TRANSCRIPT OF INTERVIEW WITH DR. TIM PALMER, OXFORD – June 24, 2020
This is Radio Ecoshock with your host Alex Smith.
Alex Smith: Tim Palmer is Royal Society Research Professor in Climate Physics at Oxford University, formerly at the U.K. Government’s Met Office and the European Centre for Medium-Range Weather Forecasts. Tim’s work is now used by climate modelers all over the world. Palmer has won top scientific awards in both climate and weather forecasting. He is an internationally-known expert in his field. When Tim Palmer warns extreme climate change is advancing faster and farther than predicted, we need to listen very closely. From the U.K., Dr. Tim Palmer, welcome to Radio Ecoshock.
Tim Palmer: Thank you very much.
Alex Smith: So it’s been a maxim, if humans double the amount of carbon dioxide in the atmosphere compared to 1750, before we started all of this, scientists expect the mean temperature here on earth would be three degrees centigrade hotter, and that’s been a bedrock for decades. Why are multiple scientists and institutions having to rethink that?
Tim Palmer: Well, the first thing to say is that we don’t actually know, even before the latest research, we didn’t know precisely, and we didn’t have precise figures on how much the Earth would warm as a result of a doubling of carbon dioxide concentration. So the number three degrees that you just mentioned, you should really think of that as an average of a range of estimates that went from somewhere between one and a half degrees at the very lowest end, up to about four and a half degrees at the upper end. And we always thought that those two numbers, one and a half and four and a half with brackets, the range of possible warmings that we would see as a result of carbon dioxide.
So what’s happened is that quite unexpectedly, a number of models from the very latest round of climate models that feed into these IPCC Intergovernmental Panel on Climate Change assessments are now producing warmings that go outside that range and they go at the warmer end, they’re going up to five or even five and a half degrees warming for a doubling of carbon dioxide. And that is certainly unprecedented amongst the models to date. And we’ve been running models for, I don’t know, best part of 40 years probably. So it’s both a concern and a surprise that now these most sophisticated, latest generation models, a number of them, at least, are producing these much larger warmings.
AS: Who are some of the official research groups who are encountering this higher possibility?
TP: Well, it’s the U.K. Met Office, the Hadley Center from the U.K. Met office, which is certainly one of the top models, the National Center for Atmospheric Research in the U.S., the U.S. Department of Energy’s climate model. And I think there are one or two others. So obviously, the question is what’s changed about these models?
One of the things that’s changed is the way in which the models represent clouds. Now, clouds are a critical part of the whole climate change question because how clouds adapt or adjust, rather I should say, to increasing levels of carbon dioxide will determine whether clouds act as a kind of damper on climate change, whether they will make it less warm than it would otherwise be, or whether they will amplify climate change by making it warmer than would otherwise be. And this has always been the single biggest uncertainty in projections of climate change and the range of values that I talked about before, one and a half to four and a half, the previous range of values was mostly predicated on the fact that we did not, and we still don’t, to this day have a complete understanding of how clouds will affect the global warming that we see from our emissions of carbon dioxide.
AS: I’d like to come back to the cloud problem in more detail, but first let’s talk about your article published in that top journal Nature. The title is “Short Term Tests Validate Long-term Estimates of Climate Change”. It seems incredible to me, that weather forecasts for a mere six hours could tell us anything about climate change hundreds of years from now. Please take as long as you need to explain how is that possible?
TP: Well, it is indeed peculiar, but nevertheless, this is the case. And the story behind this Nature article, actually, goes back several years. When some colleagues of mine published a paper claiming that it was possible that this climate sensitivity, so how much warming we got from a doubling of carbon dioxide, could be as large as 10 or 11 degrees. And this was a very interesting project. This is from a few years ago now, but they took a Met Office, U.K. Met Office, climate model and configured it so that it could run on people’s PCs. So it was a relatively low resolution version of the climate model, but nevertheless, a credible one. And what they did was they, it was kind of a citizen science project, where people could download a version of the Met Office climate model and do their own climate change experiments on their own personal computer.
