into climate activism. You know, I was dropping banners off buildings, chaining myself to protestors and corporate offices. It was a fun time, and we won a lot of campaigns. But, you know, my focus back then was more on climate policy and economics. And I ultimately ended up going into the tech startup world and helping start a company in 2008 called Climate Culture that helped young people understand and reduce their carbon footprints. You know, this is right after "An Inconvenient Truth" came out. It was sort of the zeitgeist of the times. That then morphed, when the financial crisis hit, into a company called Efficiency 2.0 that was helping utility customers reduce their energy use. 'Cause a lot of the, you know, analysis I did around carbon footprint calculations also applies to home energy savings. That company got acquired in 2012. But during this period where I was working in the cleantech startup world, I was getting more and more interested in climate science myself. You know, I had started reading a blog called RealClimate that came out in 2006 by a bunch of early career climate scientists. Like, at the time, Gavin Schmidt, Michael Mann, Stefan Rahmstorf got involved in the climate blog wars in the early days, which was a very different era of the internet. Did a bunch of my own blogging around 2010 on surface temperature records and all the debates that were happening about urban heat islands and other issues at the time, and ultimately ended up teaming up with the group at Berkeley Earth, which was developing their own surface temperature record, in part 'cause I was doing a lot of work that paralleled theirs. We can talk about that work a bit later. But around 2014, 2015, I was spending, you know, a fair amount of my time moonlighting as a climate scientist, writing academic papers, going to conferences, despite not having a PhD, and working full-time in tech startups, and ultimately decided that my hobby was more fun than my day job. And so went back, I got a PhD, and, you know, did a fair bit of academic work around that. And now I am back in the private sector but in a role that's a bit more directly climate science focused. - What is it about the science that appeals to you? I mean, it's a lot of math, in the case of the sort of work you do. Was it the math that drew you in? Or was it really the questions that you were able to answer? - I mean, it's a combination of both, right? The climate is actually a really cool system. It's super complicated. There's so many moving parts. There's a lot of things that we can simulate at a high level but are really hard to understand at a granular level, or to model at a granular level. And so just trying to extract signals out of chaotic systems itself is really interesting for me. But I think part of it is also, it is such an important question for society, right? You know, there's a lot of science that is fascinating from a pure science perspective, but, like, you're the expert on one plant that grows on the top of this one region, in one part of the world, that probably isn't ever gonna change, you know, the course of human history. Whereas climate is an issue that, you know, because it poses such a big challenge if we keep putting more CO2 and greenhouse gases in the atmosphere, you know, better understanding what is going to happen in a warming world has a real impact on, you know, the choices we make as a society. And so it's really exciting to be working on a science that's so relevant for the big issues we face today. - All right, and I'd like to take a moment to remind the readers that if they have a... If anything Zeke says raises a question you'd like answered, please put it in the YouTube chat. And we have people paying attention to that who will try to get it answered before our time winds up. So you mentioned people like Mike Mann and Gavin Schmidt, who, you know, when they started in climate science, it was this quiet academic backwater in some ways. And you came in at a very different time, when it was already the subject of public controversy. And you mentioned blog wars and things like that. Has that sort of influenced how you view the science as you're doing it? Or do you try to keep the sort of public controversy and what you think about separate? - I mean, the science itself is obviously the science. You know, we follow what the data shows us. And if it shows, you know, results that are inconvenient for one particular side of a debate or another, you know, that is what it is. At the same time, I personally like exploring what questions are being debated. I think, in some ways, it makes an interesting paper to look out and say, hey, you know, everyone's arguing about urban heat islands, for example. Let's actually take all the temperature data and break it out by, you know, urban and rural, by a bunch of different metrics. Like, how bright is this spot at night based on satellite measurements? How much impermeable surface area is there in particular region? And let's analyze those separately and see if there's any difference in the rate of warming between urban and rural areas. That was actually my first climate science paper back in 2011, driven in part by that being the big argument at the time. But another paper I did, with Gavin Schmidt actually, later on was looking at the performance of old climate models. And it was inspired by a similar thing. You know, there's always arguments online about, you know, are climate models overestimating warming or underestimating warming? And so we decided to go back and actually collect every single climate model that had ever been run, you know, starting in 1970, when the first climate model came out that had some sort of future prediction, and going through the early '90s. 'Cause after that, you know, it really blew up. And looking at, you know, not how did those models perform in their sort of hind-cast test period, but rather, how did the model do after it was actually published? Because, obviously, the modelers themselves don't know the future. And so it's a really good independent test of the performance of a model to say, hey, this model in 1970 said this thing would happen over the next, you know, 50 years. Here's what actually happened. And it turned out, when we did that, we found that the vast majority of models, you know, got it pretty spot-on, that the warming that occurred after they were published was pretty much in line with what they expected to happen. Yeah, I think that's something that is clear if you go through the academic literature, but I don't think the public has really come to recognize that or what that portends to the future. - Yeah, I mean, it's pretty important because we don't have a perfect crystal ball. You know, we are reliant on simulations of what we think is gonna happen that are informed by the past. But, you know, they're certainly necessarily cloudy. - Yeah. So how did you end up at Berkeley Earth? You've mentioned that it sort of was part of your transition into serious climate science. - Yeah, so Berkeley Earth itself was a very interesting organization in the way it started up. So, again, in late 2000s, early 2010s, there's a huge amount of focus and debate, particularly online, about temperature records writ large. In some ways there's less of it now, to be honest, but back then a lot of people argued about things like, you know, the reliability of the temperature record, the amount of measurements we had over time changing, things like urban heat island biases. A lot of this sort of came out in the brief and somewhat overblown Climategate scandal in 2010 around, you know, proxy data and how that was treated. And so Berkeley Earth started around that period to provide a independent assessment of surface temperatures that was not produced by a government. 