Climate Science 102: The Global Climate System and Climate Modeling

1273.3.Climate Science 102- Modeling Systems - Justin Bandoro

[00:00:00:12] Thanks, Christoph, for the introduction and welcome back all of the people who were here yesterday and all of the new faces I see in the crowd today. So as Christoph mentioned, I'm pre-presenting Climate Science 102 today. It's building on what we learned yesterday, but I'm going to do a short recap at the beginning to get everyone up to speed. So everyone should be able to follow along.

[00:00:26:03] And to introduce myself, I'm Justin Bandoro. I'm a PHD student in the program in atmospheres, oceans and climate here at MIT. So that's part of course 12. And I study part of the layer in the atmosphere that we learned yesterday, the stratosphere, so from 20 to 40 or 50 kilometers up.

[00:00:49:16] So today's topics, we're going to look climate forcings and feedbacks These are probably terms many of you have heard, but we'll look into them some more. And to be able to understand them, we'll also look at what are our global climate models, because these are being used for predictions and also, for better understanding what's going on or what has happened in the climate system.

[00:01:14:17] And parts of these that we'll look are what's known as parametrizations, so what we do on processes these are too small to be resolved in the model. And also, we'll look at uncertainty. And then, the last topic we'll touch on is the detection and attribution of human-caused climate change and how scientists can do this.

[00:01:40:03] So yes, a quick recap from yesterday. What this plot is showing is the rate Earth's radium fluxes. So you can think of this as the total energy coming into this system, and then, how the energy is distributed by different process within Earth's system and then the energy going out of the system.

[00:02:01:29] So as we learned, there's around 340 watts per meter squared. So all these units are in watts per meter squared, so it's an energy flux through an area coming in. Roughly 100 is reflected back out. That's through both clouds and through the surface from ice, snow and ice and snow from the land, so there's 100 going back out.

[00:02:26:20] 79 is absorbed by the atmosphere and so it's just from ozone and oxygen. And this absorbs in the ultraviolet. And then, as we learned yesterday, 161 reaches and is absorbed by the surface. And then, the surface--

[00:02:45:25] 79 what?

[00:02:47:09] Sorry?

[00:02:48:18] 79 what?

[00:02:50:00] Watts per meter squared.

[00:02:51:09] Oh.

[00:02:51:21] So it's intense radiation that's absorbed by the atmosphere. So the surface, you can see, it radiates up at 398, so double what's coming in from the sun, but this is because in the atmosphere, as we learned yesterday, we have greenhouse gases. And these absorb the longwave radiation from the surface and they remit it both upwards and downwards.

[00:03:25:26] So they reairiate it back down to the surface. And since the pressure or density of air decrease exponentially with height, there's more coming back down to the surface than what's going up. And there's also this little window here where some infrared or long-wave radiation can escape back out to space. So what's important to understand is this system should be in balance, so if you add up both these, 100 and 239 you get 340, so that's roughly 340 watts per meter squared coming in and 340 watts per meter squared going up.

[00:04:10:02] But what we're going to look at today is what different factors can disrupt this balance of incoming and outgoing energy from the system. So you can think about this as adding more greenhouse gases. What would this do? And then, also if you change the number of incoming solar radiation or you maybe add some more clouds.

[00:04:33:23] There are other particles in that material that will reflect the sunlight back out. And this is basically what climate forcings are. It's the changes in the amount of energy that enters or leaves the system. And this then alters earth's radiated equilibrium and can force surface temperatures, or temperatures of the atmosphere, to either rise or fall.

[00:05:00:13] And these can be either natural or man-made. And like we saw yesterday, we have the solar cycle, so that's an 11 year cycle where the output of the sun varies. And so that will change-- if you go back-- it will change this number coming in. And there's also changes in Earth's orbit like the milankovitch cycles we say yesterday, but those are much longer time scales from 20 to 100,000 years. And that will also effect the amount of solar insulation during different season.

