 So, good morning, and thank you for inviting me to this conference. I'm going to talk about interstellar connections with the region of the North Atlantic, but in general the results I'm going to present can be extrapolated to the whole globe, teleconnection with Enso, and the results are in the line of what Enki Shah Khan has presented before, because I am going to focus in non-stationalities of the teleconnections with Enso. These non-stationalities occur at multi-decay dial time scale, and I would like to propose like a mechanism that will maybe try to entangle the question of if it's, we kind of take this kind of pattern by internal variability of we can have an external forcing that can make that the teleconnection change at this time scale. So this work is going to be a summary of many research work that we have in our group in Madrid. I have a problem with that, and the rest of the authors are not. So I'm going to focus in two regions in the North Atlantic, one in the tropical areas and another in the stratropics because the teleconnection mechanisms are different, and teleconnection provide a paradigm for understanding connection between remote areas, and they can be interpreted in these regions in terms of Rosby waves, equatorial Rosby waves, and in this tropical and stratropical Rosby waves. So I'm going to talk about this kind of teleconnection. This is a paper in Alexander, in the journal Climate to Transcendent 2 in which the Enso in the stream function in 200 topazcal is indicated in contours and rainfall in shading. We can see this, the typical response with two anti-cyclists is trying to indicate and then we can see also waves all around the world that indicates the teleconnection associated with Enso. In the tropics, this teleconnection is in the form of equatorial Rosby waves. So there is equatorial heating, then convergence in surface level and divergence in upper level this changes the coriolis parameter slightly around the equator and then two ions and cyclones appear and then forming a Rosby wave, a symmetric to the equator and then a Kelvin wave to the east. In the Sahelian rainfall, this response affect over the Sahelian rainfall that was stated in a paper in 2001 by Robert and he found that the response to a linear over Sahel was a reduction of rainfall. Later on, other authors have found this relation and in this paper from Moino et al. in 2011 we can see this response when imposing a linear pattern over the Pacific, over the Indo-Pacific realm and we can see these trying cyclones in upper levels and a subsidence over the Sahel that reduce rainfall. And then in the estratropics, what happened is that we have this grid response around the equator, which is baroclinic, but then in the estratropics there is a barotropic wave. This barotropic wave can teleconnect the answer with remote areas and this happens because of the conservation of the absolute vorticity. In the region of Europe, I'm going to focus in the season in which there is the strong teleconnection, which is the late winter and early spring, although there are also works focusing in fall. Morona and Guillard has found this kind of teleconnection related locally with a negative NEO pattern, but it's more global and here we can see the sea temperature and sea level pressure in a paper of Broneman in 2007 and we can see the Rosby wave coming from the Pacific up to the region of Europe, which is stronger in late winter and spring. But this is a stationary view. All of this work have been done using long time series, they calculate the teleconnections without taking into account what happened with the mean flow. So then it's important to know that the connection depends of the upper level mean flow as Rosby waves propagation depend of its intensity and location. So we can see here that the wave number depends of the meanders and intensity of the jet, the mean jet, and the propagation of Rosby waves in arcane patterns are stated in Hoskin and Breeze and other papers, take place in the regions when there are weakening in the jet, weakening in the jet, and when are weakening in the jet, we can find weakening in the jet when there are strong SSD gradients and when there are strong SSD gradients depending on how is the gradient, there is a strengthening or a weakening of the upper level jets. So we can think that, well, if we think in these decadal patterns, they receive it marionel and sonar gradients in their structure. So they can weaken or they can strengthen the jet stream or also they can weaken or strengthen the upper level flow in the equator and then the worker's circulation. This can be seen in the thermal equation, thermal wind equation, is not working, I guess. So we have a hypothesis that the connections in terms of Rosby waves depend of the climatological upper level winds, changing the intensity of the upper level winds are related to changes in the SSD gradients, teleconnections with esthetropical North Atlantic requires weakening of the jet, teleconnection with the tropics require changes in the mean worker's circulation. So decadal SSD modes exceed the strong barrier of the sonar gradients which could change the way in which the teleconnection takes place. So we are going to see this in observations or the first question is do we see this in observations? So in the paper of López Parajes and Rodrique Fonseca in 2012 and also in the whole PhD pieces of López Parajes, who is going to present it in the next week, he has focused in the teleconnection in late winter spring and if you look at the leading EOF at the rainfall, you see the PC and the standard deviation in 20 years window, there is a change in the amplitude. This is the filter, this is the rainfall filter, but the amplitude exceeded changes in the standard deviation at the decadal time scale. And then if we take the principal component on project and project it in the SSDs wall dwells like a Nino finger print, but if we do moving correlations between the Nino and the reading EOF, we can find that at the beginning of the 20th century and after the 70s and in phase with the evolution of the A and B, there are teleconnection or there are no teleconnection because in some decades, like here or here, there are no significant or even zero. And, sorry, so in this paper in the López Parajes and Rodrique Fonseca, there are some composites in which we can see that there is no signal of ENSO in the rainfall variability associated to the leading EOF and it's internal variability which is leading this