 And our next speaker is Jean-François LaVar of NCAR, who will be talking about modeling the couplings across the Earth's surface in CES and the community Earth's surface. Thank you. Good morning, everyone. So, I think I did a great job at introducing some of the aspects of the current model or the system model that the grid developing in NCAR and throughout the community. So, I was chosen in the same building as Jean-François LaVar, knowing what he was looking for. So, it's something that's maybe a little bit general, but hopefully it will just generate parts that people might get back to me. Unfortunately, I won't have to run right after this meeting, but please contact me. You can find my email in the description using the different word. So, CES. So, CES stands for Community Earth's System Model. And so, the real works is that even though we have a really large room at NCAR working on this model, this model can only be made and work and work well because there's a very large community of people outside NCAR that is developing and using testing writing papers. All sorts of ways we can actually make this thing work. And what is this thing? So, a slightly different version of the one I'm going to show. So, we have basically five main boxes. One is the atmosphere, land, sea, ice, ocean. We now have the land ice representation of glaciers in the model. All has been taught to each other through a coupler. And in addition to those big boxes, we have some variations. We have some versions of the atmosphere, whether it's interactive chemistry, tropospheric aerosols, ozone. We also have the very high top, which we call the whole atmosphere, which extends right now to 150 kilometers for the people who are interested in the atmosphere. I don't think there's too many in this room. But depending on the number, we're starting to get up about space weather. So, we have some variations of the stellar ability that's influencing the atmosphere. Very important for people who have stellar lights up there. And then, so there's another thing that I'm going to talk about, the biology of chemistry. So, having the representations of biology both in the ocean and on the land. And we'll be discussing today. One of the reasons being that knowing that energy is in the bottom of the ocean, most of my talk will be about land. But it can be applicable to just your raw structure to different species. So, we have this big system that we can make work with earners being coupled, or some of them being coupled. So there's a lot of flexibility because we want to start to understand how the system works. We don't just have the most complicated system all the time. We want to be able to cut things up. So, we try to start cutting those couplings and feedbacks that were discussed before, to see if I'm part of the data. And so, this system we started running for a long time at simulations. And the type of IPCC that are being used in IPCC reports and numerous papers. This is through a number of courses. And we're starting to get fairly limited in what is being done. And so, for example, we might specify greenhouse gases, some of the longer greenhouse gases and emissions. We also have volcano eruptions. We don't do these new volcanoes in use. And I started out with the event of really changing certain things that we don't have internally done. But then what we try to do is to have everything else. Everything else being explicitly represented in the model. And we've already seen attractions. And indicated that we are most of the job at talking about and talking about new attractions. So, this is just a very brief... So, this is the way we actually get things done. We have working groups which tend to focus on specific aspects. We have a group on atmosphere model. And if you want to know how the convection is being done, then this is the group. And we'll be discussing this and where to go next. And then all those groups, such as we have weekly meetings or bi-weekly meetings now, trying to work together on moving forward into the best model. And that, again, the rest of us have some work to do with this. This will be about a degree in size. Two annual meetings. We had one in February. And then we would have one in February. And then we'd make sure we go to the mountains. And then we have about 400 people. And for three days, this is where we can... And the people can talk to each other and get the coupling. There will be a couple of mechanisms to get all those people to put together. Because this is a lot of expertise. People are interested. This is the most common version of a description paper to what is to be a community versus a model. What are the boxes? And all that's done, and what it is in this model. We're generating thousands of years of simulations. For example, this is the climate model comparison project is live. So this is usually done in support of IPCC. And so this is a model in comparison. This is where we try to see where we're at. And of course I think this figure, because it's putting a lot of models at the top. Not all of them are dead. But the model is doing pretty good. This is a representation of the climate that we have a lot of people from a lot of places and a lot of different expertise using and testing the model. This is not just a small group of people. But right now, the standard configuration, similar to what Eric Lee has been discussing, the first version of the community or system model, is the atmosphere resolution. We have a degree, one degree, two degrees. Our work quality is the one degree. There's something that when we're running the intercom care, we can get about 20 years to double. Which is pretty good. Because we're going to be doing thousands of years. We do that with a seven minute time step. 32 levels from the surface to 40 kilometers. 