[00:05:38:13] And there's also other natural ones. There's volcano eruption, so, again, like we saw yesterday, these are eruptions where you get input of sulfur dioxide into the atmosphere and this condenses and it's able to actually get up all the way up into the stratosphere. And then, it condenses and forms particles that are like little mirrors that will reflect the incoming solar radiation back up to space. These are all examples of natural climate forcings.

[00:06:06:17] And then, you look at man-made ones. These are ones shown here in red. And these are pollution, so aerosols. When we omit pollution, these can also form little particles.

[00:06:19:02] And then, they have the same effect similar to volcanic eruption, so it can reflect solar radiation back out. And these converse bond with it. What we'll see is a negative forcing, but then you have these positive forcings, such as from greenhouse gases, so if you put more greenhouse gases into that atmosphere, you'll have more radiation coming back down to the surface. Land use changes and deforestation are also man-made, or also called anthropogenic, climate forcings. And so yes--

[00:07:00:18] Can you match on the last word and deforestation?

[00:07:04:20] Deforestation?

[00:07:05:18] Yeah.

[00:07:06:00] So we can think about services like cutting down forests and that will change the surface albedo. So it's how much some more sun light will be absorbed by the surface as compared to if you have a canopy. That will reflect more. So it's the changes in albedo or how much is reflected from the surface.

[00:07:25:27] Thank you.

[00:07:27:09] And so positive radiative forcing, so more incoming energy warms the system while a negative one, which corresponds to more outgoing energy, cools it. And qualitatively, these are given as changes in the energy fluxes, energy flux at the top of the atmosphere. But one more thing to take into account is that this does not entirely predictable the climate response. And this comes back to feedbacks, which we'll touch on later.

[00:07:56:17] But I want to first give an analogy to what a climate forcing is and this is a simple one that will help you understand. So you can think of having a bucket with water flowing into it, and then it fills. And then if you have a little hole at the bottom with water flowing out of it.

[00:08:19:27] After a while if we leave this on at a constant rate, we will reach a certain level where the level will be constant and then, the amount of water entering will be equal to that coming up. And if you can think of this in terms of the earth system, the water flowing in is the incoming solar energy, what's coming up is earth's outgoing long-wave radiation, so that's what's coming back out. And this level here is analogous to the global heat content, so how much heat is in earth's system.

[00:08:55:28] And then, studies state, once again, the amount entering is the same. as the amount leaving. However, you can think of these examples for radiative forcing it's-- say if we were to block this off a little bit and make the pole at the bottom smaller, that would cause the level in the bucket to rise. And so it will be more water in the bucket, so it will rise until it reaches a new equilibrium. And you can think of this as a positive forcing such as if we introduced greenhouse gases to the atmosphere where it impedes the escape of long-wave radiation to space, so you have more of it coming back down.

[00:09:38:11] And then, the level rising will be similar to the surface temperature rising. So this is a good analogy. So what are the different forcings?

[00:09:49:12] So this is a figure from the most recent IPCC report. And as we mentioned yesterday, it's an intergovernmental panel on climate change. So this is where they assimilate all the reports or the currents understanding of science of climate change and to work report every five to seven years and this one was in 2013. And this is showing quantitatively the radium forcing associated with different anthropogenic and natural forcings. And this is all relative to 1750.

[00:10:26:06] So on the top here, we have well mixed greenhouse gases, so we have CO2 and all the other ones, which includes methane, nitric oxide and halocarbons. What's important to also note here are the confidence intervals associate with this. This shows the 5 to 95% confidence interval, so you can see how much certainty we have on that and with the amount of radiative forcing associated with these particular gases. So we can see that CO2 has around just over two watts per meter squared, so that's the initial energy coming into this system and these other gases from all or half of them are a bit less, but it's still net positive.

[00:11:20:06] And then, there's also the fact of ozone and how it's pulled from the stratosphere and the troposphere. And then, go through all these, so there's surface albedo and that's associated with the land use change I was talking earlier. And you can see for some of these, such as the aerosol radiation interaction, so these are emission of aerosols and population.

