[Scott] Coming up on "Energy Switch," the computer models that predict climate.

- Climate modeling, it's really to explore what could happen and then contrast them, okay?

So we could really cut emissions.

This is how low we could stay and look at the impacts.

You know, how much does that affect people?

Who does it affect and how does it affect them?

There's already going to be a certain amount of climate change and so we already need to be adapting and different places need to understand what they need to be adapting against.

- That's why we use models because we aren't going to mess up the world by-- - Yeah.

[all laugh] But we do it in the past.

So this is our tool and we can change something in the future without messing the system up and see what would happen.

And we are giving information to people so they can choose their future.

[Scott] Up next, we look at climate models, how they work, and what they tell us.

[Announcer] Major funding for this program was provided by Arizona State University.

Shaping global leaders, driving innovation, and transforming the future.

Arizona State, The New American University.

[upbeat music] - I'm Scott Tinker, and I'm an energy scientist.

I work in the field, lead research, speak around the world, write articles, and make films about energy.

This show brings together leading experts on vital topics in energy and climate.

They may have different perspectives, but my goal is to learn, and illuminate, and bring diverging views together towards solutions.

Welcome to the "Energy Switch."

Climate models are large, complex computer programs using atmospheric science and temperature and weather data to try to simulate Earth's climate.

Climate scientists built them because they can't run physical experiments on the actual atmosphere.

So instead, they input carbon emissions, warming, rainfall, or other data and see how the model responds to try to predict the future climate.

We'll learn how they work and some of their limits with Natalie Mahawald.

She's a professor in engineering and director of graduate studies for atmospheric science in the College of Engineering at Cornell University.

Alan Robock is a distinguished professor of atmospheric science at Rutgers University and co-director of the Rutgers Impact Studies of Climate Intervention Laboratory.

Up next, an inside look at climate modeling with two climate scientists.

Well, welcome.

Natalie, Alan, we're really glad you're here.

Just real concisely, why should our viewers watch this episode?

Natalie, start with you.

[Natalie] Climate modeling does not sound that interesting, but it's really one of the most important tools we use to try to understand what climate change has occurred and what will happen in the future.

And we combine it with a lot of other tools as well, but that's the main reason.

- Alan, what do you add to that?

- Chemists have laboratories with test tubes and physicists have accelerators, but our laboratory is the whole world.

And so we can't go and do experiments there.

So we write down equations that describe how it happens and we simulate it in a computer.

And that's what we do.

- Equations of physics.

- Of physics and chemistry and about water and clouds and wind.

[Scott] Interesting.

- Ocean currents.

- A lot going on.

We hear about weather all the time.

We can see weather out our window.

We feel it.

What's the difference between climate and weather?

- Weather is the instantaneous state of the atmosphere.

Like today it's, right now it's 66 degrees, but climate is the statistics of it.

So what's the average?

Traditionally the weather service uses 30 years and they average the last 30 years and that's the average.

But it's not just temperature, it's amount of precipitation.

It's what the wind is, where the storms are.

[Scott] Right.

- And the distinction between weather and climate can be confusing.

For example, you can have colder time periods under a warming climate, right?

There's no problem with that, right?.

Because it's the average that really matters for the climate.

And individual weather events can be colder or warmer even under a warming climate.

[Scott] Right.

I mean, we hear quite often, I'll read, it's the warmest year on record again.

- Yeah.

- The second warmest.

How much has that happened over the last decade about?

- It's warming by about two hundredths of a degree per year.

- Two hundredths of a degree?

- Yeah.

- So two tenths over a decade.

- Yeah.

Yeah.

- In Celsius.

[Scott] That seems pretty precise.

How how are we able to know that?

- I mean, we have a lot of observations covering all over that we get.

Then, you know, we do have to combine them intelligently because we're missing data over the ocean.

- Sure.

- So we have to kind of, you know, we don't have that much data over Antarctica, Greenland.

So somehow we have to, you know, extrapolate what we know in the most consistent fashion.

And so that's done by different groups and they do it differently.

And then we take a look at how the temperature trends have gone.

- Right.

- And you want to measure temperature at places where there's not a human influence.

So you don't put a thermometer right in the middle of a city.

You find places that are outside of cities that are affected by the global climate, not by some local action.

But it's not just temperature.

Again, we measure how much precipitation do we get.

We measure how cloudy it is.

We measure what the wind is.

And so there's lots of different elements of the weather, not just temperature.

