 Good morning everyone. So we have the first 10 plan areas this morning and they're a little bit focused on human dimensions. And the first one is by Mark Ransomal at the University of Edinburgh and he is our new co-chair of the human dimensions focus to research groups. So with that, I invite Mark to come up. Thank you. Good morning everyone. So I'm going to explain in a moment what the title actually means. I'm going to start with the immortal poet from Monty Python and now for something completely different. So I've been listening very carefully to the talks that I've heard so far here. What you're going to hear from me is really quite different from what you've heard from the other people so far because I'm going to be talking about people, people and society, the societies within which people operate. We have a whole set of ways of thinking about how we understand how people contribute to their system and I'm going to give you some examples of how we try to model people and societies. The title said integrated modeling. So again, I'm going to give you some examples of this but it means different things. It means integration across natural and human finds disciplines and that's absolutely crucial. So we're really dealing with a multidisciplinary perspective. It's integration across areas of human endeavor, different social economic sectors which again we have to account for when looking at human systems. It's also really crucial for understanding people and its integration across spatial scales. I'm going to give you some examples from the global right down to the local scale of human dimensions modeling. There's also integration across multiple global change drivers. So the theme of this meeting is climate change which of course is an incredibly important topic but we can only think of climate change within the context of a wide range of different social economic change drivers and that's crucial in terms of understanding the human dimensions aspects. So I'm going to give you examples today which are really drawn mostly from the land and water resource sectors mostly looking at the topic of land use and land cover change because one of the areas where this integration across the biophysical and human dimension sciences has progressed the most in my opinion. Okay, so the human dimensions of the earth system. We know that human actions impact on the earth system and changes in the earth system. That's what climate change is all about. Climate change is about people. We are the cause of climate change and we may or may not find solutions to the problem of climate change. So we impact on the earth system in many different ways but we're also highly responsive as individuals and societies to changes in the earth system itself. An important component of human dimensions modelling is trying to understand those responses how the individuals and societies respond to a changing balance around them. So we have this sort of dual set of impacts between human systems and the broader earth system. So the question is can we model human principles? I'm often asked or given statements by my colleagues who are physicists engaged in modelling physical components of the earth system. They say, how can you model people? They're sort of random beings that just make things up as they go along. There's no basic laws or theories or anything like this. How can you model people? But we can model people. It may be that we don't have a particularly well-developed understanding of the processes or the underlying theories. So when we think that people are behaving randomly, that probably just means that we don't understand really what underpins their decision making. So it's important to understand that human systems and interactions with a broader environment are reflected as a complex system, like in many other complex systems that all of you are working on. And that we can build models to try and understand more about how human systems operate and their role in the earth system. And there's no gentleman who always thinks that we've been working on here. So the climate system we know is a complex system. We know that we can build models to do experiments to understand how the climate system works, and we can have a fairly good idea of the processes within the climate system. That doesn't mean we can predict a typical cycle at any given moment in time, but we can understand a system quite well in the sun. Pectit Thomas, another example, we can understand physics a bit, we can understand the theories, we can build models, but it doesn't mean we can predict earthquakes. And it's the same with human dimensions. We can try and build models and understand how human dimension systems operate and function, but we can never predict Donald Trump. Okay, let's start from the global scale. I'm just going to give you some global right down to local. I'll give you some examples of what's going on. One of the things you need to understand about human dimensions modeling is that there are many different paradigms, many different theories, and many different methods that are used in modeling different human systems. You're probably familiar that there's a lot of work on modeling done by economists. Well, economists have a very fixed view of the world. There's a lot of modeling that goes on in terms of physiologists, for example, psychologists. They have quite a different vision of how human dimensions operate. So let's keep that in mind. There's no single set of models. If you did a review of a human dimensions model, I suspect there would be someone between 10 and 20 different, completely different paradigms, how you go about undertaking modeling. So I just want to give you some examples of some of our studies of these different scale models. And scale is important because what you can represent in terms of human processes is really dependent upon the scale of which you're undertaking a study. So that's crucial to bear in mind. Right, so we're far away from modeling. We're planning the process on this land use model. This is a global scale model, or a global agricultural system. I'll have to show you a picture with arrows and boxes behind the words, the relationships that are modeled here. And again, you can see some of this relates to the physical environment, how crops respond to climate change, for example, in terms of yields, but also very much the social side. You know, how do people, or