[00:11:45:00] And you can see just how big the uncertainty is associated with this. And this is because we're not sure how these aerosol particular effect cloud formations and where these clouds form. So they're low or high level clouds and this will effect their magnitude and the sign of the radiate forcing, so you can send them when it reaches over to zero for these aerosol cloud interactions. But if we sum up all the anthropogenic ones, we have a net positive just under 2.5 and with the uncertainty from around 1, 1.2 to just under 3.5. Anything compared to what we know of the forcing from the sun, so this is a change in the solar and radiance, actually this includes quote orbital and the solar cycle. And you can see since the 1750 it's very small compared to the total anthropogenic forcing in the climate system.

[00:12:55:27] The shade areas represent-- the shading is that some certainty?

[00:13:01:10] So this shading is just showing the troposphere versus the stratosphere.

[00:13:07:12] The shading itself represents uncertainty, doesn't it?

[00:13:10:14] Oh, yes.

[00:13:12:12] As compared with the solid bars.

[00:13:14:16] Yeah. I forget. There's uncertainty associated with that as well. I can't remember, but the solid bars are the 5 to 95% confidence interval. The next thing that we going to talk about are forcing feedbacks. So these feedbacks, they're showing a little chart showing what they are. They're internal climate processes that can either amplify or dampen a response to initial forcing.

[00:13:47:24] So say if we have any of these, if we go back, if we had any one these forcing how would the climate system respond, because, as we'll see, where there's many different feedbacks where it can either amplify, so it means the surface temperatures might get warmer than they initially were from the initial forcing or they might be dampened. So that might prevent it from warming as much from these feedback processes. So the positive feedback will increase the initial warming and the negative feedback will reduce the initial warming.

[00:14:21:08] So positive feedback here is a bad thing, isn't it?

[00:14:26:07] I don't know if it-- Well, it's one that amplifies the initial response.

[00:14:31:11] But since that means a warmer climate, in the present contact, it's a very bad thing. So that's something that's bad called positive. It's just difficult for me to wrap my head around.

[00:14:47:11] Oh, I see where you're coming. Yeah, you can label it, but here we're going to stick to positive and negative just because we'll see how it works, just the amplification bit. So an example of positive forcing or positive feedback is this ice-albedo effect.

[00:15:02:20] So say we have a positive radiate forcing, this causes surface warming. And then, associated with this you have melting of ice, glaciers, and snow. And then, that you can see, a little of after the ice retreats.

[00:15:17:27] You have a much darker ocean and you have a darker land. And that'll absorb more of the incoming solar radiation. And that will come back and amplify the initial surface warming, so this is an example of a positive feedback.

[00:15:34:19] An example of a negative feedback has to do with a temperature lapse rate. So like we saw yesterday for the structure of the atmosphere, so you can just think of this blue curve. The temperature decreases in the troposphere from the ground up to the tropopause. And what we know--

[00:15:51:03] What's the troposphere? Sorry.

[00:15:52:01] Oh, so it's the layer from zero to ten kilometers.

[00:15:57:15] It's below the stratosphere?

[00:15:58:20] Yeah, below the stratosphere.

[00:15:59:09] OK.

[00:16:00:01] Where temperature decreases with height. And in the future climate we know that the amount of warming isn't uniform across all the troposphere where it warms more in the upper troposphere compared to closer to the ground. And this has to do with changes in water vapor.

[00:16:22:16] So this increase is higher up and then at the surface and what this causes-- yes, so the warmer air up high can radiate more, heat away the space more easily. And so that means you're losing more energy or you can think about it as losing more energy then if it warmed if you just shifted it and it warmed uniformly with height. And so this is primarily in the tropics, which has most of the area of the earth. And so we can think about this, you have introduce a positive radiative forcing, you get troposphere warming.

[00:17:02:04] You get a decrease in lapse rate. This is the lapse rate, so decrease means it doesn't fall off as much of a height. And this means you get more increased cooling to space. And then, as well, feedback to dampen the initial warming.

[00:17:19:11] So there are examples of all these feedbacks in the climate system and that's what this figure shows. A positive is one where it intensifies the initial warming. And then, the negative is the negative feedback. So [INAUDIBLE] we saw a snow-ice-albedo, but there are also other ones with heat, and permafrost, and water vapor changes, cloud changes, ocean circulation, and this bottom part here might be hard to read, but showing the time scale associated with all these different feedbacks.