[Natalie] We don't... The climate to a very large extent is really driven by energy balance.

Okay?

It's really about the energy in the system.

And the energy depends on, say, the cloud distribution.

It's very important if you have clouds, because clouds are white.

And so for example, the ocean is very dark blue.

These are really different colors.

And so the incoming solar radiation, the sunlight, you know, reflects off clouds, but is absorbed by the ocean.

So really different impacts there.

And so that's what becomes so important on the longer time scales.

It's just small changes in the energy budget.

- Right.

[Alan] So there's a graph here of... We break the world up into different boxes.

Each box we calculate the wind and the temperature and the humidity and cloudiness.

And in the box next to it, you do it the same thing.

And then you keep everything constant for say 15 minutes and you see how things are going to change.

Then you take those answers and you put them back in and you do it again over and over again.

That's how we actually do it.

And so the graph over there is a different grid showing you like higher resolution over the United States, over North America.

So you can get more details there.

You do that to save computer time, you don't make the whole world very high resolution.

- What's the stack look like?

How fine do you stratify the-- - Well, the models that we use typically go up to maybe 60 or 70 layers in the atmosphere going up to above the troposphere where it's where we live, where there's weather, but in the stratosphere, the mesosphere we go all the way up there because these layers interact also.

- Sixty to 70.

[Alan] Yeah.

- And how many grid cells are there that blanket the earth on most models today?

- The model we use is called WACCM, the whole atmosphere community climate model.

- WACCM.

[laughs] [Alan] WACCM, yeah.

W-A-C-C-M.

- That was not a acronym, that must have been... [Alan] Yeah, yeah.

I don't know.

I don't know how many.

I have to calculate how many.

If it's one degree, it's 360 times 180.

So I don't know what that number is.

- And times 70.

[Alan] Yeah.

Yeah.

- Yeah.

So you got pretty decent number of cells, if you will, that you're already trying to handle.

And those will just get better with compute power.

- Yeah.

- You mentioned wind, temperature, clouds.

Of all of those, which one do we probably understand or able to model the least?

[Natalie] The clouds.

Just look outside.

The clouds are constantly changing and they have very different properties.

And understanding what causes the cloud to do what, and then also their interactions with the energy coming in and out of the system.

That's a really hard process.

And so we don't have like, you know, Newton's law about that.

We have to, we call it parameterize, but really what we're saying is, you know, what assumptions am I going to say controls the clouds?

- Right.

- And different people have different views on what the right way to do that is.

And so they put them in their models differently, basically.

- So clouds are very small in space, but also in time.

And so if you have a grid box like one of these, you have lots of clouds there and you can't model each cloud.

So you have to have the average of cloudiness based on some temperature or humidity considerations.

- Yeah, yeah, yeah.

[Alan] So it turns out if you make the grid small enough to just model clouds, say one to three kilometers or a couple miles, then you can model cloud systems individually and you get a much more accurate forecast.

But we don't have big enough computers to do that globally right now.

We're getting there.

- How many accepted sort of climate models are there in the world?

Give me a feel for that.

Is it five or 50 or 500 or?

[laughs] Yeah, no, it's in between there.

So between 20 to 40.

- Okay.

- And there's more and more models.

And a lot of what that is, is actually more and more countries investing in their own model, which is really good because they shouldn't just believe us, okay?

They should have their own models, their own internal ability to make predictions and contribute to the science.

[Alan] And there really aren't 40 different climate models because when you start a new modeling scheme, you usually take an existing model, which is available for free, and then you modify it a little bit and have your own version of it.

And even the big modeling centers in the United States, National Center for Atmospheric Research and Geophysical Fluid Dynamics Lab, they talk to each other.

But so there are different families of models.

There aren't 40 independent models.

- Now, what's the range of future outputs from the great modelers in the world?

- The range in the future can actually be fairly large.

Okay?

Over the past 20 to 40 years, we have substantially more observations.

So we compare what's going on and maybe we take a look at how the seasonal cycle evolves or how different regions respond differently.

To try to figure out which models are doing a better job.

And then we take a look at the future.

We might only use the models that do a good job with that when we go into the future, right?

We might sample the models for quality, basically.

But we do like that the models give us a different result.

And we use that information to figure out which- - Why?

- Of the responses, why?

Why are the models different?

But also, the models will agree on some things and you're like, oh, you know, it really doesn't matter what you do, you're going to get that answer.

Okay?