'Cause, at least at the time, all the other temperature records were produced by government agencies. And so it was a group of physicists out of Berkeley, including a Nobel Prize winner, Saul Perlmutter, Art Rosenfeld, who was sort of the godfather of energy efficiency, Richard Mueller, who led the group, who wrote the book "Physics for Future Presidents," Robert Rohde, who was Rich's doctoral student, who, you know, sort of did all the actual work. [chuckles] And they were initially funded primarily by the Gates Foundation and the Koch brothers, which is a very weird mix of philanthropic funding. And so part of the idea is it's this, you know, group of very austere, qualified physicists funded by both the left and the right-leaning philanthropies that could really settle these issues once and for all. And so, as I mentioned, around the time, I was doing some of my own coding and analysis on surface temperature questions. And so I got connected to the team at Berkeley Earth, started initially attending their meetings as sort of an interested volunteer, and eventually, you know, started working with them on developing the temperature product. And, you know, fast-forward, what has it been now, almost 14 years, 13 years, you know, a lot of those physicists have retired, and it's sort of Robert, myself, and a few other folks maintaining that temperature record and improving it in various ways. You know, we recently launched a high-resolution product that uses sort of a downscaling, machine learning-based approach to get 25-kilometer resolution temperature fields for the whole planet back to 1850. We've, you know, incorporated a huge amount of additional data. We have about 53,000 land-based weather stations now in the data product and, you know, millions upon millions of ocean-based observations from ships and buoys. So it's constantly improving those temperature records. I should say that, you know, when Berkeley Earth initially published, they, you know, perhaps unsurprisingly, found the same result that NASA, NOAA, the UK Met Office, everyone else had found that the Earth is warming, [laughs] and warming more or less as fast as we thought. In fact, we showed slightly more warming than the other groups since pre-industrial, in part because we had a bit more data in the early part of the record. - So what is the importance of the surface temperature record? What does that tell us? And how do we use that in terms of planning and just general understanding of what the climate is doing? - So the surface temperature record is both actually not a great measure of climate change, but the one that's most relevant for us, 'cause we live on the surface, and we experience the heat that's being measured in those surface temperature records. So when we talk about surface temperature records, essentially what we're doing is taking a huge amount of measurements from weather stations, from ships, from buoys, and using it to figure out how the temperature of different parts of the planet, and aggregated up to the planet as a whole, has changed over time. There's other approaches you can use called reanalysis that is more based on weather models, essentially feeding all these observations into a weather model every hour, including satellite and pressure and other data that's not directly temperature, and using the weather model to fill in the gaps. Groups like Copernicus in the UK do that. We at Berkeley just rely on direct temperature measurements. But it is, you know, it does produce this very iconic graph that shows the world has warmed by about 1.4 degrees since preindustrial, with most of that, about a degree or a little more than a degree of that warming happening since 1970. And, you know, it shows a pretty robust acceleration in the rate of warming since 1970 and also over the last decade, even compared to the rate since 1970. That said, you know, I mentioned initially that it wasn't actually that great a measure of global warming. And the reason for that is that, you know, what's actually happening with global warming, with climate change is that, you know, greenhouse gases concentrations are increasing in the atmosphere, that traps more heat in the climate system that would otherwise escape to space, and the vast majority of that heat, about 97%, does not actually go into the atmosphere. It goes into the ocean. It goes into melting ice. It goes into the land. And so if you want a really, really good measure of what's happening to the Earth's climate, you should actually look at the heat content of the ocean. 'Cause that's really where the energy is going. And unlike the surface, which varies a lot year to year due to things like El Nino and La Nina events and volcanic eruptions, you know, ocean heat content pretty much monotonically increases. Every year is setting a new record. Because that is really the measure of the thermostat of the planet. The challenge, of course, is that, while we have a reasonably good surface temperature record back to 1850, you know, ocean heat content, we really only have excellent data for the last 20 years or so and more sparse data back to 1940 or so. So it's not quite as complete in time, but it is a more reliable measure of what's actually happening to the climate system. - So one of the striking things about be Berkeley Earth to me is less that it wasn't a government organization and more that it came up with a completely different way of handling all the collection of complications that go in a temperature record, where people move the station that's taking the temperatures. They changed the time of day that the temperature's read, things like that. And you guys came up with a completely independent way of handling that and yet came out with the same answer more or less. So what do you feel is the importance of that sort of parallel replication of the other results? - Yeah, so one of the big challenges of constructing surface temperatures for the planet is that the way we've measured temperature has changed a lot over time, as you alluded to. You know, back in the day, they used to throw wooden buckets over the side of sailing ships, and pull them up and stick a thermometer in the bucket, and write down the measurement in the log. You know, from 1850 through 1930 or so, that was the primary way that ocean temperatures were taken. But it turns out that if you're pulling a bucket up the side of a ship, water in that bucket is evaporating, and that evaporation cools down the rest of the water in that bucket. And so you actually get a cooler temperature, the taller the deck of the ship. This is a pretty consistent relationship