[00:17:54:01] So it ranges anywhere from hours to centuries long. But how we estimate earth's response with so many of these feedbacks going on? You can see just from here it's a very complicated system. And this is where global climate models come in.

[00:18:15:21] So given how diverse and all the processes in the system, how do we make a qualitative assessment of global climate change? So you can picture the earth if be divided up into small little grid in 3D like in latitude, longitude and also in height. So you have a 3D grid.

[00:18:37:28] And then, we know the governing equations for fluid motion, radium transfer, and then, we also have chemical composition. And then, we can calculate all these equations and how it evolves with time through partial differential equations in each grid box. And then, we can solve it.

[00:18:58:09] And there's also processes that are, like I mentioned earlier, they're parametrized, so these are ones that are sub-grid scale. And I'll touch on this in a few slides on how we can deal with these processes that we can't resolve. So if you think about It, if we have an area that's maybe 200 by 200 kilometers, we're not going to be a resolves a cloud that's significantly smaller than that area, like a single cloud.

[00:19:28:15] So we've got to figure out how we can deal with these processes. So these climate models are fully coupled and this means that they have different systems working with each other. And so there's an example, you have the atmosphere, ocean, land, sea, ice and by coupling they're talking to each other. So they can send information about what's going on in one to another system, so that the whole earth system is in sync and coupled or talking to each other.

[00:20:08:05] And this is really neat, because we can apply changes in external forcings, so say if we just in the solar cycle you can see how the system responds to just the solar cycle or we could add in greenhouse gases and see what happens and similarly volcanic eruptions. And this is interesting, because we only have one earth now and we can't these cool experiments with it, unless we find another planet that has the exact same conditions, but maybe does have greenhouse gases.

[00:20:45:23] So this provides a virtual laboratory for experimentation. And this is showing the complexity of global climate models over time. So is showing at the first assignment report, so the IPCC's first assessment report.

[00:21:04:23] And then, going through to AR4. It doesn't have AR5 in this one, but you can see the grid scale changes from these grid boxes were around 500 kilometers. And now, they're below 110 kilometers.

[00:21:18:26] And then, going through with time. At first it was only an atmosphere and an ocean. And then, they started adding in different components to the model.

[00:21:27:21] And you can see the sea ice and then, added coupling, aerosols, dust, carbon, biogeochemical cycles, ice sheets, marine ecosystems, adding in an upper atmosphere, a and also atmospheric chemistry. So they've gotten more complex with time. And the skills with our computational power, because that's what allows us to have all these complexities in the system.

[00:21:58:16] When you say these different areas talk to each other you mean they influence each other?

[00:22:03:25] Yes, yes, that's exactly what I mean.

[00:22:06:11] OK.

[00:22:06:23] So like the temperature in the atmosphere can influence the temperature at sea surface of the ocean. Similar like that. So they give an example of parametrization like I talked about before. These are these processes that occur at small spatial scales that these models can't resolved.

[00:22:45:27] So I'm giving an example here. So if we just took a photo of the sky and we see these clouds, and we dived these up into grid boxes. So you can see in this grid box, if we look in terms of total cloud area in the grid box, this one has very little, so we can put this as 0.1 the fraction of clouds. And then, this one's 0.2. This one has a lot more clouds, so it's 0.7.

[00:23:12:05] And you can represent this one as 0.5. And so to do this the physics needs to be represented. And then, the grid box.

[00:23:21:11] But it also is empirically supported, so that we know what we're approximating it by is right. And so examples are cloud processes, radiative transfer, and boundary-layer processes. So how you would do this is in any given grid box we know the average temperature, water vapor concentration, and pressure.

[00:23:43:25] And from these three variables, we can get an estimate of the relative humidity in a grid cell. And then, we can use observations of how we can derive statistical relationship between relative humidity, temperature, and cloud formation conditions. So this is what we can observe in the real world. And then, we can transfer what we know into the model to parametrize this. And it's also important whether it's over land or ocean as well. And from this we can get a cloud fraction in the grid cell.

