 OK, so thanks, Ellen. Thanks, everybody. I'm Paolo Rogeri, third year PhD student from the University of Laguil. I'll try to discuss a bit how teleconnections between the Arctic and the Middle Attitude work. So I would like to, so we had a nice, regular teleconnections after yesterday I think. But I'd like to point out how teleconnections between the Arctic and the Middle Attitude are summarizing this recent review, Overland Attitude. So they say that remote forcing, that is changes outside the Middle Attitude, remote in space and perhaps in times, can influence the Middle Attitude circulation from linear and nonlinear atmospheric patterns known as teleconnections. So what I'll try to discuss in this talk is basically how I will focus on ice variability, but how high latitude forcing can interact with the Middle Attitude atmosphere. I will focus mostly on autumn, winter, and the north of planting as far and scarer sea ice. And I will try to emphasize a bit also the role of the stratosphere. So I'll give a very short introduction, then I will show an experiment from a speedy, as a result from a speedy experiment. And then I'll talk a bit, perhaps, about the simple setup in speed and simple climate experiment that we recently did. So one reason why there's a lot of interest on the Arctic at the moment is that Arctic amplification, that is the rapid warming of the Arctic regions that has been observed recently, has drawn attention on the processes that are going on in the Arctic. So sea ice loss is a very important role in the conjecture of the Arctic amplification feedback. But the interaction of the Middle Attitude jet in general with high latitude forcing, so how latitude surface fluxes is not well understood, and that makes essentially the link with mid-latitude uncertain. So one of the prominent patterns of mid-latitude weather that has been observed associated with Arctic amplification is the so-called war-martical continent pattern, which essentially says that when the Arctic is warm, the continents are cold. So it can be a proper definition of the war-martical continent pattern that probably does not exist yet, but it can be identified in many ways in observation and reanalysis. It also turns out that if you look simply at the difference between the low-level temperature in late winter in the last decade compared to the rest of era interim, for example, what you see is a very strong Arctic warming signal with the cooling over the continents. This is true for late winter, but if you look only at December, for example, what you see is a very strong signal over the Barron-Sancara region, which is a region that is being studied a lot at the moment, with probably some cooling over Siberia. Now, one point is that models seems to do not reproduce this pattern associated to surface forcing in the Arctic region. And that is more or less what I'll try to discuss. It's also worth to point out that there's a lot, at least in observations and reanalysis, there's a lot of intracesional variability in these recent winters. So it's fair to say that in early winter and in late autumn, the signal is confined in the Arctic mostly. But then the evolution seems to point out that there's some interaction with the middle Arctic. And it happens that in these recent years, we also had a very strong and clear signal in the stratosphere, it seems to point out the warming in the lower stratosphere, essentially. So the reason why we need a lot of effort to understand how this connection work is probably there, as has been pointed out again in this recent paper, the interaction is essentially nonlinear and state-dependent. The response of the atmosphere to surface forcing in the Arctic is nonlinear and state-dependent. So somehow we need complexity in models to understand how the interaction works in the real world. Because we need to represent in the model essentially the real state of the atmosphere. So ultimately, we need to use realistic models. But on the other hand, we also need simplicity, because state-of-the-art models essentially disagrees. So recently, this authors pointed out that the state-of-the-art AGCMs that were forced with identical CIS laws produce significantly different circulation responses. And none of them was able to reproduce essentially the negative Arctic oscillation response that has been found by many model experiments. On the other hand, simple models are able to show the weakening and the greater shift of the jet stream that can also be used to understand why the warm article continent patterns arises. So one reason why we need complexity is also that at this stage, it seems that the community has not identified the precise mechanism for this interaction. And essentially, all the components that are important for military weather seems to be involved. So the basic idea that I'm trying to simplify and summarize a lot from this paper call and tell is that when you warm the article essentially, you change the equator to pole temperature gradient. But you also induce a direct interaction with the low level and then the driven jet. And these changes in the jet eventually can affect the circulation in the stratosphere, which can propagate downward the signal and give a delayed response. So these are more or less the ingredients of what I'm going to say and show in this talk. So I guess more or less you all know speedy at the moment, so I don't want to talk a lot about this. But we are using the full model at T30, only AGCM, so we prescribe SST, essentially. What we do is to make an ensemble of short runs. So we produce an initial condition from a long run with climatological surface force conditions. Then we produce short runs, one with climatological sea ice cover, and one where we reduce the sea ice cover in Valence and Carla. But we only apply this forcing for a very short time, only for we made two experiments, one where the forcing persists for two weeks and one where it persists for six weeks. And we start the simulation in midwinter in January the first. So what we see in the response is that in the first two or three weeks, here we see temperature in collars and zonal wind