Now, the thing was that each model that was downloaded by the public had very, very slightly different parameter settings. Now, I need to explain what parameters are now. So when you run a climate model, you’re trying to represent the atmosphere and the oceans and deviations in atmosphere as a fluid. And we have well-defined equations which govern how fluids behave, Newton’s laws of motion and all that sort of thing. But there are always processes and phenomenon like clouds. A cloud’s a good example of a process which you can’t resolve by the grid in your model. So you can’t represent a cloud as a sort of fluid dynamical object, and clouds have to be represented in approximate ways in climate models. And that was true of this particular study. And it’s still true today even of the most advanced models in the world that clouds are represented by what I would call approximate formulas rather than strict fluid dynamical equations.
And these formally have free parameters that you just have to set, you have to appeal to observations and maybe appeal to theory and appeal in a way to your gut feelings about what the values for these parameters are. And an example of a parameter could be how much air is mixed inside the cloud from the outside environment? The models can’t represent cloud processes. So the mixing between what’s called cloudy air and environmental air. The dry environmental air and the moist cloudy air is something that’s just set by a parameter in the model.
So when, in the citizen science project, people were downloading different versions of the models and some of them this particular parameter actually, how air is mixed between the environmental air and the cloudy air, was changed slightly. And the ones where this parameter was smaller than it typically is tended to produce very large warmings due to a doubling of carbon dioxide. And the paper that my colleagues published said that we can’t rule out these large amounts of warming because the basic climate that the models produced with these lowered parameter values is really not much different to the climate of the model where the parameters have their standard settings.
Now, I got kind of curious about this result and one good… Well, if you’re forecasting the weather tomorrow, let’s say, a key thing you’ll want to know about the weather is what is likely to be the daily maximum temperature, how hot will it get at the hottest part of the day tomorrow. And if you get your clouds, the representation of your clouds, completely wrong in your model, so you get it much too cloudy on a day when in reality there’s very little cloud, then that will affect your prediction of tomorrow’s daily, maximum temperature. Or conversely, if you have not enough clouds and on a hot summer’s day, you might find your daily maximum temperature as much to warm.
So I asked the question with these reduced values of the cloud parameters, what would actually happen to a weather forecast? So if you just ran a weather forecast for six hours would it get better or would it get worse? And what we found, I did this with a colleague, I should say, Mark Rodwell from the European Center for Medium-Range Weather Forecasts. We took that model, which incidentally, is, literally, the world’s premier weather forecast model and changed the parameters that led to this large global warming that my colleagues were seeing with the citizen science project. And we showed that, in fact, the kind of cloudiness was so degraded even after six hours, that the weather forecasts were pretty much lousy, I would say. They were really poor.
And you would never, if you were in the weather forecast center, you would never in a million years adopt the parameter settings, the reduced values of this mixing of cloudy air parameter. You’d never introduce that into your weather forecast model because you knew that it would make the forecast worse. So in this way, we could discount the models that were producing this very large global warming. Anyway, so that was a study that was done a few years ago. And I kind of made the case in the paper that this should be a pretty standard test to do on climate models. But the problem is that most climate centers don’t have that capability to run weather forecasts. It requires some additional software that most climate centers don’t have, but the one exception to that was the Met Office. The Met Office in the U.K. have, in fact, they call it a unified model because it does those whether… The same model system is used for both weather and climate predictions.
So when these new results came out with this larger global warming, the very latest results with five and a half degrees. I asked them, can you do this weather forecast test? The changes you made for the model, that made it produce more global warming, would they actually produce a better weather forecast or a worse weather forecast? And if they produce a worse weather forecast, you can discount these values. You can say, well, okay, this is probably not the way the climate is going to go. The result, however, turned out to be the opposite. In fact, in this particular case, the changes that were made to the model made the weather forecast better, not worse. So the conclusion is, therefore, that one actually should be genuinely concerned about these higher values. They probably do represent improved representations of clouds and improved physics.
AS: Your paper talks about a catastrophic warming of perhaps five degrees C when atmospheric carbon dioxide reaches 560 parts per million, that’s the doubling. On our current emissions path, do we know what year that doubling is expected to occur?
TP: Well, the term climate sensitivity is a fairly abstract idea because, as you say, what it means is it’s the amount of warming that will happen in the atmosphere once the atmosphere equilibrates to this new level of double carbon dioxide. So if we doubled it overnight, we’re not going to double it overnight, but if we were to double it overnight, we wouldn’t see the atmosphere warm by five degrees overnight. And the equilibration, actually, takes hundreds of years. I mean, by the time… It really depends on which parts of the climate system you want to equilibrate. The deep oceans would maybe take even thousands of years to equilibrate to that. But I would say as far as the atmosphere is concerned, most of the equilibration probably occurs within once 200 years, that sort of thing.