[00:24:23:07] What does that mean, cloud fraction in the-- oh.

[00:24:25:12] Oh, sorry.

[00:24:26:06] The percentages of clouds.

[00:24:27:21] Yeah, so it's the fraction of percent. What percent of the grid box is covered by clouds.

[00:24:33:27] All right, you covered that.

[00:24:38:07] Yes. So the difference in how these sub-grid scales processes are parametrized is an important reason why climate models differ from one another. So you can have group, say over here in the United States having their own model. Then, another group somewhere in Europe with their own climate model. And they may choose to represent these processes in a different way. And that can have an effect on their results and their predictions.

[00:25:10:24] And that's what's shown here. So the CM5 is a model inter comparison project. And in here there's roughly 20 different modeling groups.

[00:25:21:14] And so to walk you through this, the black line, this is global means surface temperatures, anomalies relative to 1986 to 2005. So these three black curves are showing observations. And you can see, the spread around it is the historical model spread, so how these models were able to historically capture the changes in temperature.

[00:25:49:18] And then, this is a projection, so they can just keep on running into the future. So this is showing how much uncertainty there is with the future projection. At one part of it, the orange part, is internal variability. Like I talked about yesterday with El Nino, Enso and other internal variability in the climate system, but what you can see is the width of this uncertainty stays relatively constant we time, because we don't expect the variability to change, the main bit of it.

[00:26:29:06] But a large portion is also from model spread and this has to do with how the sub-grid scale processes are parametrized, like I mentioned before. And that represents a large uncertainty. But the largest one is the scenario spread, so RCP spenser red representative concentration pathway

[00:26:48:15] So we don't know how much greenhouse gases we're going to emit in the future, so we can estimate it. So say if we have business as usual and we keep on ramping up our emissions we might get up here, but if we decide to cut back on all CO2 emissions, we might get something like this. So it just depends on what path we decide to go, but there's no way we can get this one to reduce uncertainty. It will be less accurate. We don't actually know what we're going to do in the future, but this one's the largest contributor to future temperatures.

[00:27:28:24] You told us to ask about things like RCP.

[00:27:32:07] Yes, so like I mentioned it's a representative concentration pathway. So you can think about it in how much are emissions going to change in the future. So you might have a emissions scenario where you just keep ramping up greenhouse gas emissions or one where we level off and just keep it constant. That's what's contributing to spread in the future and it gets larger with time.

[00:27:59:26] So the last part of the talk I'm going to focus on is detection and attribution of climate change, so just take this down. Climate change is the change in the state of climate that can be identified. So this is through statistical tests and looking at trends.

[00:28:18:06] So this is both changes in the mean or the variability and how long it persists. It can be decades or longer. And climate variability is the variations beyond just weather events in the mean state and others the physics of climate like standard deviations or occurrence of extreme events. And the variability is both on spacial, so looking spatially and temporal scales. So detection is when you can definite-- yes?

[00:28:50:10] Is it possibly to run through that again?

[00:28:52:24] Oh, sorry.

[00:28:54:03] I don't know. Is there a different way to explain that?

[00:29:01:28] I'm just going to show it a bit more in detail in the next slide. It might help.

[00:29:06:08] All right, go ahead then. Thanks, yeah.

[00:29:08:09] So we're going to look at the change versus the variability. So say we have a set of observations. So these can be anything. They can be temperature, precipitation, or just any measure we have of the climate system.

[00:29:24:12] And we look at all these different ones, and we pick out this red one. And we can see that it has a positive trend and that this trend is statistically significant. That's the detection I was referring to but the attribution is more complicated, because for detection we only need observations.

[00:29:42:24] For our attribution we need climate models to do this, because it's establishing the most likely cause for these detected changes, and that they're consistent with the forcing. And it can't be due to variability alone. So this could be we have volcanoes, we have the 11-year solar cycle, and we have CO2 emissions.