in contours, essentially in the first two or three weeks, the response is shallow, it's confined in the Arctic, and it's essentially linear and thermodynamic. I think Tito will talk more about this later this morning. But it's interesting to see that after a while, we observed the equator was shift of the jet. We also see that the warming in the troposphere is just shallow and confined somehow, but the warming is found also in the lower stratosphere. And it actually persists until the end of February, more or less. And the signal is dominated by the stratosphere warming. So if we look what's going on in the area where we are warming, the system, we always see the Trianotropascalzonal wind in collars and transient heat fluxes in V prime, T prime, essentially in shadings. So as I said, at the beginning, the response is essentially confined in the surroundings of the heating region. It's essentially linear, and it's a heat flow. Then after a while, you see that the jet in the mid-Latid is perturbed. Somehow, eddies tend to compensate the heating in the warming area. But also, there's an indication of a reduction of eddy activity in the region associated with the North Atlantic storm tracks. Then in February, which is the last plot that we see here, the signal is essentially confined in the North Atlantic where it persists for a while. And it's basically a negative NEO. So we think that we can separate the response in a range of two months, essentially, into a direct and linear one, a large scale response, which is somehow confined in the troposphere, and a delayed response where we think the stratosphere is important. So to look and to explain why we see, essentially, the strongest response when the forcing is not active anymore, here we quantify the impact of the anomalous tropospheric circulation in the stratosphere. And I don't want to go a lot into the details of this, but essentially what we did is to quantify, so this is V star, it's the eddy heat flux explained by anomalous waves. And we try to quantify from how the stratosphere is perturbed. And the only thing that I want to point out here is that if you separate this into components like anomalous temperatures and climatological velocity and vice versa, assuming that nonlinear terms are not important and indeed they are important, but anyway. So what you see is that this link can be explained by a linear interaction of anomalous waves and climatological waves, which is something that has been speculated and pointed out quite often in literature recently. I just want to point this out because if you look at the red line, which is the interaction of anomalous temperature and climatological wind, essentially, the signal that you find there is associated to, so the result that you get here is consistent with a perturbation in the circulation in the balance and carasease. So studies that look at how the, for example, blocking in high latitudes can perturb the stratosphere have also found this result. Without perturbing CIs, simply looking at how is the circulation when there's a blocking in those regions. So the perspective that I'd like to point out that it seems that our response is both tropospheric and stratospheric. So what I did here is to regress the signal onto the anomalous heat flux that we have just seen before and to look at the lagged relationships. So the dashed line that you see here is essentially the heat flux, while the black line is the geopotential light in the 30-acto-pascale, and the blue line is the geopotential light in the upper troposphere over the North Atlantic, while the red one is over Bar and Sankara. So what you see essentially is that as soon as you have a burst of heat flux, the stratosphere is perturbed, this burst of heat flux is preceded by a signal in the troposphere, which is in the North Atlantic and in Bar and Sankara, and which looks like this one. But after a while, you find the signal again in the North Atlantic. So the picture that we get, although we are mixing force and internal variability, the picture that we get is that if you get this response in your model, then you expect to see a perturbation in the stratosphere and also delayed response afterwards. So the crucial thing here is whether you get this or not, because this is a deep response in the troposphere. Indeed, this is a simple model with a simple stratosphere and not many vertical levels. But it seems that essentially we find a lag relationship, a lag connection between Bar and Sankara and the CIS and a negative, an AO, so that the equator was shift of the jet in the North Atlantic, if you want. It seems that state of the art models are able to reproduce this teleconnection. So this results from a paper written by Havier Gershia Servan et al. Recently, they looked at simip models and they made a statistical analysis to see whether they can find statistically significant teleconnections between the Arctic and the North Atlantic. And it seems that they find, so what you see here is a collection of these links in these models and these points here indicate the stratospheric pathway. So it seems that this kind of interaction can explain the connection that is found in state of the art models. So one could say that, well, the linear response is essentially not really active to perturb the state in the middle altitude. That most of the response that we have seen is explained by the laid response done also by the stratosphere. But that the key to understand whether this connection is real or not is to understand what happens in the troposphere and to understand whether your forcing is able to perturb and to change the jet. And people are trying to understand this because as I pointed out previously, the spread of the response that we get in models seems to indicate clearly that it's hard to determine whether this equator was shift of the jet in response to sea ice reduction is real or not. And recently people at the UK Met Office tried to propose a constraint to somehow classify and understand more the behavior. So what they did is they approached the problem with a Namib-like round experiment where they essentially perturbed