how much food do people consume or their preferences to do that? We're doing that. So we're really looking at the supply side and the demand side. This is a crazy economic approach. So we're modeling a global agricultural management system. And the types of things we generate are maps like this, very straightforward. So this is for the IPCC, S-Rest scenarios. So we're looking at crop flow and it's modeled in the year 2000 from the baseline projected forward to 2050 for the H2 emissions scenario. And you can probably see if you look carefully, that the odd differences between the top map and the bottom map, that's good. So we can model some of the changes that come about due to a whole range of scenario assumptions about how the climate will evolve but also how socio-economic development trajectories will evolve as well, both important in contributing to the differences between those maps. Again, one of the reasons we try and do this in a spatial context is to understand the hotspots of change. So which parts of the world are under increasing pressures due to land use change under this set of scenario assumptions. The most important thing about doing modeling is to confront our models with observational information. And from that we can learn something. So as you can see from this graphic, this just shows a period of 1990 to 2010. This is the serial land area, the thousand hectares for different regions of the world. And what you can see is that the little circles, the circles are the model simulation outcomes and the squares are observational data collected from the United Nations Food and Agriculture Organization, the FAO. And as you can see, that model is fantastic, isn't it? Isn't it really good? It's only a count of years, I'll give you that, but it's really good. Well, that's not bad, actually, given the complexity and difficulties in the model and human system that people have experienced. Let's point out where the difficulties are. You can see this one here doesn't quite work. There always has to be a difficult one, doesn't there? Somewhere along the line. And unfortunately, it's Russia. Russia's being a bit difficult in this particular situation. Now, what's happening actually is really interesting because if you look at the time horizon here, we were unable to model Russian serial land evolution over that period of time. You can see it's significant that this graphic starts in the year 1990s. So this simply reflects geopolitical changes that happened with the collapse of communism and the Soviet Union over that period of time. And we don't have that as a process in our models. So actually, this is a good result. It would be a shame, actually, if we were able to replicate what's happening in Russia because we're not measuring those processes. So this is just a little warning that, you know, when we think we've got some handle on how these systems are operating, there are always processes which we don't know are going to have an influence. And the crucial thing there is that even the CIA couldn't predict the collapse of the Soviet Union. So how are we going to do that as models? We need to understand something better about the evolution of government structures, geopolitical structures, and how they reach tipping points, such as these geopolitical collapses. And that's something which is industry-difficult to model and practice. Well, this is a lovely table because you won't be able to read any of this. But I can say with a watch. So this is, I'm just showing you one model, which happens to be one of my models. And it's great, as you've seen. But of course, we want to be reflective of how we look at how we model these types of systems across a wide range of different models. So what we're doing in this case, this is a study where we did a land use change for global land use change model into the power system exercise, just to try and understand what's going on out there. This is just used for listed models. Actually, this is a listed model that operates at the global scale and also at the continental scale called the consulate of Europe. Within this list, there are about a dozen global scale models that will model land use and land cover change as an outcome. And most of those are what are called integrated assessment models. These are models that are used strongly by the climate change community in understanding the energy economy feedbacks to the climate system. So they're quite important models. They're also going to land use land cover change so that you probably realize land use land cover change is incredibly important to the impact on the climate system and the climate system. Something like a third of the observed climate change today compared to the industrial. So it's really important to be aware of the climate change. So there's a whole range of different models. So what we did is we took on these models into a wide range of different scenarios. Some of them would be the old IPCC as well as some of those. But we've also got here a lot of new ones today. AXSSP and RCP based scenarios. If you're familiar with those, some of you are, some of you aren't. And we try to compare. We try to understand the purpose of this. It's not to say, oh, this is a great model. This is a bad model. This is the interception that's afterwards. It's not really to say, this is good. This is bad. It's actually to understand why there are differences between them and see where we are so we can learn to take things towards the next step. So this is the type of data you can obtain. I'm going to take you through this very quickly. We have a data that opens very full class of land use land cover. We have problems with pasture and forest. We put it in top graphics now. This shows the area of each of those particular classes. The mass of classes. And we're going two times in the future. So we're running from a multitude of scenarios. Now, if you look at the first one, it's a problem. And you're getting exactly what you would expect. So these curves represent different individual models, which are listed in the legend at the bottom. Hopefully, you can't see that. Otherwise, you'd be trying to see which is good and which is bad. If you just ignore that, it's just data for this purpose. And it's going to get different scenarios. So what you would expect is exactly what you see in a problem, which is this tone of diversity as you go forward, because there are different