So it’s not, when we talk about five degrees warming, it’s not something that will happen probably in this century, but if you’re equilibrating to a five degree warming, then it means for example, by the time we reach 2100, we may, well have got three or four degrees of warming out of that. So knowing that the climate sensitivity is high is a matter of concern, even if we’re only bothered about what the climate is going to be like in the coming 80 years or the last of this century.
AS: But we’re committing the rest of humanity in the future to living in a much hotter world.
TP: We are indeed. We’re committing humanity for the foreseeable future. I think this is an issue people don’t fully appreciate about carbon dioxide. The current pandemic, actually, this is quite a good example of when we stopped emitting, for example, particulates of pollution and these enhanced nitrogen dioxide levels from vehicles, we’ve all stopped driving, or at least we did in the early days or the pandemic. And we saw very quickly that the atmosphere would kind of clean itself out. The particulates would rain out or drop out relatively quickly. So on timescales of a few weeks, the standard levels of pollution would drop to almost to pollution free levels very quickly.
So the carbon dioxide is not like that at all. Carbon dioxide is a very well mixed gas in the atmosphere. And once we put it into the atmosphere, it’s basically, the enhanced levels of carbon dioxide stay there for, potentially, thousands of years. So when we emit carbon dioxide in the atmosphere, unless we can find very clever ways of sucking it out, which we haven’t really done so far, then we are committing humanity to a very long-term future of enhanced warnings. And of course all the changed weather patterns that go with that.
AS: What do these new and more serious results from the models tell us about how soon we might reach the two degrees C of warming, which was the alleged danger zone and how much time we have left to act upon this?
TP: Well, certainly, it means the two degrees C is happening much quicker, it will happen this century for sure. I think there’s very little doubt about that. I mean, if these results are right, I think it’s going to be very hard indeed to try to keep this one and a half degree target of Paris, and even two degrees may be pretty tough. And that’s why it’s important to remember there isn’t a kind of a formula which says for a given amount of carbon emissions we will warm by a certain amount. There is a range. And I think the latest results is suggesting the range should be pushed upwards to greater values. So I think it puts the emphasis on cutting our emissions to zero and hopefully even having technology which can suck CO2 out, although that may be very, very difficult to do at scale, as urgently as possible. I think that’s all I can say about that.
Alex Smith: You are tuned to Radio Ecoshock. I’m Alex Smith with our guest, Oxford professor in climate physics, Tim Palmer. Well, due to the pandemic and subsequent economic difficulty, some universities are closed while they try to remodel into online institutions. Budgets will be tighter than ever, even at research institutions. You are talking about unification of weather forecasting and climate modeling, but that’s expensive and it needs supercomputers in some cases. How can we do that?
Tim Palmer: When you say it’s expensive, I would say, well, I don’t know, compared to what? Just the other week, Elon Musk sent a rocket up to visit to the space station, which I looked it up. We, the taxpayers, and I think that includes Canada, the U.K., the U.S. and other countries, have spent $150 billion on. $150 billion. Now, I don’t know exactly what science has been done in the space station that warrants $150 billion. But I think if we could spend a tiny fraction of that, 1% maybe would be enough, on trying to pool our international resources to create, I said this on another interview, instead of an international space station an international climate station, and this doesn’t have to be in space. This can be on Earth, funded with one or two of those $150 billion, the object of which would be to develop climate models with a level of precision and accuracy and resolution that is just impossible to attain as an institutional level.
Then, I think actually, we could beat down a lot of these uncertainties about how much warming will happen, what will happen to clouds and above all, what will happen to local weather patterns, which is crucial for how society adapts and adjusts to climate change. So I don’t think this is an expensive… Compared for example, to what we spend on space, on astronomy or, indeed, particle physics. Think about the large Hadron Collider, again, that’s another multi-billion type of project. I think we haven’t treated climate modeling with anything like the same level of ambition and commitment. So I don’t think in the big scheme of things, putting some funding into an international climate prediction center would be expensive compared with many other projects that we do fund internationally.
