So what I wanna talk about today is how is this forecast of climate change made? We talk about global warming, but what does that mean for the regional scale? And we talk about climate change, but what does it mean for the weather we experience? And I'm gonna touch on those two things. But really, because my job is a climate model and I am a scientist, I wanna mostly give you a sense of how do we know these things? So here is an example of a simulation by a climate model and it starts our before the forecast. It starts out looking back from 1950 and going forward to the end of this century. And what's plot here is the anomaly in surface temperature compared to the end of the 1800s. So you see that we start on 1950, it wasn't that warm. And then as you go into the forecast, it gets warmer and warmer and warmer. There's a little spot in the North Atlantic that the model project is still gonna be cooler, but the rest is getting warmer. How is this forecast made? We use climate models. Let me give you an example of a model that we all use that is the statistical model. You save for retirement and they tell you that you can expect a 4% income coming from it year after year after year and it's based on things like the S and P 500. So you see, going up and that's how you decide how much money you're gonna put in your 401K. Now there's a reason why they tell you past performance doesn't necessarily predict future results, because you can see that the S and P 500 can do something very, very different. Not only that, but you get a sense that maybe the system changed somewhere around that year. God knows what? The regulation, the global trade, I don't know. But this teaches you that it's risky to base your projections on fast data. And so when we look at the climate, this is the series that tells you what the global temperature has been doing from 1850 all the way to now or 2012. You could extrapolate, you could say, it's gonna keep warming. But you don't really want to do that, cuz you don't know over what period you should be extrapolating. If you start it out in the 70s, it would look very different if you can start right now and we know that the system has changed. We know that there is something else that will probably make the climate behave in a way that is different from the past. And of course, that something is the fact that the composition of the atmosphere has changed. So you saw the ice core. This is the CO2 from the ice core, just from the same period i n the 1850. And then this attached rate is Mauna Loa, the record from Hawaii. So we are changing the system. How do we put everything together in a model that is not a statistical model? Where in much better shape than economists, because we have chemistry, we have physics, we have math. [LAUGH] They have math. So that's the apparatus that was used in 1861 to measure the interaction of CO2 with infrared radiation. So we've known about the fact that CO2 is a greenhouse gas since 1861. 1843, the laws of fluid dynamics and how to solve them in equations were also known. They're complicated, it took us some time to figure out how to solve them in a way that was useful. 1950, a computer in Princeton that did the first weather forecasting and that's where our models are coming from. A few years have gone by, these are the supercomputers that we now use and the kind of climate model that we now have is a much more refined one. So here is another little movie. This is from NCAR, a National Center for Atmospheric Research. What is plotted here is the water vapor in the atmospheric column. So you see that the tropics have a lot of humidity, obviously and you see all of the features. The things swirling around and so and so forth. If you are geeky enough to look at solid images of the atmosphere, they look like this, but this is actually made in those very big computer models. So we can use these models to look forward, but you wanna make sure that you've done your homework right. So you wanna validate them with the past. You wanna know that given the right forcing, the model is doing what you expect it to do. So Debra talked a little bit about the four scenes of the climate system. The very basic ones, like the geometry of the Earth, those don't change. Other things do change, so you wanna know what the sun has been doing. There's a sunspot cycle. If you're looking very far in time, then there are other issues with the amount of energy that the sun has been putting out. There are volcanic eruptions, there's anthropogenic emissions from factories and cars. There's also other forces like soot that come from, for example, from fire that can be naturally occurring, but also coming from deforestation and so on and so forth. The state of the land, whether you have a vegetated land, agricultural land. If you go from forest to barren land, that also matters. So you need to know what these forces have been. If you know those, you give them to these models. You can know what the models think the past was and you can validate your model. Everybody with me on this one? And then once you've done that you say, okay and then I'm making a wild guess about what we're gonna do in the future with those forcings and see what happens. So what the community has been doing is to do this exercise over and over and over again and we and the US have a few modeling centers. There's one in Princeton, there's the anchor that I mentioned before and others. There's the UK, China, all over the globe. People have been doing these exercise. So here is the average of all those simulations for the 1950 