[00:30:08:07] And we can model it. We can bottle it. And this is just showing if we put all these together in the model what we get. This one looks like it's all from CO2 emissions, the trend we're seeing, but I'll talk about more on how we can identify the human caused component.

[00:30:30:03] When seeing that graph, there was a period in which CO2 emissions went down. Oh, no, I mean the graph on the right. What is the graph on the right?

[00:30:41:08] Oh, this is just showing temperature probably. It's just showing a cartoon example, but probably based on temperature.

[00:30:49:22] So there wasn't temperature when we had global cool.

[00:30:53:12] So there was a period where it cooled a little bit. And that was due to all the aerosol pollution in the '50s and '60s because we had lots of pollution and [INAUDIBLE], because like I mentioned before the aerosol and pollution is a negative forcing where it causes cooling, because it's reflecting more radiation back out to space. So to start off with detection.

[00:31:23:14] So how can we do this? So this is showing what we've detected for climate change. So this is showing observed surface temperature changes from 1901 to 2012 and each one of these dots in here shows statistical significance at the 5% level. And you can see most of the globe-- so these white spots are ones where we don't have sufficient records to say anything about it-- but for most of the world where you do have observations, you can see there's been an increase in surface temperature.

[00:31:59:25] But there are areas, like you see this spot here, where there's a cooling, even though it's not statistically significant. But the detection. So we've detected a change in surface temperature and spatially as well. And you can also look at it in terms of mean sea level change.

[00:32:20:16] So this is from 1900 to 2010 and the same thing. The different curves here are showing from different estimates, so different data that's capturing the global sea level change. And you can also look at annual precipitation over land. And this one you see while in these two cases you can detect a change in precipitation it's much more sparse. So if you look at the speckling where the dots are, there's only some areas over the land where there's significant trends in precipitation compared to here where it's a much more robust response over the globe.

[00:33:10:16] Can I ask you a question about the map on the left?

[00:33:13:17] Yes.

[00:33:15:08] There seems to be a home in what is the Great Lakes area and the Great Plains of the United States, there seem to be a relative-- yeah, up there. Right there. Why would that are be-- I suppose it's blank not because of lack of observation, but because there wasn't any change?

[00:33:32:15] No, it's showing this color, light blue. There's no statically significant changes in surface temperatures over southeaster US shown here. This is annual temperatures.

[00:33:47:01] Where as most of the blank areas are just lack of--

[00:33:52:00] Yeah, it's hard to tell, but like this one we do have measurements, but whenever it's clear white there's a lack of observation over the period, so like here, here. But this area and that area is showing a small negative slight positive, but not significant change. But, yeah, it's hard here to see the difference between the lack of data versus what's significant.

[00:34:24:05] So how do we do the attribution of climate change? So how do we attribute the cause of what we're detecting? So first, we need observations of climate indicators. And this is what we just saw in the last slide. And then, we need an estimate of external forcing, so how are these external drivers in climate change have evolved both before and during the period of investigation.

[00:34:47:13] So before is very important, because you want to know how it happened without human influence. We want to know what the level of greenhouse gases were pre-industrial, before they start ramping up. So we need information on that. And also, you need an example for greenhouse gases and solar. And we need a quantitative, physically based understanding, so how external forcings might affect climate indicators.

[00:35:16:12] And this is where the global climate models come into play for the attribution. And it comes into here. And we also use them for the estimates of internal climate variability. So we need to just do the simulations and we'll see where we just have the paleoclimate or you just have the system running unforced just to see how many ups and downs in these indicators we can get and whether we can ever get say a 50 year, 30 years period that corresponds with what we've seen in observations.

[00:35:57:20] And before I go into an example of it, I'll just say how the attribution of climate change has evolved over the various IPCC reports. So this is the first one in '95. It says the balance of evidence suggests a discernible human influence on global climate. Then, we can go onto the AR2, where they say, there is now new and stronger evidence that most of the warming observed over last 50 years is attributable to human activity. And then, looking into this one where it's very likely. And then, the most recent one says it's extremely likely.