sea ice in the Arctic, everywhere almost, on the ice age. So then they look essentially at the atmospheric response to ice loss in a similar way to what we have done with speedy. But what they get, what you see here is the response in the Amy front, which is, so this is the jet in the response in the North Atlantic at high latitude, between 50 and 65 or something like that. So what they get is essentially a positive, a weakly positive NaO response. Interestingly, if they use the coupled model, they get a negative NaO response. And what they argue that, well, maybe it's coupling that is important, or maybe it's the background state of the model that is important. And what they find that, if they simply change the sea surface temperature of the climatological surface temperature of the atmospheric model, and they replace them with the equilibrium the average sea surface temperatures in the coupled model, they get, again, the negative NaO response. So what you see here is essentially how one can infer the relationship between the response in the jet and the state of the system. They are trying actually to quantify the state dependence of the response. Observations like here, the one may be tempted to say, well, basically models that gives you the equatorial shift are wrong, because observation stays closer to the thermodynamic response, which is small. And it's not an equatorial shift. But one thing that they have also pointed out in their paper amongst many interesting things is that it's not really the position of the jet that is important to get that response. It's more how the model propagates waves in the troposphere, essentially. So they want to go very much into this plot here. But what they look at is the difference between the climatological refractive index between high latitude and mid-latitude. And essentially what they say, well, it's not really the relative position of the jet and the forcing that is important. What is important is where the model is able to propagate waves. So if the waves go more equatorial, essentially, you won't get this response. Anyway, what we're trying to do in this is to try to understand how the position of the jet in your system determines the response to sea ice forcing. So we think that these two points of view are not completely independent. They're not completely separated. Because the preferred direction of propagation of waves also depends, or at least they are not too independent measures of your system. And indeed, if you think about it, so what I'm showing here is just the average of the zonal wind from era interim at 300 and 150. If you think about it, surface forcing comes from many regions of many different latitudes and different longitudes. Some are upstream with respect to the North Atlantic jet. Some are downstream. Some are far. Some are in the Pacific. What could argue also that no cover represents an important forcing, which is in the middle of the continent. So we designed an experiment with a simple climate. I think many of you have seen the poster that Manuele presented on Monday, where he's trying to reproduce the same setup in open IFS. This is also the idealized speedy setup that we have seen before, I think, in the stone track stone. So what we do is, the idea is to introduce a thermal forcing in the system. So we use speedy dynamical core in alpha planet coupled to a slab ocean. And we change the cue flex in the slab ocean to warm an area in the mid-latitude. And then we perturb the high-latitude that introducing small and weaker forcing that should resemble the effect of sea ice loss on the ice age, or warmances of the temperatures, or any surface forcing. So the system, essentially, I think Lenka probably will show some results about this. But essentially, introducing this triangle, we tilt the jet, we perturb the jet, and we create a wave number one structure in the troposphere. And one advantage of this is that you also get rid of the stratospheric signal, because with the triangle we introduce stationary waves. But it turns out that they are not able to interact efficiently with the forcing and to propagate the signal in the stratosphere. So a second advantage of this is that you are not really perturbing the stratosphere. So whatever happens is happening really due to the interaction of the jet. And we repeat a couple of similar experiments. So again, a long run, we skip 20 years just to pick initial conditions from an equilibrated distribution and then we impose a localized forcing that is something like what we have seen before here. So a forcing in a latitudinal band of 15 degrees, more or less, and a longitudinal band of 50 degrees or something. And then we prescribe the forcing, which is zonally symmetric, and over the polar cap, up to the lower boundary of this localized forcing. So we've introduced the same amount of heat into the atmosphere, essentially. And we are starting sensitivity to many aspects. Actually, I will only show results for longitudinal and meridional position, but we have also seen how the magnitude of the heating affects the response, the temporal evolution of the response. And we plan also to change slightly the climate, the simple climate of our system, using a smaller thermal forcing in the Q flux. But unfortunately, the results have not really yet. That would be an interesting experiment, and it would be directly comparable with what people at the metal have done. So what I show here is the latitude of the upper level jet in this diagram. So the y-axis is the latitude of the upper level jet minus the zonal mean. This is more or less the position of the triangle. And on this axis, somehow I have longitude, but I'm plotting the number of the experiments that I use to label the experiments, which means that at every point here, I show results from a different experiment. So if we look at this plot now, experiment four is performed using the simple system with the triangle in the middle latitude, and a high latitude forcing, which is at 10 degrees, so upstream with respect to the triangle. While experiment eight is equivalent to experiment four, but the forcing is downstream at 210. So