to 2000. And what's shown is that all the vagaries of climate kind of get averaged out when you put all those datasets together and you're left with an indication of what the forcing does. And so you see the dips coming from volcanic eruptions and you see the slow warming and then you warm. Now this is the business as usual scenario, where you warm a lot. And this is the scenario, which there's immediate action in reducing emissions. So there's uncertainty in how much we're gonna warm given a certain path of emissions, but that's much smaller than the uncertainty that comes from figuring out what choices were going to make and how were gonna change the combustion of the atmosphere? That's one number. That's the global mean surface temperature, and it's very fundamental, and it means a lot, and it's a good thing to have one number that we all agree is a good indicator of this important thing that is climate change. But there's a lot more to climate than one number, and so you wanna know that your models are doing the right thing at the scales that are important to the forecast, so the regional scale and other variables, not just temperature. So let me give you an example of work that I've been involved with. So the Sahel is basically the region just South of the Sahara desert. It's not an easy place to do some farming, but in particular, in the 70s and 80s, it was subject to a terrible drought. It was one of the longest drought in the observational record. So if you look at the time series of rain falling in the Sahel, these black line is the observations, you see that in the 50s, there was what we call a pluvial, a period of abundant rains and then 73, 84, the drought. If you're my age, you remember maybe We Are The World and other [LAUGH] attempts to mobilize the developed world to come in and help. So the other lines you see here are what the models do. So here in the blue line with the shading, indicates what a model that has old historical forcing and the red line is one instance of that hindcast. So they do an ensemble, meaning that they do it again and again with different weather, and just like every summer is a little bit different, but you can tell summer from winter, here's the same. Every realization, every example of climate that this model produces is a little bit different, but overall, you can tell where it's going and one particular realization looks eerily similar to what the Sahel historic record tells us. So you can use this to say, well, the models have scales on this variable and where does that come from? And so I'm not gonna go through the whole thing, but we know that this decline in rainfall was linked to the changes in the surface temperature of the ocean. The first idea that people had was that it was actually caused by overgrazing or other local practices, and because of climate models, we know that the cause of that drought was attributed to the ocean temperature. And those are partially just bad luck and partially a response to the forces that we put into the atmosphere, not as much the CO2 itself, but the soot and the sulfate and all the other pollutants that we put in. Now once you've done this and you say, okay, I trust this model, you look at the future, and this is, for example, what's expected for the region. So here's a map of Africa, and this is a plot of difference in summertime rainfall, which is the only time it rains here because it's a monsoon region. The greens are positive anomalies that we're expecting for the future, the browns are negative anomalies, and then here is the seasonality of those anomalies. And so the expectation is that at the beginning of the rainy season, the rain is gonna be less and then it's gonna be more at the end of the rainy season. So the rains are gonna come later, a longer dry season, but when the rains come, it's gonna pour. If you think of the past in the Sahel and you think, oh, there was that terrible drought, so you wanna say, oh, so if that was manmade, then it's gonna keep going in the same direction. That's really one of those instances in which the past is not necessarily a good model for the future. We really wanna put it in the model that has all the physics and has all the chemistry and has all the forcing and then see what happens. So in this case, we see that the expectation is not necessarily more drought everywhere, but a more complex pattern of change. And that's why you wanna really base it on a good climate model that was evaluated properly. So people obviously have done this work with a lot of detail for North America. I'm gonna just present a very few things. There are some things that are more controversial or difficult to measure, etc., but there's one thing that is pretty clear, and it's that it's gonna get warmer. And so there's aspects of that that you can follow through, and maybe there's an uncertainty in when things will happen, but at this point, not really that they will happen. So for example, for California, the snow is gonna start melting because it's warmer. And so depending again on whether we're gonna have less or more emissions, you're gonna have these different patterns of remaining snow pack. And of course, for California the snow pack over the winter is water over the summer. In terms of how to experience this warmth, this is a map that tells you by the time you hit 2080, your average summer, how hot is it going to be? And what this tells you is that in a large part of the globe, every single summer between 2080 and 2100 is expected to be hotter than anything the local population have experienced