 So welcome to the late morning session in the S2S workshop. It is my pleasure to introduce Heilin. Heilin works as a researcher at Environment Canada. And his research focuses on dynamic processes with a time scale longer than a week that influence medium and extended range forecasts. I thought that was a really nice description. He's also an adjunct professor at McGill, and he will be talking about the NAO influence on the MDO. Welcome Heilin and thank you so much for giving a talk at this workshop. Thank you. Thank you Judith and Abish for the invitation. So today I'm going to talk about the NAO influence on the MGO. It's a bit different from what we have been listening and heard. And I also want to acknowledge my collaborators, Zhiyun Huang, Harry Handen and GeoBear Blunei. So first of all I want to do a brief summary of the MGO impact on the NAO. It's quite well known and we also heard quite a lot yesterday. And after that I will show you some observed NAO influence on the MGO. So it's like another direction influence. And after that I will show you the impact of this influence on the forecast scale of the MGO from the S2S models. And then I'll give you some possible explanation for why this happens. And finally it's a summary. So this diagram we saw it yesterday a couple of times but in different format. I believe Friedrich showed this and Jaime also showed the seminar figures. It is the composite precipitation for eight phases of the MGO. So what you can see is the phase one, two, three up to phase eight. You see the eastward propagation of the precipitation anomaly. And there's two groups of phases two, three and six, seven. What you can see is there is a dipole structure of the precipitation anomaly with enhanced precipitation and convection in the Indian Ocean and reduced connection in the western Pacific. And the phase seven, six is just opposite. Quite a lot of studies demonstrate that these two phases, these two groups, the dipole structure of the precipite convection has the greatest impact on the teleconnection and the Lawson-Hemisphere exotrapics. So the NAO, and also you may also know about NAO, it's one of the most important mode of variability in the Lawson-Hemisphere. It's stronger in the winter season than in summer. And the variants explained by this single mode is quite a lot. There's about 31% of the variants in winter. Air temperature can be explained by this mode alone. And also not only temperature, there is also precipitation and humidity in North America and Europe. It's quite significantly influenced by this mode. All right, so I started by the influence of MGO and the NAO because the MGO is a tropical mode and the NAO is a variability in the Northern Hemisphere and with centers in the high latitude. So these two modes, they have teleconnections. So this is from early study. We show that the lag composite of the probability of occurrence of NAO phase. For example, this table shows the phase 1, 2, 3, to 8 of the MGO and the lag in pentaths, that's five days. So zero lag is the same pentath as the MGO. For example, the MGO phase 4, there is a 45% chance of aperture sale, that's above normal NAO index. So there is an indication that when the MGO is in phase 2 and 3, after about 10, 15 days, there is a high probability of positive NAO occurrence. So that's very consistent with what Frederick showed yesterday and also in the study of Kassu. And after MGO phase 6 and 7, you see the negative, the lower aperture sale of the NAO index. So that's the negative NAO. And the mechanism that there is an explanation for this lag teleconnection is one is the possible Raspberry propagation. And another recent study showed that there is a stratosphere pathway and the MGO can influence the loss of hemisphere NAO. So that is known. So the MGO influence in 2D dimension is for the MGO phase 3, zero lag, that's simultaneous. You see that the positive anomaly for geopotential height anomaly in North Pacific and the wave train just arc to the North Atlantic. And five days later, you see the further development of the geopotential height anomaly in North Atlantic. And another five days later, you see the development of positive NAO. So this is quite the opposite for the MGO phase 7. You see after about 10 days, there is a negative NAO development. So on the other direction, so today my focus is on the NAO influence on the MGO. So it's the opposite direction influence. So the similar in that paper we published in 2009, we did the lead lag for this MGO and NAO. But then this is again the phase of the MGO and this is the lead. That's when the NAO leaded MGO by one, two, three, four, two, five pentats. So what you can see is that the negative NAO leaded MGO by about three, four pentats. And the positive NAO leaded MGO phase 6 and 7 by about two to four pentats. But people can argue that because the MGO propagate eastward and they have influence. And when there's a cycle that can have this kind of like behavior relationship. But actually, as I will show, even when the MGO at first is very weak or there's no MGO at the beginning, the NAO can influence, can amplify the MGO afterwards. Okay, so the schematic figure of this interaction is the MGO. And when it reaches the western Pacific, it can send the Rosby wave and influence the NAO. And after the NAO is amplified or it's changed, it can influence the MGO again. And also if there's no this branch, the NAO variability because there is other mechanisms can generate the NAO. For