[00:36:40:25] So what has changed over all this? So one thing is that there we have more observations, but they've also become better where we've corrected for any biases or drifts in these observations. Satellites are very prone to drifting, so even though we put then up there in orbit, they do drift with time, so we have to correct for that as well. So we better correct it for these drifts.

[00:37:07:27] And also for stations on land as well, to make sure that they're all calibrated to each other. And the big thing also is the advancement in global climate models and how we can model climate change. So I'm going to give a quick example before time runs out of one way we can do this. And this is called the fingerprinting method.

[00:37:40:04] So the way it's set up is that each one of these forcings has a different fingerprint on climate. So the one from greenhouse gases will be different from solar variability. And so we search for these different fingerprints and observations and see how their signal has changed with time.

[00:38:09:03] Great, so this is an example. So this is showing 1979 to 2012 atmospheric temperature trends. And there's a lot going on here, but I just want to focus on these bottom two ones.

[00:38:23:02] So these are from different sattelites which have microwave lens sounders, so they've captured temperature data in the atmosphere. So it's showing it over form 90 south to 90 north and also in height. So it's going from the ground up to 100 hectopascals. So in right here around 17 or 18 kilometers. And you can see--

[00:38:47:28] Is 90 south to 90 north south pole to north pole?

[00:38:51:06] Yeah, in south pole to north pole, yes.

[00:38:53:12] OK, so this is showing what we've observed.

[00:38:55:28] Everything.

[00:38:56:18] Yes, everything, but it's zonal. It's averaged all over the globe, so you're not seeing it in longitude, because it's averaged in longitude. So this is showing temperature changes that we've observed with time. And you can see what the pattern looks like. You have a warming on the surface and then you have this Arctic amplification that's associated with sea ice changes. And you can compare this to what we see in the model world.

[00:39:27:11] So this is showing all 20 historical CMIP-5, so this is the intercomparision project I was talking about earlier. And you look at all these simulations and look at their average changes. So each one of these simulations there's multiple realizations, so each one of these groups might have done anywhere from three to maybe over ten on some their call ensembles. So the same forcings, but they change the initial conditions a bit.

[00:40:01:21] So they average over that and see this is the model response for changes. Just one thing to note, the scale on here is different, so this only goes up to 0.3. This one goes up to 0.6, because some of these models have much larger temperature changes. And what's shown up here is if we average over all these different modeling groups, so we get this model response, model change. And what's shown here is if we just perform specified coursings.

[00:40:37:06] So here on this one that's called ANT, that means anthropogenic, so if we only have human caused changes or forcings, so if we only have greenhouse gases, aerosols and ozone changes, if we only have that going on, what happens in the model world And then, this one similarity is NAT, which stands for natural, so that includes both solar and volcanic. You can see the separate volcanic forcing and also the solar forcing, because there are a few large eruptions in this time. Particularly Mt. Pinatubo in 1991. And you also have the solar ones, so you can see what happens in these model situations where you only have natural forcings.

[00:41:35:00] And the last thing that we need is these unforced pre-industrial control runs, so these are hundreds and even thousands of years of ones where we don't force the climate system at all, so there's none of these forcings, no natural, no anthropogenic forcing. And this is just showing time series of the temperature around up here in the lower stratosphere and on the right temperatures in the lower troposphere globally from all these different modeling groups. And you can see they go around and the wiggle. And you can see the magnitude comparing, so it's showing from zero to 15 Celsius how much changes you get.

[00:42:18:02] So this is important, because we need an estimate of variability and noise in the system. So how do you do this? So We take these model simulations and then, we try to find these fingerprints. And one way to find the fingerprints is this analysis called empirical orthogonal functions, so you can think of it as breaking it up at the time series or the pattern that we see into the dominant modes of variability.

[00:42:47:17] So this is trying for the ANT, or the anthropogenic signal, you can see what its fingerprint looks like. And then, you can compare that to one with old historical forcing, so that includes natural. And then, down here is showing the control.

[00:43:05:15] So if we take all those thousands of years of control run and we look at what its dominate mode of variability is for temperature. And can see you don't get anything like this where you get the warming in the lower troposphere and then cooling in the lower stratosphere. And it's just showing the different modes.