across these measurements. And so if you don't account for the deck height of the ship, you might even end up a biased temperature record, particularly if the deck height of ships has changed over time. Those bucket measurements then got replaced primarily with ship engine intake valves. The water that's used to cool the engine, you measure it as it's coming in. But it turns out engine rooms are a little warmer. And so that, in turn, produced a bias. And then we switched from ship engine room intakes to autonomous buoys that automatically communicate with satellites, which are much, much better instruments. But it turns out that those buoys measure temperatures consistently about a 10th of degree C cooler than ship engine room measurements. And so there's all these sort of things associated with changes in measurement over time that are gonna mess up your ability to construct a long-term temperature record unless you account for them. You know, that's the oceans. Land has had their own challenges around, you know, there's very, very few stations that have been in place and unchanged for 150 years or 170 years. You know, in the vast majority of stations, we've replaced old liquid and glass thermometers with electronic thermistors. The re-temperature's a little different. And so what we did for the Berkeley Earth approach was essentially say, okay, if there's a break in the temperature measurements compared to nearby stations, if something has changed at a station, at one station, that's not reflected in any of its neighbors, and that change persists over time, it's probably due to a non-climactic bias of some sort or another. You know, climate change happens over pretty broad spatial regions. It's not like, you know, one city is warming a degree, and the city, you know, 10 miles away is cooling a degree over 30 years. Like, that doesn't happen. The thermodynamics don't work. And so if you see something weird happening at one station that's not happening at any of its neighbors, you can pretty easily identify that as, you know, something like an instrument change or a time of observation change or a station move. And the approach we took with Berkeley Earth was to say, okay, when we detect those, we essentially cut the record there and treat everything after that as a brand new station with its own sort of unique starting point. That differs a bit from what previous groups had done, where they just tried to sort of collapse those differences and homogenize the data. The broader approach that Berkeley Earth uses, which is a statistical method called kriging, allows us to incorporate a bunch of more fragmentary records than other groups can do. So it really, you know, benefits from this approach. So, as you mentioned, it is a completely different approach to correction of changing measurement techniques over time. But it does give broadly the same results, particularly in places that are, you know, very messed up, like the US. Ironically, the US has more measurement stations than anywhere else in the world, but we have one of the more messed-up long-term records. And the reason for that is because, until the last couple decades, the US didn't actually really care about measuring climate. You know, these were primarily weather stations. And so if your weather station drifts a degree, you know, because something changes, that's not a big problem. You really just care about measuring day-to-day weather than longer-term changes. And so the US set up something called the Cooperative Observation Network. They had, essentially, volunteers set up weather stations in their backyard. And over time, the National Weather Service changed the directions for how those were operated. So starting in the 1950s, they said, oh, actually, instead of taking your temperature measurement in the evening, we want you to take it in the morning, because we want better precipitation gauge measurements before it has time to evaporate. And it turns out that shifting your measurements across the whole network from 5:00 PM to 9:00 AM or 7:00 AM has a pretty big impact on the results, even if we're talking about min/max thermometers. And, you know, there's other changes like the change from liquid and glass thermometers to electric thermistors that had big impacts on the network. Regardless, the US, for various reasons, has a pretty messed-up long-term temperature record if you just look at the raw data. But both Berkeley Earth and NOAA and NASA and other groups all get very similar results when accounting for these changes in measurement techniques. - All right, just wanna remind the people watching that if you, anything Zeke says raises questions for you, please drop them into the YouTube chat. We have people watching there for them, and we'll get to them before the hour is up. So the surface temperature record's been making headlines pretty consistently over the last couple of years. What's gone on that's been so striking about the temperatures we're seeing? - So the last two years in particular have been quite anonymously warm, you know, a lot warmer than many of us expected going into it. So a number of different groups around the world produce sort of year-ahead forecasts of what they think global temperatures are going to be. And this is based on, you know, depending on the group, either a dynamical model or statistical model. And going into 2023, you know, we all thought it would be a warm year but probably, you know, maybe the second warmest year on record. You know, it had a chance of setting a new record but only barely. What ultimately happened is 2023 was the warmest year by a huge margin. And, you know, it blew away anything that had happened previously. And it was really far outside of what we thought would happen. In fact, the only other year that these sort of models have missed predicting by such a big margin was 1992. And what happened in 1992 is, or right before 1992, is Mount Pinatubo erupted and put a huge amount of sulfur dioxide in the stratosphere, which cooled the planet a bunch. And, you know, the year-ahead projections didn't account for that, obviously, 'cause they didn't know the volcano would erupt. But the problem, of course, is there's not a Pinatubo, or a reverse Pinatubo, as the case may be, erupting in 2023. And so it led to a lot of debate and questions in the community of why temperature skyrocketed quite so much. And then, of course, 2024 came along, which was even warmer than 2023 and was the first year above 1.5 degrees globally. And so we as a community have spent a lot of time studying all the different potential causes. And so we've identified a couple potential culprits. You know, one is El Nino potentially behaving a little weirdly. So there's a seven-year cycle in the Earth's climate between El Nino, which is associated with warmer global temperatures and La Nina, which is associated with cooler global temperatures. These are driven by changes in the Tropical Pacific that are mediated by wind patterns. And, essentially, we've been in an unusually extended La Nina period since the end of 2020. It was a sort of rare triple-dip La Nina between 2020 and the start of 2023. And so because we had such a long La Nina, some