assumptions about how the world is evolving, including if that's a climate change. So that's great. Well, you can see that in the event of caution, we're looking at quite scooters, obviously. So we can not quite right. And actually, well, by the way, the parallelized just showed the observational data from the data from the data I showed you earlier. So that's just a reference point. And there's a data from a different model as well, so we're not reacting. So what you can see here, oh, my gosh. So what we've got here is the initial starting positions for these curves. It's completely different. And actually, what you can see is there's a smaller range of values at the end of the simulation period than there is at the beginning. Oh, that's not good. So what's going on here is that there are two main things. First of all, it's the initial data that people use as inputs for models. They have different starting points. And there's no accepted global data set but past year, round shoes. The second thing is all of these modeling groups define past year differently, like different definitions. What is past year? When you've got a nice, intensive last one, that might be past year, but when you start putting a few trees in it, the energy range runs in a few more trees. But which point is it grossing, and which point is it porous? So it has no problems in thematic definition, which really constrains what we can do. And you can see in this that this is an enormous source of uncertainty. That's the same for porous, right? Again, look, there's no divergence. You always see divergence across those range of values. That's not happening. The bottom curve is to show when we've re-based, we've removed that initial condition. We've just re-based everything to the observational base by the near 2010. So again, you do know a little bit more about that too. But we have to keep in mind that when we're using any of these model outputs, those initial conditions are really important. And we're trying to quantify this. What we're taking are these raw data. We've done our decomposition of their own statistical study to try and break this down into whatever factors that explain that availability that we're seeing. We're going to have to try and show you this. We'll take a little bit of a faster course again. So what we can see here is the initial conditions in blue were what I was showing. So those are really important to start with. So this is a fraction of what I'm going to explain about these different variables. But then we've also got, in real here, the model type. The model type is the paradigm that underpins, of course, if it underpins the model, as well as the resolution, the number of cells which are in a grid, which is used for all these models for simulation. And then we have the sort of ready colors here which are the social economics values. And finally, we have a representation of the climate input through the CO2 concentration, which is really important for these models because it affects our tropical activity. So what we can see is if we start looking at problems begin with, the green band, the model type, is afraid of being important. The social economics is quite important. The climate change is not very important at all. And let me move to pasture and forest. It's even the last situation, the initial conditions we saw from the original grass were really crucial. The scenarios are really not much more important than the model type itself. So what we've got when we compare these models is that the differences between the models is probably at least as large as the differences between the assumptions in the scenarios. And that makes it very, very difficult to make any sensible, robust judgment about future projections, lunges, and that kind of change, really quite significant as an outcome of the study. And just what I like to call you, again, this is cropped in the areas through times of future. You don't need to understand what's going on. But what you need to know from this is there are blue curves and there are orange curves. You can see that, can't you? There are two curves, but the blue curves are what are called partial equilibrium models. The orange curves are called the general computable equilibrium models. They are both two styles of economics models, right? It's keeping them in the economic community. These are just two ways in which economists model global manager use change. And you can see there's a little bit of a stratification across those two models. So the model side is incredibly important, even from a range of scenarios, in determining this. So really crucial for those of you who are involved in climate change studies, this is really quite significant because you look at the IPCC and the way it reports so managers can't ever change from these assessment models. It's reporting the results from a single model for a single SSP future scenario, right? Which is quite an amount of months and 20 click because all you're showing there is the difference between the models, not the difference between the SSP scenarios. Okay, we'll pull it. Okay, so two key methods used in global studies, great divergence in human processes, and the implication is that we're not actually doing very well at modeling global non-juice learning climate change, the human dimensions at this moment in time. That's all we've come down in scale. And that's what's coming into the European scale now, and just give you some other ideas about how we move. So now we're looking really at what I'm calling regional interracial assessment models opposed to global interracial assessment models. Similar types of paradigms, whereas much greater possibilities to include more refined data status inputs and a wider range of processes that can be found with trials at that level. What's important actually when you get down to these levels is the interaction between different sectors, right? We know that water resources are highly dependent on what's happening with the land use change or with agriculture. We know that biodiversity depends explicitly on the growth of available to water resources, on agricultural management, and the whole range of things. And in many, many cases, this is only the level. It's not least in climate change studies. And I'll give you some quite annoying results that demonstrate some of the problems at this in a moment. So we need to do that. So in this particular