example, the atmospheric internal dynamics and that can also influence the initialization of the MGO. So we look at the market regulation for 200 manabar zonal wind with respect to the NAO index. For example, this is the zero lag, that's the simultaneous. The positive NAO, you see the increased western wind in the North Atlantic. And 5 days later, 10 days later, what you can see is that this signal is moving southward. And at 4 and 5, what you can see is that the increase of western wind in the tropical Africa and Indian Ocean. So this is quite significant after the NAO, there is a change of zonal wind in the tropics. So that is connected to the influence that the MGO. So what we show here is the average between zero and 90 degrees east. That's over Africa and the tropical Indian Ocean, the zonal wind. The X axis is the latitude from 30 degrees south to 90 degrees north. And the Y axis goes up is time in days. So what you can see is this is the difference between positive and negative NAO. So it's the same, it's very similar to the positive after positive NAO. So what you can see is that after positive NAO, about two weeks, 14 days, there is quite a sudden increase of zonal western wind over the equator. And this also happened in the model, it's an ECCC model, it's very similar. So this behavior after the NAO, there is about two weeks after the increase of zonal wind in the tropical Indian Ocean. So now we will ask if this kind of connection between the influence of NAO can impact the MGO prediction scale. So when we look at the S2S models, there are 11 models, we look at the hind cast for the winter period. So I don't want to go into detail about these models. So what I want to show is that we select those cases when the forecast is initialized with strong NAO, and that's when the NAO index is greater than one. And then compare that with the forecast with the weak NAO initialization. And also compare the positive NAO forecast and the negative NAO forecast. So there's quite a lot of cases for these two kind of categories of forecasts. So this shows that after about 20 to 30 days after the NAO, you see the probability of the phase of the MGO. For the error interim analysis, you see that after 20 to 30 days after positive NAO, you see the increase of the MGO phase 6, phase 7. And when the NAO is negative, you see the increase for occurrence of MGO phase 3 and phase 2 and phase 3. So this kind of relationship exists in almost all the models. You see the some models, there are some more sensitive, some models is less sensitive. But it shows that after positive NAO, the MGO tends to happen in phase 6 and 7 after negative NAO, MGO phase 2 and 3 tends to happen after about 20 to 30 days. All right, so that's a phase. So about the amplitude. So this is the evolution of the MGO amplitude after strong and weak NAO. So for the error interim, that's reanalysis, you see there's about 20 to 30 days, the strong NAO lead to a stronger MGO. And also for the most of the models, but some models like yesterday we see that the most of the model, the amplitude MGO decays. But there's one model at increase. But no matter it's increase or decrease, but the relative amplitude between a strong NAO and a weak NAO, it's quite consistent. Okay, again, this shows the 0 to 90 degree average zonal wind at 200 millimeter bar for all those models. So what you can see this after about 15 days, most of the models can capture this like increase of zonal wind over the equator. All right. Okay, so that's the high level. How about the lower level? 850 millimeter bar, zonal wind and ORR. So for the reanalysis, you see that the blue contour, it's the elective zonal wind at 850 millimeter bar. So you see the, like yesterday wind alamony and the reduced connection and this corresponding to the negative to MGO phase seven. So what we observed. And this kind of distribution is also like assimilated quite well in the most of the S to S models for the KMA model, they didn't provide the ORR. And the wind structure is quite similar. So over Indian Ocean, there's East Italy and over Western Pacific, there's West wind alamony. All right. So let's look at the forecast scale, like grouped by strong NAO initialization and the weak NAO initialization. The red curve, that's the strong NAO start and the green is weak NAO start. So what you can see is that after about two weeks, there is a separation when the strong NAO lead to a better forecast scale. For most of the models, almost all the models except the one the JMA models, probably at the end there's some signal and also the French model. But for the other models, the scale is quite like the difference is quite clear. And we look at the lead time in days, how many days it dropped to 0.5 so that can be regarded as the scale of the MGO. So the orange bar is the strong NAO start and the light blue is the weak NAO start. So what you can see is that for most models, the strong NAO has a better, longer forecast scale for the MGO. And for the Australian model, ECMWF and NSIP, the strong NAO, they all can reach the forecast scale of 30 days of MGO. So how about the general scale instead of the MGO? So this is the average correlation scale of the Zonno wind, 200 meter bar, that's in black, and the Zonno wind at 850, that's in red, and the ORR, that's in green. So that's the average in the strong MGO area. So what you can see, that's the dashed line that starts with the weak NAO and the solid line that's with the strong NAO. So almost