Whereas in other regions, it's much more sensitive to what's going on.

So then we find out what is more robust and what is less certain.

And then we always combine these results with observations, with kind of simple idea models, with, you know, all these other results.

And we don't blindly believe the models and the climate model predictions.

We really combine it with all the different ways that we take a look at the problem.

[Scott] Interesting.

I mean, I guess classic scientific method would say, "Well, here's my thought, hypothesis, blah, blah, blah, and I'll set up an experiment and test it."

Basically, we can't do that.

You can't set up an experiment and test whether the model was right.

[Alan] We do it in the past.

So for the past 150 years, we have a very good idea what the concentrations of different gases and particles were, when the volcanic eruptions were, and we have observations of what happened.

So we run the models, see how well they simulate the past.

- Okay.

How do they do?

Are there places where great and others places where you're going, gee, I wish we could get that a little better?

[Alan] They do a really good job, but of course they're sort of cheating because they know what the answer is.

If they get a very strange result, they say, "That doesn't look right.

Let me go and find a bug in the model or fix it."

- Oh.

- And so they won't show you the first runs they did.

So it sort of doesn't assure you that they're going to be right in the future, but that one of the big unknowns is how many particles there were in the atmosphere.

How much pollution was reflecting in sunlight.

And some of them put more or less in to make sure that the temperature rise is about right.

And we're not quite sure.

The observations aren't very good, even today of particles in the atmosphere.

[Scott] Right.

Your models are also looking at future tropical cyclones.

They're looking at drought conditions, I would assume, fire weather, different kinds of big extreme weather events that matter to us.

- Right.

- What's the range of future outputs there?

And how should we act based on that range?

You know what I'm asking?

- Yeah, so extreme events are really kind of, you know, where the impact of climate change is really felt by humans.

- Right.

- And on the other hand, they're very rare.

So our statistics are terrible.

When you go historically because they're rare, they're harder to do.

And then the models also have a harder time because often they're, you know, really short time period or really smaller areas.

And so it's a harder problem and it's a more important problem.

- Right.

[Natalie] And so I would say that's where a lot of the evolution recently has been is really trying to get extreme events better.

[Scott] Okay.

- Some things are easier.

Extreme heat happens to be easier.

- Yeah.

- Precipitation is harder, right?

- Right.

- Or, you know, some models, I'd say a lot of the climate models really can't resolve the tropical, the hurricanes, which matter so much for people.

So I would say we're at different stages in that.

Fire, for example, you know, put fire into the model to try to get predictions.

And we're not there where you want to believe everything that comes out.

We can do maybe fire weather a little bit better.

[Scott] Okay.

- But the problem is also, you know, how complicated a process, you know, humans suppress fires and human light fires and-- - Right.

- So it's complicated.

I would say we're moving towards, say, getting fire episodicity right in the models.

So they're all at a different stage.

- So when I read, this is worse because of climate change.

Is is that right?

- So they shouldn't phrase it that way.

I mean, there's a higher odds that such an event would occur under climate change.

- Okay.

[Natalie] You can say that.

Or you can also... Sometimes people are using a shortcut.

- The reporter is or the-- - Yeah, the words.

Like you're saying it's climate change.

And most of the time what we say is, you know, that extreme event increased in likelihood by this amount.

And I think that's a fair way to pose the problem or decrease.

- One thing that really annoys me, I turn on the TV weather forecast and the forecaster says it's going to be 10 degrees warmer than it should be today.

And I wrote to them, I said, "Don't say should be.

What you mean is it's going to be 10 degrees warmer than the average temperature for this day of the year for the past 30 years."

And we use the term normal for average.

It's a statistical term.

It's jargon.

So if it's 10 degrees warmer than normal, that gives the wrong impression that it thinks normal is not just weather it's going to be should be.

The good weather forecasters say it's going to be 10 degrees warmer than the long term average.

That's a much more accurate way of saying it.

- Interesting.

Yeah, I like that.

The language matters, doesn't it?

- Yeah.

Yeah.

[Scott] It really does.

And I think it helps people learn too.

We've talked a lot about climate modeling.

What's a climate scenario?

- Well, a climate scenario is basically a whole bunch of assumptions and put them into a socioeconomic and kind of an emission model and create scenarios of what could happen.

So really what we want to do is kind of span the possibilities.

Like nobody works together.

Everybody emits as much as they can.

That's one scenario.

And then we have to adapt.