72 to 250 kilometers. The ocean has two standard configurations. One for the degree and one degree. Ocean anglopters talk to the international. We don't want to do anything in between. So, either way, it's very resolving or we're not. So, there's no integration that we're really using routinely, at least. And then, Brownwell is... So, I wonder if it's about primary care boxes. This is a one and a half million lines of code. So, this is a big, big, really difficult, really tricky to keep track of. There's some time when we're trying to do events to people. And you can actually run this model on your laptop. But you're not going to run it at a pre-over degree. Or you might run it four times that. But, you can actually run it at a kind of a kind of breeze resolution. Really, really coarse. And just the atmosphere, not the ocean. But it's ridiculous, though. We're running a lot. And that's something that we really try hard. It's to make it usable by a large group of people. And you'll still have to compile one and a half million lines of code. Or a virtual code. And then, we try to do two. Generate good access that people can use. Because the model is only going to be as good as what you can make under the science with it. So, that's a good idea. And it's used, really, really by hundreds of people. And that will have a few years and two, really, at the end of the year. And so, we're working really hard right now on getting this going. And inevitably, not all, such as the USM, we have to deal with. For unfortunately, limited resources. We only have a computer as big as we have. But we have to balance complexity on some of the size of the line of resolution. And I mean, I've run a little bit of time discussing the sum of size and making it. Talk about it a little bit. I will make that point. And then, this is the complexity. The amount of processes that are being included into the model. And we're going to have a completely new model kind of climate models. And from the previous IPCC report. Where are the previous models? We're going to be giving different climate simulations. And to me, it's like, wow, this is amazing that we actually get something that we look at from the previous one. Because in the meantime, we've introduced so many processes that we are occurring in the real world. We're trying to represent it. The chances of things just exploding were extremely high. And yet, we were able to get something. So it's right at the base. And we're doing it through introducing processes. We're introducing them with enough information on the observations that we actually are learning something. We are representing more of a system than less of just putting things so that it works. And so we're going to continue this course towards complexity. And then this is going to standard IPCC. I could actually make this even more complicated. But this is going to blow me up having all sorts of pieces being added. The importance of internal mobility and the necessity for ensemble is critical. And it's critical in actually designing what the model should be. This is something that was done at entire and with a lot of people from Toronto. We're doing a very long control round. So this is something that's 1850 conditions. Nothing changes. The circumstances are the same. Conditions are the same. Nothing changes. Then we'll let the model go and just compute its own thing for everything else. And so I think of those regals. They're small. So for global surface temperature, they're small. They're not zero, right? They're regals. And so it's showing up here. And then regionally, it's going to be a lot more than that. So we have, I'm actually, this is an old slide. We actually have now 2,000 years. And then we started one ensemble member in 15, 19, 20. In 1920, we just started. And then we started with about 30 different ensemble members. We're defined by 10 to the minus 14 in one level audience. 10 to the minus 14 channel. So it's a bit better, huh? It's a better plan. And then we'll let the model go. So many areas are showing up. This is ensemble mean. So all those regals, all those gray lines are different ensemble. Because this is the first cross-system, most of it will kind of look the same. But that happens when you start looking at here for the trends over the 30 years. And over time, focus temperature. And so this is not a very high, we're not looking at extremes or anything here. This is just a linear trend over the 30 years of the surface temperature. And these are different ensemble members. And I'm only putting 10. And this is how different each of those can be for the early single-all calculation. And this is the ensemble mean. So if you put all of them together, these are the observations. The ensemble mean doesn't look like the observations, and lots of blue here, and not a number of blue here. But some of the ensemble members are showing. So you can start trying to, they said, what is going on? What is this one different than this one? And we need to try to understand why the odd, which in a way is a single realization of the world. And if the model was perfect, this would be just 30 realizations of the real world. And then you can start understanding how the system works, and why it's leading to the differences that you're seeing. In the meantime, when we're trying to answer what is the trend over the 30 years of the period by model, then we can just say, we can just run the model once and say this is this number. Even if we run it and change one digit in the temperature 60 years before the calculation, we'll get this. So which one is right? Is it this one or this one? So