study, this is another of the models I've been involved in Did it get further? Oh, well, that's interesting. Well, I think it was helpful. Oh, here we go. We've just got stuck. So this is a model where what we're trying to do here builds individual sectoral models, you know, hydrology models, agricultural models, agriculture models, flooding models, et cetera. They're based on matter modeling. So they're the youth-conversions of the original total process model. And the reason we do that is to facilitate the coupling between these different sectors. And we want to be able to do really rapid run times of these models to explore the scenario of space and the relationships between those different components. So that's more or less what the model looks like. And what we're doing is we're trying to model for the whole of the European continental region. It's actually the U27 plus 2, right? And that's the sort of thing we could do. By the way, this model, just a little slide step from an educational perspective, I think it's a little bit more process-free. This is the website. This is available online. You can go and run this model through user-friendly interface. There's a whole range of sliders and checkboxes that you can implement in those scenarios and run these models yourself. You may not, most of you in this room, be interested in Europe, but just have a play in a way if you want to, it's quite a bit fun. You get some interesting results there. As is the impact on biodiversity and threats, winners and losers across Europe due to climate change and other social economic changes. Just an example. But what I really want to show you is this. And this is quite difficult to understand. So let me just, just listen to me and I'll go back to them. I'll tell you really what I'm trying to illustrate here with some of these data. What we did in the study was try to explore how important is it to consider the indirect effects of things like climate change and other social economic change drivers. How important is it? As I said earlier, we know that agriculture, we walk around packs of water resources and vice versa and biodiversity. So how important are those indirect effects compared with the direct effects of, say, climate change or changes in the economy? And that's the study we wanted to do. So what we did with our integrated model is we round each of the single sets of models as sound and low models. Before that, we referenced that against the report, the pre-assessment report of the IPCC, to ensure that our four sets of models are consistent with benchmark against the literature sources that are out there. We'll see about that. Take it from the Vibra. Very good, very good match. So we're sort of running our models as sound and low models for the same scenarios and as a coupled model system, we're looking at the indirect effects. And this is some of the results we got. What this shows is the, just in simple terms, whether across the whole map of Europe, to all the pixels of the modeling, how many of those get the same values in the single set of models we're in the space of model? And then you can see different types of indicators that we're outputting here, but there are different levels of concurrence or agreement between a single set or an individual set. If you look at the full provision of a random full provision, across a range of different volition scenarios, a range of different GCNs we used as well, you can see that actually there's only about 20% of this area of Europe actually has similar values between the two modeling components. That changes as you go to different types of indicators. You have to keep in mind what the indicators are going to be really important. But some of the changes are really significant. And this is a little future to some point. I haven't got so much time to explain this, but just look at the arrows, right, look at that, so it's 50% change, right, that's what you need to know. And you can see that a lot of these ups and downs arrows, again, this is comparing across the minimum maximum values, changing the range, et cetera, across these different variables, how different the single set can be integrated into the world. And the result is they are very, very different. They are completely different. What's significant about that is that if you look in the IPCC working with two assessment reports on impacts and adaptation, 99% of the reported studies would be based, would use the models, will be based on single set to models. It'll be based on hydrology models, agriculture models, et cetera. And what we find from this study is those assessments will either go to very strongly over or very strongly under estimate the impacts of climate change. And that's quite a growing issue. Jay, do you spot that? How many things did you have up there? No. All right, all right, I'll keep going. All right. Okay, so let me tell you again, we're now down to the national level, and I'm just gonna, just may work on it. Just gonna click on this, and just quick explain it to you. When we get down to the, is there another two person in the house? Oh, there's only this one. Oh, there's just another. Put the glasses on, you can see. Okay, once we get down to the national level, we can use completely different parameters of how we model land use, land climate change, in this example. So what you've seen here, this is some, this is great, but small island off the coast of Europe, where we don't know where it is. And what we're seeing here, this is the simulation through time into the future of the, well, the uptake by farmers of two bioenergy crops. So this is actually climate mitigation. So I think that's climate change. So it's really important for climate mitigation. So let me start with some short rotation. What we want to see, the red dots are the core evolution of production plants, where you burn the stuff to produce energy. The two of us go together. This is based on understanding of the individual processes that are underpinned decisions by individual farmers about whether they uptake this crop or not. The UK government, since part of its UK climate mitigation strategy is providing our subsidies to farmers to encourage the integrated bioenergy crops. So they've got big interest in doing this. What you might have seen if you were watching the video here, just a video of an outcome from the model one, is you get this action, what is a spatial