Another scenario is we all work together and, you know, the poor countries get richer and the rich countries help the poor countries not emit so much.

- So the scenario includes the human component.

- Absolutely.

[Scott] The socioeconomic human component.

[Alan] That gives policymakers a chance to choose which scenario do they want?

What future do you want?

We don't tell what people what the policy should be.

We tell them if you do a certain policy, this is what kind of world you'll get.

[Scott] Interesting.

What are the scenarios estimating?

Carbon emissions or is it more complicated than that?

- What the climate models used to want is just the concentration of CO2 and methane.

Now they're more sophisticated.

They want the emissions of CO2 and methane.

So they calculate how much carbon's going to be into the forest, how much is going to be dissolved in the oceans and how much is going to last.

And nitrogen and methane too.

So all of those are now calculated in the chemistry of these models.

- And they even, they now, a lot of the models are going farther back where instead of saying, well, this is how much, you know, deforestation will emit, it's actually they take the land use and you know, oh, okay, if you have that land use conversion, this is the climate models predict how much CO2 is emitted from the forest when the forest is cut down or you're using the forest for harvesting what the carbon is.

So the models are getting more and more complicated to try to have more of these interactions as well.

[Scott] Do they model, I mean, CO2 is food.

Do they model that actually plants are taking up more CO2 as well?

- The crop models include carbon dioxide concentration and how that's going to affect them.

- Right.

- Why do scientists use a range of these scenarios then?

What's the advantage of a range?

- Well, it's really to explore what could happen and then contrast them, okay?

So we could really cut emissions.

THis is how low we could stay.

Or we could continue, you know, business as usual, we often say.

And so then we can contrast them and look at the impacts, you know, how much does that affect people?

Who does it affect?

And how does it affect them?

And there's kind of two reasons for that, maybe to excite people into more action and more mitigation, but also prepare people because there's already going to be a certain amount of climate change.

And so we already need to be adapting.

And different places need to understand what they need to be adapting against.

And that's a really hard problem also.

[Scott] Okay.

I think if I remember right, in the IPCC, Intergovernmental Panel on Climate Change, they ran four scenarios.

A low one, a couple in the middle, and then a real high one.

Is the high one still realistic?

- But people debate that.

- You want us to predict human behavior.

We have hard enough time predicting the climate system.

[Natalie] But the high one assumed quite a bit of coal usage.

- Coal.

- On the other hand, the land use scenarios in all of the scenarios, I would argue are too low.

They don't have as much tropical deforestation as we see.

That doesn't fully offset, but it partially offsets.

So for the next IPCC, we'll probably use a little bit of a lower trajectory.

[Scott] For the high one.

- For the high one.

- If you think about the most likely scenarios, we've been talking a lot about scenarios, what are they telling us about the future?

[Natalie] We'll be in good shape if we stay below three degrees C, including the 1.5 we pit.

[Scott] Okay.

- Yeah.

I think that that's going to be hard.

- Three degrees you said.

- Yeah, to peak at three.

- Yeah.

- You really have to have some, not just emission cuts, but also the carbon dioxide removal.

If you're still going to allow, you know poor people to come out of poverty, which I hope we will.

[Scott] Right.

Right.

Right.

Right.

- I mean, there really has been kind of a success that I don't think people really celebrate as much as they should, that the absolute number of people who live in poverty has actually gone down.

And that's amazing and we should celebrate it.

It's just incredible.

Now, of course, the way they've done that is the way we did it, which is emitting a lot of CO2.

But that's a huge thing.

You know, sometimes people I think are a little bit naive.

They think that we can solve climate easily without thinking about the implications for poverty, for biodiversity, you know, for all these different things.

They're all so connected.

- Yes.

- And that we have to think about it in the big picture.

- Absolutely.

It's trade-offs.

It's an optimization problem is what it is, right?

- And we keep hoping for synergies.

We always want to talk about synergies, but at some point there's trade offs, and we just have to face that, yeah.

[Scott] Yep.

What could be done to improve actual climate modeling, modeling itself?

- We need to continue observations with the NASA satellites and NOAA satellites.

Keep replacing the ones that are old.

So we have to keep the observational network going.

- And as well, the in situ, the actual temperature measurements, not just the satellites.

- On the ground.

[Alan] On the ground, the balloon, weather balloons, all those have to keep going.

And there has to be a support for those.

And then we need faster computers.

And so that's happening slowly.