this is extremely limiting what we can do, because this is the only way you can actually get an answer. So I'm going to spend some time just to thank you to discuss this in the context of this drinking model, the carbon cycle in the US system. So again, when beyond just this kind of climate, temperatures, and precipitation, this is trying to do an air system and looking at the biology. So what is the carbon cycle? Well, there's a little bit in the atmosphere, there's a lot in reservoirs, whether they're in the ocean, they're in the salt, all sorts of flexes to going up and down. And there's actually very, very small net flexes, very large up and down flexes, very, very limited, very small numbers. And similar to what, and it was a lot of different time scales. So you can't be able to capture what is going on on the one to 10-year time scales, one to 100-year time scales on the land. And yet it will be important to know how much is being transported to the bottom of the ocean. So which is a very, very, very important for all roads. So in a cartoonish representation of here, what I'm going to be discussing, the land model, the first thing that we can include is ecosystems, how do plants actually work? How do they get similar and updated at the future and respiration? And then lastly, there are a lot of biogeochemical cycles that are going to be involved with biology. Seem to will be one, but then plants will also be emitting all sorts of stuff. Biogenics, botanical organic compounds. Soil, nitrogen emissions, methane coming out of woodwinds. So if you want to have a big picture of what is going on, then you want to start looking at the representation of those cycles. And then you can represent the background, the low ground, all sorts of processes. And then one of the big genres is obviously water. So you have to have glaciers, water, river routing, flooding, because flooding will make wetlands. Wetlands is where you're going to have wetlands emissions. So you have to be able to take all those aspects and then water ends up in the ocean. So you want to close the budget and whatever comes down from precipitation has to reach the ocean. And it's a critical part of the work that we do is very simply to just ensure conservation. Very rarely is it extremely simple that it takes hours and hours of people's time to make sure that we actually keep track of all the sources. And then it's all done under the pressure of human systems and human prohibitions. So instead of having natural vegetation, which we might be able to represent, we have to take into account the fact that crops will be used, there will be irrigation, there will be harvesting, we will make seeds which will change the albedo and the river. And so the land by where it does that is by really having a grid cell, which tends to be, in what we use for climate simulations, a one-degree grid. So that will always be one-degree grid. And then we pilot. We just try, okay, so we're over an hour, there will be a little bit of this, a little bit of this, a little bit of this, none of that, and a little bit of this. And then you start breaking down. Because then you can start having representation of all the processes in all the interactions that are going to be important to drive something that ends up being piled by groups. So we do nothing like the way that the real world is organized. But we move them and then this is then averaged. And this is what the information is given to the atmosphere. So the atmosphere will see the average sensible heat flux see what you want to represent. And so as I said, there's been a lot of time looking at everything that is being done. But these are all the processes that are now being present in the current version of CLM. A lot of work has been done on a really good representation of hydrology. Hydrology has had been a really major factor in having a good representation of biology. And so you didn't have to write the way CLP was being taken because you didn't have something as simple. It's a good representation of the soil, the distribution of soil, and where the water is. A lot of work has been done trying to get members of a couple of people with really hydrologists to have a really good hydrology model to be included in this. And this is one of the aspects of this work is that we're reaching out to more and more of the communities outside the standard climate workers to get all the information that we can. And then one of this is also there's this little piece here which actually ends up being a very large driver of the system. One of the biggest responses is the nitrogen deposition. There's more than C2 than carbon that the plants care about. They have all sorts of nutrients that they need. Nitrogen is going to be the first in order to have enough of carbon. Nitrogen. And then in nitrogen, you have to have a spectrum stream. You have to have emissions. And we can start doing those simulations in trying to have the representation over the Northeast U.S. and the pollution and what's being deposited. And then there is now nutrients that the plants can use. So we're going to be discussing the exchange of CO2 because I'm going to give you that. The other aspect of the resolution is this is a one-degree model. This is a 10-degree model. If you think of this trend of the world as being seen through this amount of kinetic energy or just the structures in this path, it's hard to know if the mean of this really leads to mean of this. So there is a need for high resolution. We have to push, we have to understand how much we are missing by using this model or our projections. And we're doing the same thing on the MSL. One of the approaches to trying to be smart once in a while at