people have argued that maybe that led to a more robust heat signature showing up from the El Nino event that developed in mid to late 2023, essentially the world getting warmer faster than it normally would, because normally there's a bit of a lag between when El Nino develops and when the peak surface temperature effect is felt. So that's one potential explanation. Another is there's been a pretty big change in the sulfur content of shipping fuels over the ocean. So in 2020, the International Maritime Organization required a 80% reduction in the sulfur content of shipping fuels. Sulfur dioxide itself is not a greenhouse gas, but it is highly climate-impacting. It both reflects sunlight in the form of sulfur dioxide, aerosols in the atmosphere, and that sunlight then bounces back to space. And it increases cloud formation, and those clouds themselves are reflective and bounce more sunlight back to space. And so by reducing the amount of sulfur and marine fuel, we reduced the reflectivity of clouds and the amount of sort of sunlight being reflected back to space and potentially caused some additional ocean heating. There's been seven or so studies on that published in the last year, year and a half. Most of them find a real but moderate effect, you know, maybe something like 4/100 to 8/100 of a degree C. So it doesn't explain the entire change we saw in 2023 but is part of the story, certainly. There is one study by James Hansen that finds a much, much bigger number of 0.2 C, which would explain pretty much all the anomaly we experienced in 2023. But that one is still a little bit of an outlier compared to all the other approaches in the literature. Another thing we looked at is there's an unusual volcanic eruption in 2022 of the Hunga Tonga volcano. It was unusual because most volcanic eruptions put, you know, big ones at least, you know, put some sulfur dioxide in the atmosphere and actually have a cooling effect on the climate. This one, because it occurred underwater, put a lot of water vapor in the upper atmosphere in addition to sulfur dioxide. Water vapor in the upper atmosphere warms the planet. Sulfur oxide cools it. And so the balance of those two has been a bit of a debate in the community. Now, I think there's been about six or seven papers simulating it, and they found that the effects are probably pretty small, you know, somewhere between a very slight warming of a couple hundredths of a degree C to a slight cooling of a couple hundredths of a degree C because of the balance between water vapor and sulfur dioxide in the stratosphere. So that one seems to probably not be the primary cause. But it's certainly been one that's had a lot of debate around it. Another change we've seen is that China has reduced its emissions of sulfur dioxide from burning coal, in its steel plants and things like that, about 70%, seven-zero percent, in the last 15 years. It's been a huge increase in air quality, or a huge recovery in air quality, I should say, over China, which has been a great story for public health but has contributed to some additional warming in the climate system, though that's happened more gradually. So it's hard to use that to specifically explain the jump in the last two years. And then the final cause that we've looked at is that there's been an uptick in the solar cycle. So there's about an 11-year cycle in the sun's output that affects the Earth's climate. The current solar cycle, number 25, is a bit stronger than anticipated. And so that might also add a couple hundredths of a degree C to the record over the past two years. So if you take all those different factors and add them together, essentially you can explain 2024 quite well. So the last year, you know, we're pretty much spot-on if you add the expected magnitude of all those things. 2023 is still a little mysterious. You know, we still can only maybe explain about half of the exceptional warmth. And, you know, 2023 also saw some truly crazy things happen in the summer. Like, September 2023 was half a degree C warmer than any previous September. You know, I think, at the time, I called it gobsmackingly bananas, which went a little viral. - That's the technical term for it. - [laughs] So yeah, there still is some mystery here, but I think the community has done a pretty good job of, you know, looking across all these potential causes and better isolating, you know, what the real drivers might be. - All right. So how important is that? Because, you know, if you go back into history along the climate record, you see lots of year-to-year variation, and you see some big jumps in the past. How important is it for us to understand each... You know, despite all that, the long-term trend is pretty clear. So how important is it for us to understand when these, you know, what's causing these occasional shifts beyond the trend? - I mean, I think it's very useful from a broader understanding of the climate system perspective to see what's driving these changes. I think whether or not it's important for society depends on if what's driving these changes are persistent or not, right? You know, if it's a weird volcanic eruption, or if it's a solar cycle, it's probably not much of a concern, 'cause that will go away in a couple years, right? It's cyclical, or it's, you know, short-lived. If it's a, you know, cut in sulfur emissions that's going to persist, that's gonna drive continued higher warming temperatures than we'd otherwise expect. There is also this sort of broader question around how much of the changes we're seeing are driven by what we call forcings, so human emissions, versus feedbacks, you know, responses to those emissions, of the climate system. And so there's been a lot of debate in particular about cloud changes that we've seen. You know, we have satellite measurements of clouds. And we've seen a pretty significant decline in the reflectivity of clouds over the last 10 years or so. There's a very high-profile paper in "Science" about this last year. And that paper essentially said, you know, we see this in the satellite data. We don't know exactly what's causing it. But there's sort of three broad possibilities. You know, one is that maybe this is just natural variability, right? We only have a satellite cloud record going back to 1970 or so. Before that, we didn't have satellites. You know, we haven't seen anything like this since 1970, but maybe something like this happened in 1930 or 1900. You know, it's unusually persistent. We don't have a physical mechanism, or a mechanism to explain why it would be so persistent due to internal variability, but we can't completely rule that out. Another option is that this is being driven by the aerosol changes I mentioned. And certainly that's part of the story, right? You know, we know that aerosols, sulfur dioxide emissions in particular, increase cloud formation and reflectivity. And so by cutting those emissions, we're reducing cloud reflectivity. But, at the same time, it's hard to explain the total magnitude of cloud changes we've seen based on our best understanding of aerosol forcings. And so the third option that they brought up in that "Science" paper is, well, maybe this production of cloud reflectivity