Maybe quantum computers, maybe AI will help us make our models faster.

- I really think the AI is going to be a revolution.

It's already to some extent, it already is in weather forecasting.

And I think it's going to really change things.

I guess as an educator, I'm always excited to say that the future I find bright just because all these really bright young people are moving into the field.

And they're just so smart.

They just think about things totally differently than we do.

And you know, they can use these tools better.

So I do think that they'll make not just incremental changes, but with these new tools, kind of revolutionary changes.

I agree.

- I'm hopeful.

- I agree.

- We still need to support education and support climate science research.

The climate scientists have to be there to use the AI.

- So funding matters.

- Yeah.

Yeah.

And it's cheap.

The research isn't very expensive at all.

- Right.

Right.

Technology.

What have we not talked about that you feel is important for our listeners, Alan?

- Scientists are curious.

They're trying to figure out what's going to happen.

And they're also concerned about the impacts of climate change.

And so there aren't scientists who are trying to scare people or trying to give bogus information because of some political beliefs.

The scientific method is self-correcting.

So if you publish something, first of all, to get it published, you have reviewers look at to make sure it's okay.

And if you made the wrong assumption, then it'll be corrected.

My PhD dissertation was wrong.

It said natural variability was a major cause of climate change.

I had a very simple climate model and now I've got better data and I can say that forcing the greenhouse gases and volcanoes are more important.

So trust the scientists in the scientific method to give you an honest answer, the best that we can do.

[Scott] Okay, good.

Appreciate your saying that.

How about you, Natalie?

Anything we haven't talked about that you feel is-- - Something that I think is really important to emphasize is climate models are one of the tools we use.

And we use many different tools.

And you know, we throw all these hypotheses in there and we test the hypotheses and see if they work against a lot of observations.

And you can really just think of them as a tool to extrapolate our knowledge in some way.

But we don't rely on them for the projections or any kind of substantial, I don't know, changes in theory unless they're supported by the observations, by simple theories, by simpler models.

We need multiple lines of evidence always in order to improve our understanding of what's going on.

[Scott] Right.

What gives you hope in the climate modeling space?

As you look out a decade or so, what makes you hopeful, Alan?

- In the short term, I'm not hopeful because climate science is being attacked in the United States.

They're closing the main research labs that do the climate science.

So that makes me afraid and I'm not sure that it's going to be able to be recreated well after things change.

So that doesn't give me hope.

But in the broad scheme of things, things keep getting better.

We have better tools, better information, more ideas, bigger computers.

And so we make a lot of progress and we're going to continue to make progress.

- How about you, Natalie?

- I would say in some sense then the next generation and then also the spreading of climate knowledge outside of just the United States and Europe and Japan, that more and more countries are developing their own scientists and they have their own scientists.

And that, you know, spreading of the information and spreading of the knowledge I think is going to lead to better climate models, but also better understanding the climate models and use of the climate models in different regions.

- Yeah, I like that.

- I'm working with a group called Degrees in England that funds scientists in the global South to do climate research.

Then they'll be able to talk to their governments because they'll understand what's happening where before there weren't local experts.

So we're getting feedback and they're going to help produce better science.

- I love that.

Fantastic.

Well, look, really enjoyed the visit that we've had.

I appreciate you both being here.

Scott Tinker, "Energy Switch."

Thanks, Alan.

- Thank you.

- Really appreciate it.

- Thank you, Natalie.

- Thank you.

[Scott] Climate models divide Earth and our atmosphere into hundreds of thousands of three-dimensional cells.

Then use mathematical equations to represent the material and energy in each cell.

Climate scientists input variables into the equations like heat, precipitation, wind, and clouds, and use supercomputers to solve the equations for all the cells thousands of times to represent the climate far into the future.

Different models use different math and different assumptions of climate sensitivity to inputs, so produce different predictions.

Scientists model a range of low to high carbon emission scenarios, each one assuming different energy types, land use, population, and greenhouse gas emissions drivers.

Their intent, my guest said, is to help legislators and other stakeholders understand how following each scenario pathway could affect the climate, to motivate people to mitigate carbon emissions and to prepare them to adapt to changes in temperature and rainfall through the science of climate modeling.

[upbeat music] ♪ ♪ ♪ ♪ ♪ ♪ [Announcer] Major funding provided by Arizona State University.

Home to the Thunderbird School of Global Management, redefining management education to empower transdisciplinary leaders.

Arizona State, The New American University.