least is to go with the regular time of the regions of interest. Because in reality you might be interested in looking at things everywhere. But what are we doing here? And so if you're trying to do a little tropical cyclones, there is no tropical cyclone that just don't exist. There's nothing. But if you use those, we'll define them to 25 kilometers and then in the area where you care about and again it's where there was a point of energy to communicate, depending on the question. If your question is about tropical cyclones, are you a bit up with something with high resolution? Because then you start seeing tracks that are reasonably close in amplitude in the direction frequency to the real world. But then this is helping you but if you look this and this there's very little impact of having this high resolution over this region. So it's a limited means but this is one way in a limited amount of resources to go and start answering some of the questions that we do with all that. Well, C is the one first of all, it looks like the line is first. This is called an achievement. CSN is getting the rate of increased at normal level compared to the observation. Yet, this is a huge change in balance. Nothing that we can actually justify. This is not the price. And so we'll have the process of trying to get another hydrology, better presentation of all aspects of plants. And then this is where we started. Over the 20th century vegetation was actually a source of land. It was actually a source of the sea. So explaining at least the divergence between the two. Well, actually, the best estimate was CSN. And so now the improvements to biogeochemistry and the physics and the biophysics of the land model. Now we're getting much closer to the present day. So what I've been talking about is mostly this, where you're switching and then you have also to feedback temperature and then you can have some more information. After that it's warming. But really what we're after is something more like this. These are all the land biogeochemical feedback that are actually all in the calculus to understand the kind of system. Especially if you're interested in the regional scale. The middle scale, maybe you can just assume. And you get the complete temperature. But this is what we're after. And one of my projects is more on the ozone deposition. Because ozone, as you know since the 1950s, is actually really impacting the plant's ability to survive. And the high ozone in the plants is really effective. And so overall, in the present day, this is actually a non-ignigible part of the gross primary productivity. And this is the impunity for ozone. If you're really pushing the model and trying to reproduce some of the innovations and use all the structures, you have to start including all those processes that I've tried to highlight. So in this grant tour of CESN that I tried to show, when I looked at CESN, it was a very personal tool to explore the reactions feedbacks in the whole Earth system. And we're really showing to improve our process representation which is where interacting with people in the field and connecting with very, very different disciplines and expertise is where we're making progress. And the colleges we're working with people who grow and look at the fisheries, we have to talk to the people, we have to discuss very strongly with the hydrologists, to really understand how do we get there at the moment. And so I think this is a group that would be very well that is a very good place to have this discussion because we have to work with very different disciplines but in the end we can be able to put it all back in the kind of model we can use for our comments and reactions. Thank you. I want to talk about the audience because Jean Francois and I were talking earlier about budgets just so that we're clear we get about one million base funding and maybe three million in secondary funding so for a total of four million I'm not quite sure because it's heard in a tally up like CES and others but it may be more in the order of one million to ten million base funding and maybe 20 million in secondary funding. So the two communities are actively working to create standard models that can be shared by the communities both are very open source and this is probably the first time that CES has reached out in a proactive way to CESM to start working together and Jean Francois offered us to have the say in his sort of modeling world and I think we should take advantage of that and maybe use some of the results that they're getting so I think there's a lot of honestly things we're looking at really getting these communities work we haven't maybe reaching out to the other community that we work with CIG cyber cyber cyber infrastructure which deals for a long time right but you've probably opened it to questions because there's so many that I have I'm going to steer but I'm going to use you know I'm going to steer so I've attended many meetings between the world climate research program and since I sit on that board and the world water weather research program and I've heard these two communities and most of the people in the climate community don't even know the world whether research program exists and it's larger than theirs so that's they didn't know our community existed either so even in the climate world there's not a lot across the world but in the in the weather community you know their goal is to get to three months predictions and they think this is almost impossible the way they do their modeling but that would be their goal and at this one meeting the world climate research program said well you know our physics and our equations are based on our approach is to that would