might actually be a response to human-induced warming. You know, that's something that we see in a lot of climate models, that as the world warms, we see less low lying cloud formation and less reflective low-lying clouds, which in turn leads to more sunlight being absorbed by the surface and more warming. At the same time, you know, there's a huge range of that response across climate models. Some show it very strong. Some show it very weak. You know, maybe even like 5% of models show a slight negative cloud feedback. And so that's one of the biggest areas of uncertainty when we try to project future warming is the magnitude of these cloud feedbacks. And so maybe if what we're seeing in the real world is a stronger cloud feedback, that would then indicate that the climate is more sensitive to our emissions, and that we might expect more warming in the future. And so looking at periods like this, and longer periods like the past decade, can help disentangle some of these problems. I should also mention that, you know, the last two years have been anomalous, but we've also seen a pretty robust, at this point, increase in the rate of warming over the last two decades or so compared to what we've seen, you know, broadly since 1970. You know, since 1970, the world has been warming at about a little below 0.2 centigrade per decade. Now, in the last decade, the human contribution to warming we estimate as 0.27 C per decade. So it's about a 40% increase in the rate of warming, which obviously is not what we wanna see when we want to reduce the impacts and the rate of warming. - Yeah, so one of the thing that this sort of gets into is the worry that we're approaching some tipping points, where we'll get the climate into a new state that it isn't very easy to get it back out of. Does the temperature record show any indications of that? Or are there some that you've seen in climate models that you expect we might be getting close to? - So tipping points is a complicated subject. And it's important to draw a distinction between feedbacks, which are more sort of linear and predictable responses to warming. You know, the warmer the world gets, the more snow and ice melts. The more snow and ice melts, the more dark surfaces are exposed. The more dark surfaces are exposed, the more sunlight they absorb, cause more heating. Some things that are occasionally considered tipping points like, you know, methane and carbon loss from arctic permafrost are probably better thought of as feedbacks, 'cause they respond, you know, proportionate to temperature changes. But other areas, you know, there might be parts of the climate system that, you know, are fine, until suddenly they get pushed past a critical threshold, and they're not. Examples of that would include coral reefs, which are very, very sensitive to certain temperature thresholds, where the coral expels its symbiotic algae, zooxanthellae, and dies. You know, we've seen that happening in a lot of reefs in recent decades, associated with marine heat waves. But even something like that, you know, it's a tipping point for a particular reef. But, you know, Hawaii is a lot more resilient than the Great Barrier Reef, because it has bigger variations in annual temperatures. And so, you know, coral reefs as a global system, [object scraping] sorry, [laughs] are less of a single tipping point than a gradual tipping point across different regions. You know, another that gets discussed a lot is the Amazon rainforest, where the combination of deforestation and climactic changes could favor savanna-type ecosystems over tropical forests. And so as you lose forest, it could be replaced by grasslands, which in turn would, you know, have big effects on the broader climate, both through the carbon stocks lost in the forest but also effects on precipitation patterns and other things. You know, another climate tipping point that often gets brought up is the AMOC, or the Atlantic Meridional Overturning Circulation. Essentially, there's a ocean current that's driven by evaporation and saltier water in the North Atlantic sinking down into the depths. And that can be disrupted if there's a lot of freshwater injection from, say, Greenland melting into the North Atlantic, which makes the water less salty, which makes it not sink, which potentially shuts down that ocean conveyor belt, and then that ocean- - That's the one that feeds warmer water towards Europe, which moderates temperature. - Exactly. So, in that case, ironically, you'd probably end up with a little cooler global temperatures. I mean, you'd still have warmer global temperatures 'cause the world is warming, but compared to a world where the AMOC didn't shut down, you'd actually have slightly cooler global temperatures. But you'd have really big effects in Europe, you know, extremely cold temperatures in the Nordics, for example, or the UK, and big effects on rainfall patterns and things like that. So those are some examples of the tipping points that are brought up. Those are mostly what I consider regional tipping points. They have a big impact on a particular region of the world or on a particular ecological system, but they're not necessarily gonna, like, send the Earth into runaway warming or something like that. There are some suggestions that, you know, in the Earth's more distant past, there are periods where a relatively modest change in greenhouse gases or incoming sunlight had a much bigger climate effect than we would expect based on, you know, what we know about the climate today. Those often happen in climates that are a bit different than the one we're in right now. And so that should give us a little bit of pause in terms of what happens. You know, there's a great quote by the climatologist Wally Broecker, who says, "The climate is an angry beast, and we're poking up with sticks." And so the fact that there are some of these pretty big climactic changes in the Earth's more distant past that we don't have a perfect explanation for should be a reason for caution today, as we're quickly moving the planet outside the range where it's been for the last few million years. And, certainly, there's some mechanisms proposed for that. You know, there's some modeling that shows that once we get to really high CO2 concentrations, you know, 1,200 parts per million or so, which hopefully we're never gonna get to, but fingers crossed, you could end up with a significant loss of stratocumulus clouds over the oceans. And that would, you know, lead to maybe four or five degrees additional warming. And that's more of what we'd call a Hothouse Earth scenario, where you cross some broader climactic threshold that leads to a very quick and rapid, you know, increase in temperatures. Those seem to be not imminent, thankfully. But it's an area that we as a scientific community, you know, are continuing to get a better understanding of and better calibrate the risks. I think, broadly speaking, you know, right now, we're around the temperatures the world experienced in the last interglacial 120,000 years ago. So it's sort of the warmest now that it's been in the last 120,000 years, most likely. If we keep pushing temperatures up, that quickly becomes the warmest that's