so far. So if you wanna see in a different way, this is the distribution. So zero here, means it's average over the 20th century, and you have some summers that are hotter, some summers that are cooler. When you look at the future, the idea is that most summers are gonna be way, way hotter than anything that we've seen. So that's true for the Southeast, we are pretty much in the same ballpark. For rainfall, the basic forecast is that the wetter regions will get wetter, the drier regions will get somewhat drier. There are places like the US where you are in between the places that get drier and the places that get wetter, and so there's a lot of work trying to figure out exactly where that zero line is and the difference between summer, winter and so on and so forth. The other thing that I wanna point out though is again, that even though some things are uncertain, some others aren't, and so when you think about what dries out the soil, there's two issues. One is how much rainfall gets to the soil and the other one is how much evaporation you get from the soil. So if you have a hotter day, you're gonna dry out your soil more readily. So in terms of worrying about >> Agricultural drought temperature again is the big player, Yes?. >> I understand that it's difficult to forecast weather more than a few days out because the system has chaotic dynamics and is very sensitive to initial conditions. So when it comes to climate as opposed to weather, how is it that we're able to make such good predictions so far into the future? Why don't we suffer the same pitfalls as weather predictions? >> We do and we don't. So we don't in the sense that you are absolutely certain right, that next February is gonna be colder than this July, right? And yet February is far away, right? But you know it because it's forced from the outside, is not really dependent on what happens today. You know that this sun is gonna just be such a huge driver that. Yes, one February might be slightly different from another. One day in February will definitely be different from the next, you get the snow storm on the 3rd or the 25th. But February is going to be a cold month. It's the same here, you have a forcing that is coming from outside which is a change in the atmospheric composition. That is such a strong driver that you can get away a little bit from it but not too far. So there's still gonna be the same thing as weather but at the times you'll have climate. So you might have a decade that it doesn't warm as fast. This last one, didn't warm as fast as the previous one. And it's an active research exactly what the cause is. Some people think that there might issues with. Are we putting in the atmosphere from China for example more Sulfate and other particulate matters that are important for the radiation budget. But another explanation that, it's still out there and probably the most likely simply that. This decade we lucked out and that might not be as lucky the next decade because things come and go. So, the same way. The atmosphere has weather. The ocean has weather, just slower. El Niño is an example of oceanic weather if you want, right? You have an El Niño. You have a La Niña. It takes a year or so to develop. It's a slow thing but it's something that is not forced from outside. But when you hit the system with something as strong as 700 parts per million of volume of CO2. There's no way you don't get that warming. The same way that once you hit You know the Earth with that particular insulation the fiber insulation you're gonna get winter in New York. All right so the only thing I want to say is that we are not gonna know the weather, right? So that part we're never gonna go, but we can know the statistics of weather. So we can know are we gonna get more heat waves? Yes. Are we gonna have, more downpours? And the answer is yes. And there are some kinds of weather that we don't know yet, we're working on it. So if you wanna know more on tornados, we don't know yet. Those are very much topics of active research. Finally, this is a quote that was uttered by Sherwood Rowland when he was getting the Nobel Prize for the work on the ozone hole. And we're now in the same situation. What's the use of having developed a science well enough to make predictions if, in the end, all we're willing to do is stand around and wait for them to come true? So there's three things that we can do about global warming. And the joke goes you can mitigate, so adapt or suffer. [LAUGH] So, mitigation means a future with reduced emission. And that is a big difference between business as usual and immediate action. That's the first thing. The second is adapt and so for example there's a lot of active research in understanding things like. What kind of plants do you wanna breed for a world that is so much warmer. So for example, for the Sahel again the work I'm involved with. You wanna know what is the reduction in yield for the kind of sorghum that they plant now? What are the characteristics of a different kind of sorghum that might be bred? So you have a model you go through that exercise and you try to guide the breeders. And I think we can close it here. Yesterday's though my favorite web comic came out with a cartoon about global warming. I'm gonna put it up. [LAUGH]. It's just looking at climate change in terms of ice age units. So what we are talking about 4.5 changes in global mean temperature at the end of the 21st century under business as usual. That's as much hotter in the future as we were colder at the peak of the last ice age. Thank you. [APPLAUSE]