[00:43:24:06] And then, you can do the same with the natural only. And none of these patterns look anything like the ANT signal from human cause. And then, what we can do with that is that we can search for this fingerprinting and observations. I'll just quickly skip through this, but it's going to take a while to explain, but we can search for how strong this fingerprinting is in observations, and compare that to the magnitude of the fingerprint in control and the natural only simulations.

[00:44:00:24] We have and we can see how many times we can get a 20 or 30 year trend in temperatures and the noise. And then, compare that to what we see in observations for the fingerprint. And it's just showing that you can look at the signal to noise. And if the signal to noise gets above a certain predefined level, we can identify it as detected. So we've detected an anthropogenic signal in temperatures.

[00:44:32:24] And just to give an example, we've found fingerprints of human caused climate change all over in different areas of temperature, sea level rise, continental run-off, water vapor. So we're leaving our fingerprints on all these indicators. So in summary, these large radiative forcings from greenhouse gases have contributed to a positive imbalance in the climate system, but there's uncertainty associated with aerosols in clouds and how much they've offset the positive forcing.

[00:45:07:24] The feedbacks control the magnitude of the response, and it's-- like we say in that figure-- it's very complex. There's lots of uncertainty in the feedbacks and how they interact with initial forcing. Global climate models are invaluable. They're a great tool that we scientists use for investigating. And this is what Christoph mentioned earlier in the earth system model in the ITAS.

[00:45:42:13] What?

[00:45:43:13] What Christoph was referring to earlier.

[00:45:45:14] IG?

[00:45:46:19] IGAS?

[00:45:48:03] I guess so.

[00:45:50:02] IGS?

[00:45:51:23] Yeah, global systems all of the joint program uses. So I was touching on the earth system part of it. And the human caused effects on climate system can be identified through detection and attribution studies. And like yesterday I'll leave up some of the resources if you want to learn more, because I can't go through all of everything in just this one hour, and another link, as well, to the global change site here at MIT for more educational references. That's it and I'll takes some questions. I'll take a few questions.

[00:46:30:12] What are few sort of hot topic next steps in the modeling world?

[00:46:35:12] Right, so the large one, you know how I showed you that plot with the errors associated with the uncertainty on the radiative forcing? One of the largest things is understanding how clouds will respond to climate change, because it's very complex because it depends on the type of particles and also where they are in the atmosphere. And so that's the largest one that they're focusing one, so they've been building these resolving models that can actually resolve clouds. So these are ones that can go down to 20 meter resolution and that takes a lot of computational resources. But then, they can use the output from these cloud resolving models and try to use that as parametrization into the global climate models.

[00:47:30:15] So that's one area people are working on. Any more questions? Yes?

[00:47:37:29] I want to know something about meta because this is very confusing. For example, about told that in the Arctic Sea, [INAUDIBLE]. American scientists said it is only 3.2 [INAUDIBLE] in entire world because this is [INAUDIBLE]. This is a total disaster.

[00:48:09:05] Oh, so you're talking about sea level rise, no?

[00:48:11:19] No, methane, methane. How much of this methane under the sea?

[00:48:16:17] Oh, methane, methane, OK. Oh, yes. So that's one area.

[00:48:21:20] It's going to be associated more with permafrost. That's why you said Siberia or in Canada. When it warms those deposits underground and those will be released to. So that's definitely an uncertainty, but I don't think it's going to be catastrophic like you mentioned.

[00:48:43:01] But it will be really slowly over time, those methane deposits. Compared to CO2, methane is very short lived. It's only around 10 to 11 years. Yes, like I showed yesterday, CO2, the lifetime, it can stay in the atmosphere for thousands of years.

[00:49:01:08] If would be so much methane [INAUDIBLE].

[00:49:09:05] Is there any agreement there?

[00:49:11:05] What? Sorry.

[00:49:11:22] Is there any agreement on how much?

[00:49:13:07] How much methane deposits?

[00:49:15:07] Yeah, and what time it will be released?

[00:49:18:06] No, that scenario it would have to, to his question, it's being researched how much methane will be released.