been in the last few million years. And I think the more we go into unprecedented territory for our modern climate, the bigger those risks become. - All right, so those were the things that interested me about having the chance to talk with you, but I think we're ready to turn it over to reader questions. So I've got a collection of them here. So we've got some, one from a guy named Eric Johnston, who says, "Considering we may be past the point of return for a one- to two-meter increase in sea level by the end of the century, what do you know about, you know, what low-lying countries are gonna be facing and what they could potentially plan to deal with that?" - Yeah, so that's probably a little on the high end of what we'd expect by the end of the century. Though, again, sea level rise simulations are inherently uncertain, in part because we don't have great models of ice sheet dynamics. You know, if you just assume the ice melts from the top down, that's easy, and then you don't get that much sea level rise. But it's more like how the ice sheets or glaciers speed up as melt happens that affects things like that. I'd probably say a meter is more reasonable upper-end projection for the end of the century. But at the same time, it's true that we've already locked in a huge amount of future sea level rise, you know, even if we stop warming today. Because, you know, at 1.5 degrees of preindustrial level, we're gonna continue to have lots and lots more ice melt until the Arctic and the Antarctic reach sort of a new equilibrium. So that does mean that we're gonna have to plan for a lot of future sea level rise. Part of that might be, you know, managed retreat in some areas where it is too expensive to build coastal protections. Part of that might be expensive sea walls or raising houses, things like that. And it's really, when you think about sea level rise, the impacts are less of a gradual creeping up of the tide and more what happens when you have a large storm surge, you know, a Hurricane Sandy or, you know, that sort of thing on top of higher seas. And that's when you can really get catastrophic levels of flooding and damage for homes. - All right, a reader named Troft, and I'm paraphrasing his question here, but if you could improve any one measurement we make of the climate right now to have a clearer picture of what's going on, what would it be? What do you think our big open questions are that could be tidied up? - That's a good question. I'd say better measuring aerosol effects on the climate would be a really big one. You know, we know that aerosols have a strong cooling effect. - So why don't you define aerosols, because I think the common use is probably a little different from the scientific. - Yeah, sorry. When scientists say aerosols, we don't mean spray cans, [laughs] which causes no end of confusion. We mean small particles that are suspended in the atmosphere and can stay aloft for, you know, days to weeks. So think of, like, air pollution. And so today, the primary air pollutant that's climatically significant is sulfur, which forms sulfur dioxide when it's burned. It occurs naturally in fossil fuels primarily. And so that sulfur, as I mentioned earlier, you know, both directly reflects light back to space and increases cloud formation. But we don't really know exactly how big a cooling effect it has. So our best estimate is somewhere around half a degree C cooling from human emissions of sulfur today, which is big. That's like a third of all the warming the world has experienced has been masked by these sulfur emissions. But it could be as little as, you know, 0.2 C or as big as 1.2 C. And if it's on the high end, that has huge implications for society. Because as we clean up fossil fuels, all this sulfur dioxide in the atmosphere goes away. And if we have, you know, a degree of warming that is sitting waiting for us, that's being suppressed, that's a very unpleasant surprise. So narrowing down on that, I think, would be really important. Another is getting better data on Earth's energy imbalance. So in the last decade or so, we've been able to develop satellite instruments that can more directly measure how much heat is being trapped in the Earth's system. The current one we use is something called CERES that NASA put up there, but it is an imperfect instrument. It's a single satellite, you know, getting some redundancy there. And better narrowing down that value, I think, is gonna be important. 'Cause there've been some very intriguing data coming out recently that numbers we're getting from that satellite are a bit higher in terms of how much heat's being trapped in the Earth's system, and particularly how much it's increased in recent years than we expect to see in most climate models. And if that's actually the case, and it's not, you know, a calibration issue with the satellite, for example, you know, that would have important implications on what we'd expect to happen in the future. - That sort of raises a question by, I don't even know how to pronounce this, I think it's Bearable W., who mentions that, you know, the current government is very hostile to climate science. They're getting rid of some instruments we have monitoring it. And he's wondering if the community is making backup plans to preserve data and things like that, that's currently sitting on government servers, and making plans for just adjusting to a different information availability. - Yeah, so there's a lot of work going on behind the scenes to try to make sure that there's redundancy. You know, a lot of the global temperature work has gone through NOAA in the past. And they've sort of been the clearinghouse for all the global temperature records. And sort of, that's now being replicated by Copernicus, which is the European Union sort of equivalent of NOAA in terms of climate data and monitoring. And so there'll be a completely sort of independent system operating there shortly, which will provide important redundancy if, for some reason, NOAA's operations, you know, get disrupted there. I don't think there's a huge risk of, like, existing satellites that are up there just, you know, being turned off. 