be a very tough thing for us to do to make a three-month prediction so I want to turn this over because we've had people at our meetings that have passed a doubt about the water bottles the weather research forecasting approach so maybe you could then maybe one minute on talking about climate versus weather in terms of the climate because the climate is not the weather so the answer is this is something that's been in the world over the last few years because the realization that there were very different communities and we're actually trying to test our climate models in the framework of our chemistry and usually well of course it looks terrible the head of the book is because a lot of the work that we have done in a fairly course scale but also because we've never really tried and so we might be missing some processes that we think we don't think are equivalent on the time scales that we're interested in but actually so actually the beautiful the beautiful tracks from five kilometers when they show that to the work people they're like if you get psychomes at 25 kilometers for the longest so again we actually need and right now we're building the system in such a way that we can bring the world physics into sea and the tropical cyclone projections are the way to do routinely with the weather I don't try to see what works what does and where does it break because in the end you want to have you want to make sure that what you're using is actually the representation of the physics and not just some evolution from where we started in the 1970s where we're really limited but the overall approach is as we hydrologist you sort of tried to work with CLM before and these components and parts from CLM with my own work one of the issues we've run into is that 1.5 million lines divided by the seven odd components he showed is still a really large number and so it really comes down to that each of these components are also really big monolithic frames or desires that move towards a slightly more flexible or sort of moving away from the one model to the other model and do you point with them to those pieces? Yes that's a very good point and we've heard that more than one it is the main factor some people just fight to build it and spend time to really get into it but it is a big barrier especially for students who you can't just say oh let's start looking at the code and see where your work coming from without working between modularizing the land model is actually one of them the most progress at this point is another one where for example the biogeochemistry is going to be a separate module that can exist outside the ocean model so if you're interested you can just go to it so we're working in that and then at the same time the limitation is resources because we're trying to get new releases and new things so we have just the getting dead weight of bird waste so we're moving more and more in that direction and I think the more we hear how limiting it is to actually in the end improve our model the more we can make it happen but it is in the plans and this CESM 2 we'll have moved about this far in that direction doesn't it really one of the challenges is that we face at the coastal land margin in evaluating climate change and sea level rise is how the flood plain is going to change into the future and it's really interesting the work that you're doing being able to project the tropical cyclones in the tracks and get more information on how they might intensify and what kind of regularity we're going to see into the future but do you think that what you're producing ought to be used in a design sense that it ought to be used in informing future flood plains or is it just more of a general information on how they might intensify what's your opinion on using them directly and I would probably be my overly cautious nature to be on the letter and this is there is information and this we're also trying to get statistics with different models at the county university single model and it's really with us that we can actually say that this is just a single model and it's completely dependent on our ability to present the process it really depends on the question that you're asking I had a workshop last summer with people doing this from Florida coming in a huge issue like underground water works with respect to sea water and so it looks that there is no way that whatever we can produce would be at the future for me but even in that case having the general question is it going to get drier is it going to get wet was sufficient for them to actually start thinking about what could be the issues that will be facing so ideally you would want to see things as a higher resolution with more representation of the details but in the case it was extremely good at showing this is just not insured it would be better when you start making those plots you get all those details whether it's actually working better for you is completely open and questions we don't really know it might give you the false impression that it is actually working my view has been more let's use the limited resources that we have and go with the low resolution but then with a lot more statistics that you can actually get and so again in the question you're asking it would really be more like what is it like we heard of something that would be affecting the region more than something that's specific and I think that's what we're going to be for quite a while when you go the more noise you get into your system and that's where on stumbles we can even more so you kind of start you have more resolution the more computing you have to do and the more on stumbles you have to do so that's going to be a little bit more