what you see on the right side of the right diagram is a collection of points coming from 24 different experiments. Each point there is the ensemble mean of 50 years of experiment. And we look essentially at the same quantity that has been defined here to investigate the response. But here it is highlighted to the upper level winds in the North Atlantic, while here in our experiment we use the zonal mean for probably obvious reasons, not obvious that anyway. And then I have three lines because I have used eight longitudes and three different latitudes. So the red one is highlighted, probably close to where, more or less, CI is various in the real world, while the green one and the red one has slightly more greater work. So the interesting features of this response are that if you're upstream, the response is very weak. Here we see the response after one year of integration. So the system had a lot of time. Equilibration essentially we're looking at the steady response if you want. Then there's a sharp transition as we move downstream. One could say that if you are downstream, the response is more efficient, is stronger. But the most interesting part is that when you get to these regions, not only you can have the strongest response in your system, but there's also an enormous variability with respect to the small change in the latitude of the forcing. So if I make an experiment, this would be barrens and cara, by the way. So if the triangle is the Gulf Stream, then this is barrens and cara. So imagine that this simple system could be used to explain what happens in the real interaction between barrens and cara and the North Atlantic. If I put my forcing here, I would say, well, there's a strong dynamical response in a greater shift. If I put my forcing here at 50, 70, I would say, well, there's not a big response. And probably I will find that the thermodynamic response is dominant in that case. So what I'm doing now is to average these regions that seems to have a similar behavior, so I'm averaging over long due to the central D. So I would call this region downstream, this one midstream, and this one upstream. And now we compare results with the zonally symmetric forcing experiment. And now we want to show the time series of this index so that you get sort of idea of how the response evolves. So I keep using red to indicate high latitude forcing. So we use red for this line, essentially, what I will not separate the other lines. And what you see is that it doesn't really matter if your forcing is localized or if it's zonally symmetric and if it's pressed everywhere over the polar cap. As long as you are away from the barren scarlet region in our simple model. So if I put the heating more or less where the storm tracks are induced, more or less over the triangle, or I put it everywhere in the polar cap, it doesn't make much difference. But if I put it far downstream, more or less in the region of the barren scarlet and Siberian seas, I can get a very strong, stronger response. And indeed, if now I average over longitude, if I average over latitude, the latitude of the forcing, and I look at the equivalent of a molar diagram, so we have time here, and the longitude of the forcing here. And again, we look at the response compared to zonally symmetric forcing. What you get is that also the response has a wave number one pattern. So if I put the forcing downstream, the localized forcing is more efficient to push the jet, the equator work, compared to zonally symmetric forcing. While if I put it upstream, it is less efficient. So there are many aspects that still have to be investigated in these experiments. And of course, it's a simple system. And I want to just go back quickly to what we are doing. So we don't have many ingredients of the system. As Franco also pointed out before, maybe wave number one is not enough to capture all that. You're doing half of the work. Yeah. You'll see that you have another piece downstream. Yeah, we could add another piece. We could also add a continent here, maybe to have a sharp collapse of the jet. Anyway, one could build complexity. But I think what we're really trying to do in these experiments, and we should keep it simple, because at the end we would need to get to a state of the art model. So there's no reason to probably add complexity at the moment. What we're really trying to do is to understand why there are some positions that somehow are resonant, and why the fact that the response is localized in some areas can give you a stronger response. So thinking in terms of the equator to port gradient controlling this is not enough. So the position of the heating and the response of the troposphere to that may be a wave train, essentially perturbing also the low level jet. And yet the feedbacks in the jet should be taken into account, and that's essentially the message. I hope I'll have more results soon. But I think, essentially, I can draw some rather than conclusions, maybe it's more of a perspective. So the response to ice losses, both linear and thermodynamic and dynamic, say. And this has been found out probably 10 years ago by early studies done in the United States. But there's also an important component which is driven by the stratosphere. Essentially, if you want to perturb the stratosphere, you need a deep response in the troposphere. And we've seen that it's not clear how you get this deep response in the stratosphere, whether it is real or not. It should be also pointed out that it could not be meaningful because we are assuming that the system is forced by fluxes in the high latitudes, while instead, I would say that the problem is intrinsically coupled. So not only you have to be able to respond in the right way to the forcing, but the motor should be able to prescribe the correct forcing. So we also need, ultimately, to know how sea ice variability is controlled by the atmosphere. And while it seems that the community is looking forward to doing more intercomparisons and classification of this response in models, the understanding processes could be actually also important. And I think that's the top here and just the reference.