'Cause that would be silly even for the current government. But there is a big risk of them not being replaced at the end of their life. You know, satellites only last, depending on their orbit, you know, for a matter of years or a couple decades. And so they're constantly having to be replaced by new instruments. And if there's significant cuts for NASA's budget or NOAA's budget, as have been proposed by the current administration, we may well end up with very unfortunate gaps in the monitoring of, you know, critical climate data, things like aerosols or Earth energy imbalance or the gravimetric data we're getting from GRACE that gives us estimates of sea level rise and ice sheet loss. So making sure those sort of things are continued is important. I'm sure there's discussions among Europeans and others about potentially developing their own instruments there. I'm not necessarily privy to those. But I think, if I were to worry about one thing, like, really biting us long term would be if one of those gaps arises, 'cause then we can never go back in time and fill it. - So this may be, take you back to your days in the trenches of the blog wars, but a user called Nobody at All is asking, you know, how do I even start talking about, you know, the science here with his family members, who are addicted to Newsmax and just don't believe any of it? - I mean, I think I'd emphasize that the basics of climate science are, you know, they've been around for a very long time and are pretty robust. Like, we can debate about policies, if we need to build nuclear or build renewables or do X, Y, and Z, or have a price on carbon or a border adjustment tax. Like, that's all politics, and people can have different opinions there. But I think I'd emphasized to them that, like, we discovered CO2 as a greenhouse gas in, you know, the 1870s, you know, work by Eunice Foote and John Tyndall and Svante Arrhenius. Like, this is not new science, right? And, in fact, a lot of the critical work on atmospheric radiative transfer physics, the sort of how the greenhouse effect works, was funded by the US military. In fact, if CO2 weren't a greenhouse gas, heat-seeking missiles wouldn't work. And so the fact that, you know, we still use, like, MODTRAN and HITRAN, these radiative transfer data sets, to drive a lot of climate models that were originally developed by the military, you know, should give them, you know, some reassurance that this isn't just kooky left-wing academics, you know, building computer models in their spare time. This is based on a huge amount of measurement and a huge amount of basic physics. There's also a great paper that some colleagues of mine at Lawrence Berkeley National Labs published back in 2015, about a decade ago, where they actually set up sensors pointing up at the sky in two sites, one in the Great Plains and one in- - I remember that work. - Alaska. Yeah, and they actually measured how much heat was being trapped and reflected back down by greenhouse gases. And then they compare that to what the models expected would happen. And it turned out the two matched almost perfectly. And so it's not just a simulation. We can actually directly measure with instruments what is happening to the climate system. - Possibly related, by somebody named Tune, and he said, you know, he talks to somebody who's less than 20 years old, who feels they have already experienced an altered climate in the limited number of years they've been paying attention to things. And if we're seeing palpable changes on that sort of short timescale, how is it that people can still... He wants to know how people can still be questioning the reality of this, if it's just a matter of perception, at this point? - Yeah, I mean, never doubt people's ability to rationalize away [chuckles] things that are right before their eyes, right? You know, for better or worse, we can normalize changes that are pretty terrible pretty quickly. You know, I feel like that was, certainly for me, one of the abiding lessons from the COVID period we all experienced is that we can normalize bad things pretty quickly. So I do think it's tough. I do think that the commenter has some truth that changes are happening faster now in a way that is harder to ignore. You know, for example, I lived in California, and I've lived here for, you know, 15 years or so now. And when I first moved here, you know, we didn't really have catastrophic wildfires. Like, there were wildfires, they burned some areas, but it would be like a little thing in the news. You know, over the last five, six years, we now have like a smoky season in the summer. We have to keep our doors and windows closed because the air quality is so bad outside. You know, the sky turns yellow, that sort of thing. And so during that period, or actually in the broader period since 1980, you know, the number of fires happening in California have actually declined. We have less fires today than we had 30 years ago, but the area burned for the average fire has increased by a factor of three. And so it's really not a problem of ignitions. It's a problem of conditions. It's the fact that the vegetation is so dry. And so, like, things like that, there's, you know, count it as an equivalent thing, you know, there's increased rainfall associated with tropical cyclones that contributed to damages from things like Hurricane Helene. You know, we're just seeing so many new records broken every year in different parts of the country by climate extremes that it is becoming a lot harder for people to ignore them. - Yeah, I can say that the area I grew up, where some of my family still lives, when I was young, it was a question of how many snow days you'd get a year, where you got off from school unexpectedly. And now the question is more like, do we need to put snow days into the budget for the school as something they might need to worry about? It's just shifted entirely. - I'm sure your audience will appreciate this, but there's a great "xkcd" about this, about snow days in St. Louis and how, like, back in the day, it was really common. Now, you know, snow is actually a little unusual in the winter. And then it's like, fast forward to 2030, "Look, snow! Global warming isn't real." - Yes. - It's like, no. [laughs] - So a couple of people have sort of, both in our comments when we announced this and in the Q&A on YouTube, have asked about the role of aerosols as a potential moderator of the climate. So, you know, we're seeing a potential effect of getting them out. And you said there may be some nasty surprises lurking there. At what point does, you know, using aerosols as a tool to manage the climate become something that we really need to start considering? - So it's a tricky question. And there's a lot of different views on this in the community, I mean- - Yeah, and in the- - or actually reminds me- - And in the policy world, it's hard, very hard. - And in the policy world. [chuckles] It reminds me, there's an old "Futurama" skit, I think circa 2006 or so, where it was an episode about global warming, and they solve it by dropping a larger and larger ice cube in the ocean every year, thus solving the problem once and for all. And, you know, putting sulfur dioxide purposefully in the atmosphere, as we're doing... Well, we're not putting it purposefully in the atmosphere today, but we're putting much in the atmosphere today, you know, would cause cooling. There's no question about that. The challenge is that the warming effect of CO2 is very, very persistent. So if we got all of our emissions to zero today, all of our CO2 emissions to zero today, and held everything else in the climate system more or less constant, you know, it would take at least a thousand years for the Earth to meaningfully cool back down. So we're sort of stuck with the warming we have for an extremely long period of time. And so if you try to counterbalance that by putting sulfur in the atmosphere, you know, you have to do it every single year. If you're still emitting carbon dioxide, if you don't get that down to zero, you have to put more and more and more in every year. And if you ever stop doing it, you know, in the next millennia, you're gonna immediately, you know, over the course of a couple years, shoot back up