EOF because there is an annual pattern, very, very constrained to the Atlantic and no external forcing at all, so it can be led by the mean flow and interaction, but in the other decades, there is a force, a forcing. So then, in his thesis and in a paper that is submitted to Journal of Climate, he has looked this in all the EOF industrial control simulations, semi-five models and it's a complex methodology in which we have done a cluster analysis and we have looked at this multi-decadal modulation of the teleconnection with ENSO and we have found two different periods, periods with this kind of base background state in his thesis and periods with this background state. We can see that following this thermal equation, there are rears in which the gradients are in the way that they go from a warm to a cold region, so there is strengthening of the jet. In shading, we can see the anomalies in the jet stream, so red implies strengthening, so there are periods in which there is a strengthening on the jet stream, just in the regions in which the teleconnection should take place, so it's not going to pass, it's not going, there is a lot of particular shared, it's not possible that the, there is a very straight jet and it's not possible the teleconnection to take place, but there are other periods in which this gradient favors weakening in the jet, the blue is the weakening in the jet and then the teleconnection could take place. If one look at the stationary wave number, the red indicates, this is the difference between these periods and these periods and the red indicates the regions in which the teleconnection should take place, so in these P periods, the telecollection is favors in this direction and in this direction. So then what we see is that in one period, the teleconnection almost does not have an impact in Europe and this is little and opposite than in other periods, in which there is a stronger teleconnection and the, the, the impact in rainfall is the opposite, more rainfall as a city to enter, less rainfall as a city to enter and we see this in the observations also. Going to the Sahel, in this paper in Dozada et al., we can see here different correlations between different SST index and we are going to focus in those related to El Nino and Sahel and we can see that there are periods in which there is a strong telecollection with El Nino and Sahel and rainfall, another which no, no relation. We, we regress this curve, we normalize it and regress on, on SST without, we have found this, this SST pattern and this is the, the SST pattern is like an entering is very gradient and, and could favor the, the change in the teleconnection. So we can use these to, to, to predictability, to enhance predictability. We have applied this with a model that Roberto Suarez is developing for his PhD and he has just published in the GMD journal and this is a statistical model in which we, we have called it as forecast because of SST force, statistical seasonal forecast and we, we do maximum covariance analysis between SST and rainfall or between SST and other variable and then we, we study the stationarity and see the windows of opportunity for the predictability, for the prediction. So here we can see the correlations between the SST and the rainfall and when we impose, we, we, we just focus in the Pacific and in the, we have used station data provided by Andreas Fink, they are published in these papers and I went on in 2015 and we found, have found with the station and with CRU data, periods with strong correlation and periods with no correlation between ENSO and SST. And for this period we have done a hind cast and there is an increase of a skill over the Sahel because ENSO. But for other periods we have done a hind cast and there is no skill. So this would be useful for open windows of opportunity for ENSO predictability. And then we project this, we obtain this pattern which is the same that I have found before, I have shown you before and which is characterized by anomalies in the Indian Ocean, in the Southern Ocean and also a AMB pattern. So we are now working in another paper which is in preparation and we have, go one step forward and we have said, well, we are going to do EOFs of the regression maps between the stream function and the NINU. So every 20 years we have a regression maps, map. So we have like an evolution of regression maps so we can see different picture of the teleconnection around time. And then we do EOF and see what is the teleconnection that happened more often and how does this teleconnection evolve. So we can see if the teleconnection evolved at multi-decay time scale or not. And we have found this, which is the same evolution that we found before. So for this Sahel, there is a change, so it's the last. So in contours I saw the stream function in the periods in which there is a strong teleconnection and inserting the periods with weakened teleconnection. And here is the SST forcing inserting and the upper level wins in contours. So we can see that there are periods with strengthening of the worker cells, which will be associated with a background state that is opposite to this, a cooling in the Southern Oceans and periods with a weakening of the worker cell, which is like this. And in these periods of the weakening of the worker cell, the teleconnection is favourable. Why is favourable? Because the key response requires to affect the vorticity and then it's favourable. So for conclusions. So the teleconnection should be interpreted at the sum of a stationary and a non-stationary component, the non-stationary component of NSO, teleconnection, various at multi-decay time scales, according to the cadal oceanic patterns. These decadal changes are related to changes in the upper level wins, which determine the response to a thermal forcing, changing the waveguides and associated teleconnection patterns. In Europe, the AMB global pattern seems to be responsible of changes in the upper level wins, which determine the existence of teleconnection, more effective under negative AMB. The Semid-5 models exhibit multi-decadal modulation of El Nino, according with the PDO-AMB pattern. And in Sahel, the response changes with a trend pattern related to changes in the mean worker circulation, which then determines the location of the response to an equatorial diabetes heating. And the non-stationarity of NSO teleconnection and the existence of a multi-decadal modulation opens windows of opportunity when predicting with using NSO, using El Nino.