diffusion process that occurs within the simulations. And this one is really useful in helping us understand how farmers are uptaking knowledge about these novel crops to them, this is novel, and implementing them or not. And what we find is, if you look at the black curve in this case, this black curve is the uptake by farmers of these particular crops, assimilated. And what you can see is, we get something like this, it was 20 years before we get full uptake. That's really significant. We know about the exchange of knowledge between individual farmers, which influences their risk profile and how they think about risk of a novel crop. It's literally a physical process. You see this through the landscape. You don't have to believe my model, but this is actually the red curve of the information data for a crop which you can call oil seed rate. But the people call this oil seed rate, they don't know what oil seed rate is here. Yellow flowers, smooth food, oil seed. What is it? Hello, I think that's the one that I told you about, she's right, canola. So then you have observations from the day between the 1980s, they were subsequently introduced, farmers took this crop up. And actually what you see, I want to re-face this to give you an idea. These two different crops, but what you see is the same pattern of uptake, it takes, look at the red curve, something like 20 years before you actually have a uptake of this particular crop. The UK government, I'm just a bit of a disaster, really, because it's always a lot larger than money, and we'll have lots of climate mitigation through bioenergy. But really, it takes 20 years before you really get that. So you're completely overestimating your capacities and this big climate change with your particular individual strategy. So understanding the social interactions between individual people is really important in these types of lessons. I want to give you a quick note. I'm going to switch right up to Sweden now. And we're going to look at another agent-based model. And again, I'm just going to make some points here. This is an agent-based model which has been developed for the forest sector in Sweden. Trees are quite fortunate in Sweden, if you've got a finger. I'll do a little bit about the culture in the South. And what we've been doing is working with social scientists, civilization doing social surveys so we can parameterize agent-based models and decision-making of that. Which type of agents are located where and what types of trees do they grow and what purpose are they trying to generate as outcomes? And this is what we can model through time. Well, I just wanted to show you this because we don't have much time. So again, this shows assimilation outcomes. So we're looking at, in this case, the ecosystem services which are provided between the different land uses by different agents. Just to give you a heads up on this one, you see there's a bit of a peak. This is pine timber production. This is going forward in time. I'll give you a little bit of the word, isn't it? This is a peak. This is a peak. This is really quite easy to understand. This is an emergency effect of planting of forests in Sweden just over 60 years previously. They're just all concentrated at the same time. So sometimes you can actually understand a lot about the future just by not knowing what's going on in the past. So what's an important point? I want to show you this one because you will never understand what's going on here. And I want you to feel and enjoy the randomness, apparent randomness of what's going on in these curves. And what this is showing, we've got this index of what we call coping ability for individual agents. We have agents who are productionist agents who are recreationalists that have different objectives, different attitudes, different ways of interacting with one another. They can collaborate, they can compete, they can do more sorts of things. But what these graphs show is through time the relative coping ability of those individual agents within the system changes. Right? And this is really significant because we're thinking here in terms of how people adapt to climate change, that's really a different question that we would pose with this particular study. And the issue here is you can see that coping ability changes all the time. There's this sense that adaptive capacity of individuals is some sort of inherent trait that we have. You can do it, yeah, but it's money, and quite bright, so I can adapt. The problem, what this is showing is that it doesn't, the ability to adapt to a changing environment is not just a function of your innate ability and the changing drivers as well. It's also your relationships with other individuals, other actors in the system. And that means that coping capacity is contextual on your relationship with others within the society. So the various moments in time, you are better or less able to compare, to compete relative to other agents within the system simply because of those interaction processes of collaboration and competition. So it's not a static process. Adaptation is dynamic. And you need to understand the human dimension processes to understand how that's going to unfold. Right, race towards the next generation, and then save the slide at the end, which is a product of systems. So I know that I'll get an extra minute or so to explain that one. So I can use that for time. And race towards, race towards, we need to match across those different scales. We know at the global level, we've got pretty poor representations of global human processes. We don't really do a good job, as we can see from those comparisons. We know at the local level, we've got lots of interesting knowledge, information, theory, processes represented, but it's a complete mismatch across those scales. So we need ways of better developing understanding of how we can use the theorizing process representation of those local levels based on individuals and knowledge of others individuals operating within society to offer that to larger geographic extents. That's one of the ambitions we have within the human dimensions effort. Gee, in systems, if anyone's interested in this, please engage with us. It's really just trying to improve theory, knowledge, process, and being able to upscale the global levels because we know from a political and as a scientific perspective, it's really important to understand global