to the temperatures you'd have having never, you know, done it in the first place. And so it doesn't actually solve the problem, the underlying problem, which is the greenhouse gases in the atmosphere. It just, you know, puts a bandaid on it. So in the best case, it could buy us time to figure out better solutions to actually solve the underlying problem. In the worst case, you know, people could say, oh, wait, it only costs a couple billion dollars a year to mask this problem, so we'll kick the can down the road. But the problem, of course, then is that, as long as CO2 is increasing, you're putting ever more sulfur dioxide in the atmosphere, which then starts having a bunch of knock-on effects, right? You know, at some point, you start turning the sky white, which, you know, might not be a great thing for many people. You have lower agricultural yields, 'cause you're blocking more sunlight. The crops can't get any more. You have impacts on precipitation patterns. With higher amounts added, you could potentially degrade the ozone layer. So, you know, as an option that's built into a strong emissions reduction pathway with a very limited deployment, I think there is a case there, you know, maybe if we just replace what we're removing by cutting fossil fuel emissions and nothing else, like, that might be manageable. But I do think the sort of open-ended, using it as a solution to the problem is not very plausible and potentially sets us up to just, you know, screw over future generations by, you know, buying time for ourselves now. - Right, somewhat related to this is the big issue that it doesn't affect is ocean acidification. And one of our readers named, goes by Rain Shadow, asked in the discussion, you know, is there an equivalent, are we seeing the equivalent of a hockey stick curve when it comes to ocean acidification? I know that's a little outside your area of expertise, so. - Yeah, so we've certainly seen a big drop in ocean pH, so an increase in acidity in recent decades compared to, you know, the 20th century. I'm actually less familiar with the literature on paleo-ocean acidity data. I'm sure it exists out there. I'm sure someone's done studies on that and found some proxy that correlates with ocean pH. I wouldn't be surprised if the current ocean pH changes are, you know, very much outside what we've seen in millennia, if not hundreds of thousands of years, in part because we're already at CO2 levels that are at the highest in 3 million years or so in the atmosphere. But I'm less familiar with, like, really long-term ocean pH datasets. But the broader mechanism here is that the ocean takes up about 30% of the CO2 that we emit today, which is great. If it didn't, we'd have a lot more global warming. But that taking up of CO2 then forms carbon acid among other things that, you know, changes to pH of the ocean. And it turns out, when the ocean gets more acidic, it's actually less good at taking up CO2 from the atmosphere. And so it sort of weakens that carbon sink over time as it becomes more acidic. - Yeah, I'd say the one thing I've seen is, in past events where there's a lot of CO2-driven climate change, they actually see a decrease in calcified shells being deposited in sediments. So you go from sediments being mostly white to all of a sudden turning very brown as a lot of the shells just get dissolved by the high levels of carbonic acid. - Yeah, though it is worth pointing out today that the, at least near-term, the impacts of temperature on ocean processes seem to be larger than impacts of pH. But that's obviously gonna change, you know, later in the century as pH changes become more dominant. Like coral bleaching, for example, is being primarily driven by temperature today. - All right, so we're pretty much near the end. And I'm sorry we didn't get to all the questions our readers had. But I wanna throw one more out for you from my own mind, which is, you know, what scientific questions are you most interested in answering right now? So, you know, what does your work allow you to look at that's really fascinating to you? - Hmm, that's a good question. I feel like there's- - [laughs] Sorry to surprise you with it. - No worries, there's a bunch of different projects that I'm noodling on right now. The baby's up, so sorry for the noise in the background. I think one that I've worked on off and on for a long time and currently have some things related to is this broader question of how sensitive the climate is to our emissions, which is something that climate science has not been as good as we'd hope at resolving. So back in the 1970s, there was something called the Charney Report that was put up by the US government. It was one of the first big reports on climate change. And they estimated back then that if we doubled the amount of CO2 in the atmosphere, the world would warm somewhere between 1.5 degrees and 4.5 degrees at equilibrium. Fast-forward to 2015, the IPCC Fifth Assessment Report said, if we double the amount of CO2 in the atmosphere, the world will likely warm somewhere between 1.5 degrees and 4.5 degrees. And so we had this range of, hugely important value, right? You know, 1.5 degrees warming is very different in terms of impacts than 4.5 degrees centigrade warming. But that uncertainty was largely unchanged for almost 50 years. We've made some progress recently. So the most recent IPCC report, we narrowed that range down to 2.5 to four as the likely range, and then a very likely range, meaning the 90th percent chance of two to five. And that was by combining a bunch of different lines of evidence from the paleoclimate proxy records, from physical process models, from observational records, using essentially the fact that if you have multiple independent estimates that are not related to each other, the joint combination of those using some Bayesian magic ends up being significantly narrower than any individual one would be by itself. But, you know, that still is a pretty big range and so, you know, working on various efforts to try to synthesize all these different estimates and, you know, better resolve that. 'Cause if we knew where we were actually headed by the end of the century, it would be pretty helpful, you know. 'Cause right now, even if we end up in a world where we know our future emissions, so if we end up in a current policy world where we end up at, you know, a best estimate of 2.7 degrees C by the end of the century, you know, that could still be as high as four degrees C if we roll sixes on the climate dice, so to speak, if we have high climate sensitivity and high carbon cycle feedbacks. And so if we can narrow that range, if we can say, actually, no, the highest we can expect is this instead of that, you know, or if that range expands, and it's even higher, that's more worrying. But that's hugely relevant for the decisions we make. - All right, well, I'd really like to thank you for spending all your time talking with us about your work and the general understanding of the climate. So that was Zeke Hausfather of Berkeley Earth and Stripe. And I'm John Timmer. And I'd like to thank everybody who watched for tuning in and giving us some good questions. - Thanks, John. [bright upbeat music]