AS: You’ve raised the difficulty of clouds several times and pardon a lay person’s questions, but we have mountains of satellite images of earth over the decades. Why can’t we just calculate the amount of cloud cover and then see if they’re cooling or warming the place?
TP: Well clouds are of course, very… You probably can look out the window and you see cross are extremely complicated things. And it’s actually extremely difficult to observe whether the total cloud cover… Because this is the point, the question is, is the total cloud cover of the planet systematically changing in one direction or the other? Is it going up or is it going down? And there’s a subtlety to that because, well, you need to know separately, whether low level clouds are going up or down in amounts, not in altitude but in amount, and whether high level clouds are going up or down as well separately. This is actually a very difficult thing to do. There is no overwhelming evidence yet that the clouds are systematically changing, but one of the problems is that satellites don’t last a long time stuff like instrumentation only lasts a few years.
And so if you go back over some decades, you have to try to blend together data from different satellites. And the problem is that the calibration, you have to then calibrate data from one satellite with data from another to make sure you’re not just seeing something spurious because the instrument is slightly different from the one before. And the changes, if there is a change in cloud cover over the last, say, 20 years, it’s smaller than the uncertainties due to these instrumental calibration effects. So we can’t, the bottom line is, we can’t yet appeal to observations to tell us whether climate change is systematically affecting clouds or not. The effect is currently too small.
But all the models suggest this is something that becomes more substantial when climate change becomes more important or bigger in amplitude. So by the middle of the century, I think we will know one way or another whether clouds are going to be our sort of savior, which means that the low level cloud cover increases and sort of offsets the climate change signals. Because if low level cloud cover gets more and more extensive, then these clouds are reflecting sunlight back to space. So they’re offsetting the warming. Or are clouds going to be our nemesis by going in the other direction and starting to evaporate as climate change kicks in, which means that they will be, if you like, because they’re no longer now reflecting sunlight back to space, they’re adding to the climate change problem.
So we’ll have to wait, I guess, maybe another 30 or 40 years, if we’re going to find this out from observations. And so if we want to know the answer to this question sooner than that, the only other option is to develop models where the representation of clouds is good enough to answer these questions with certainty. And at the moment, the models aren’t resolved enough. They don’t have enough detail to be able to answer these questions other than in a very kind of probabilistic way. But as I say, the latest results suggest things aren’t looking so good. It looks like, based on the fact that the weather forecasting test, if you like, that we’ve done, Met Office, I should say, have done, most recently suggests that clouds are not going to be our savior. They’re going to go the wrong direction. In other words, they’re going to start to dissipate as climate change proceeds.
AS: In 2009, the 4 Degrees and Beyond International Climate Conference convened at Oxford and although many top scientists presented there, I recall thinking they were still the daring minority outside the mainstream. Tim Palmer, how does that look now with your predictions about the newest models?
TP: Look, I don’t want it to overstate things too much by saying, categorically, we know now for certain that climate change is going to be catastrophic. It’s going to pose a… It’s going to wipe the human race out or something like that. There are still big uncertainties. And anybody that tells you categorically climate change will be on existential crisis. They don’t know, they don’t understand the uncertainties. And there still are a big uncertainties and most of the uncertainties hinge around clouds. But all I’m trying to point out is that we just have yet another new piece of evidence, which has come from a completely different, if you like, a completely different area of research than has traditionally been the case in climate. So the new area is to test the climate or test one particular climate model in a weather forecasting type of environment, which as I mentioned earlier, is not something most climate models can do.
And what we found is that the change that were made, the clouds, which would cause more climate change also improved the weather forecast. It made the weather forecast better. So this is not categorical proof. It’s not like I have a mathematical proof that therefore clouds are going to be an amplifier of climate change and they’re going to make things catastrophically bad but I’m just providing another element if you like, or another piece of work into the mix, which should make us and should make the politicians think even harder about how to cut our emissions with all possible speed in the coming years ahead.
Alex Smith: From the University of Oxford, we have been speaking with the world leader in climate and weather forecasting. Professor Tim Palmer. Find links to his work in my show blog at ecoshock.org. Tim, thank you so much for sharing your valuable time with us.
Tim Palmer: It’s a pleasure. Thank you very much.
Check out the Radio Ecoshock website. We’re at ecoshock.org.