scale changes in human systems. And this is just one way we've been thinking about doing it. You're familiar with the notion of the function of types used in global vegetation models as a way of simplifying and upscaling from sand-based models up to global scale models. Well, we're working on the notion of agent-function types where we can begin to categorize individuals according to their cognitive processes, the ways in which they use capital space and the ways in which they interact with society. And that's for the basis, it's really clear that's going to be the moment since people in each time have changed the spine here, but that's the way forward what we're trying to develop this way of thinking. And, well, that's right, that's what we're going to be doing here. So next week, some of you here will be involved in this, which is great. What we're going to be doing is hosting a meeting here in this building. That's really good. We're going to be hosting a social system model. And the whole point about this is to tackle this mismatch between how we understand human processes across scales and preceptors and think of the next generation of human dimension models that can be implemented to go beyond the capacity to model human-evil interruptions at present. And I will just leave you with that quote. Thank you very much. Yeah, I'm a communicator personally. So again, it's kind of a mismatch across scales. So if you work at a local level, you can work with collecting social information from the individual people, social surveys. You can test those processes and practice empirically because there are models to do experiments. And actually, of course, the other tricky thing with working with human dimensions is we can't do experiments with people other than the models. You can't get people to put them on the line and somewhere and see how they interact. You have to be able to do that. Let's do that. I'm very grateful for that. Unfortunately, the Ethics Committee in my university wouldn't be too pleased if I decided to do that. So we do need the models to start doing experiments again against data. And that's what really was down at the global level because you just cannot do that. Apart from anything, you don't have global scale data sets that would give you that information. And let me say long-term ambition upon which I would love to see is if we could actually initiate a global scale repository of social sciences information based on hundreds of thousands of millions, I don't know, case studies that people do in anthropology, sociology, all over the world. There's a thing that can damage you with a vegetation moment in a community that's building trait databases to try and parameterize functional types. But why don't you bring together case studies in human dimensions where we can parameterize agent functional types and use that to build. So, yeah. It should be one of the top ten in the world. So we'd like to expect a lot from you at the moment. Because the other thing is, well, by all the time it's just a lot of stuff. Right. And it's only rich in civil science. Civil science. And so it's only once in a moment we're tough to reach out to a program which included these individuals. Sorry, which individuals? Right. Which individuals? Which individuals, which individuals? The last person. Right. They are, I could say, part of the board community. Actually, I don't remember the course by that. They have got a certain role to play. And so I think one of the things I was making here is that actually we do need that to realize it about how human systems operate. We just don't, we have theories. They're mostly economic theories. And we know that they're all wrong, right? So we need to do better. And actually, a more philosophical approach, often maybe more psychologically-based approach as well, is probably really important here. And our people engage in this type of work across communities. But I've said there isn't really any community that it depends together across these different disciplines that can actually take the work forward at the moment. I think that's why, frankly, human dimensions and our sort of modeling is where you are finding many assets of the physics of the Earth system. And I think it's mostly because a lot of philosophers, if you tell a philosopher you're a modeler, they think you're the antichrist. They're not accepted, modeling is a way forward. So we have to work on it for that. I mean, there's lots of challenges in doing that. But the other thing I just want to take up in the time here, the other thing I think is really useful, I think loads of questions, is that the other thing is really interesting is how you use these models, how you compose these models with real people. And sometimes it's fun to come up with models of people with real people, that's quite fun. We're doing a lot of participatory studies in that team, where we are running models up in real time, and so you can run over two and a half times with decision makers. In my case, we've done studies, workshops in Scotland, with Scottish government officials, to understand their responses to some of the model changes that we're observing. And it's the basis of a discourse to analyze what's going on. We're not saying that a model is the best prediction of how the future world will be in Scotland, or elsewhere, but what we're saying is these are plausible future projections that come out of these models, that we can form the basis of a discussion. An example of that is the Scottish government is currently interested in re-forested in the whole of the surface area of Scotland by 25% of the surface area, part of its climate mitigation targets. We're in a great workshop with Scottish government officials, so when you can do that, there's the number of lines that show them that, so 25% of Scotland is now covering trees, that's just what you want. Oh, by the way, it's got to mention that you have to increase your food imports by 10%, and everyone has to move over to a vegan diet, right? That's the only way you can achieve that, because there are trade-offs in lunges. They look on trees and can't grow animals. So I'm like, oh my gosh, wow, that's unusual, what happened there? They'll achieve exactly, and it's not that the 10% of food imports is important as a predicting value. It's just saying, actually, you will need to increase imports. Yeah, I'll shut up. Thank you. Thank you.