for all those variables, you can see that when you start the forecast with strong NAO, you get a better scale. And the ORR has the weakest scale, so that can be expected. And the 200 meter bar Zonno wind, the scale is the best. But we do see that the scale difference between strong and weak NAO forecast is quite clear for all those models. Helen, about two or three minutes. All right, okay, I'm about to finish. So there's also dependence on the NAO phase. We compare the forecast starts with positive NAO and the start with negative NAO. So the conclusion is that when you start the forecast with negative NAO, that's a green line, for most of the models, you get a better forecast scale of the MGO. The reason, okay, this is somewhere you see that the negative NAO leads to a better scale of the MGO. So the reason, we look at the wave activity flux for the positive NAO. This is after 21 to 30 days from the initial condition in initialization. You see that in both positive NAO forecast and negative NAO start, we see a strong like a southward wave activity flux in the tropical Atlantic. But in the negative NAO forecast, the southward wave activity is much stronger. So this can be, the contour is the stream function. This can be linked to when you have a negative NAO after 20 to 30 days, you've got the cyclonic circulation. There's a westerly wind in the lower latitude in the North Atlantic. So that helps the southward propagation of the rugby wave. And the possible mechanism for this is that when you look at the DJF, climatological zonal wind at 200 millibar, you see there's quite an area over the tropical Africa and the Indian Ocean. The basic state is east wind. So the southward wave activity flux in the North Atlantic, when they reach this area, they can become a forcing mechanism for the tropical upper Kelvin wave. And there's studies, earlier studies showed that the tropical Kelvin wave can be forced in the east wind basic state. So that can be the possible reason for why the North Atlantic oscillation can propagate to influence the MGO. Okay, to summarize is that we see there is a substantial influence of NAO on the MGO. And the NAO positive NAO need to face 6, 7 and negative NAO need to face 2 and 3. And we see that the forecast when initialized with strong NAO can have a better MGO forecast scale than weak NAO. And also the face of the NAO has influence also, negative NAO tends to be more skillful than positive NAO forecast. So, okay, I will stop here. Thank you. Thank you. Thanks very much. I have, yeah, please post your questions or raise hands if you have any questions. I have a question. So a lot of the S to S community focuses on precipitation and to meet a temperature. And I'd be interested if this state dependent predictability you have beautifully shown and demonstrated here translates to those variables. Yes, the examples I show here, that's for these three variables that compose the MGO index, Zung Nguyen 200 and Zung Nguyen A50 and ORR. So I think those differences in the tropical precipitation can be influenced as well because the ORR we see this difference. And the temperature, probably as well, but I didn't look at the temperature because the Zung Nguyen at the lower level is, there's a difference. So I guess the temperature also can show some difference. Thank you. Any other questions? The question from Anjie. Anjie, do you want to unmute yourself? Yeah, I was interested in your lagged composites analysis where you may look at the MGO, NAO, NAO-MGO influence or relationship. I wonder what would happen if you looked at the different phases of the NAO in a similar way in the lagged composites analysis would you find kind of a cycle there and if so what period or what lagged time period is important there. You mean for this kind of analysis? Yeah. You separate them because this is the take all the NAO, I'm not sure how to get to the only the positive NAO, because this identify the negative NAO, not lead the MGO phase 2 and 3 and identify the positive NAO, lead MGO phase 6 and 7. So if you only select one phase, you only get part of it. So I guess, is that right? Yeah. Now I probably meant more completely independent of the MGO phases just between basically like the, like an autocorrelation of DNAO basically, but in terms of these pentate legs. Okay. So, yeah, we did this analysis for example for the difference between the forecast scale of MGO between strong and weak NAO. We did this only for for the cases that has a very weak MGO at the beginning. So we still get this influence so that that means that it does not depends on initially you have a strong MGO or not. I don't know if that answer your question. Yes, I mean it's very interesting. I think there are lots of angles. Yes, it's interesting. Thank you. Thanks so much. Jan had a question, if you'd unmute first. She had her hand up before me. So, yes, I decided to. Okay, sorry. Then I can go ahead. I was interested in it and it brought in a new aspect. Okay, thanks very much for this interesting talk. I was wondering if you have looked at it, your relationship North Atlantic to the MGO for more from including more modes of variability for the North Atlantic so including blocking over Europe and Atlantic which for example. If not, would you expect to get different results on the different impacts on the forecast skill using more modes of variability? We didn't look at that but I think it's an interesting topic to do more investigation because we picked up NAO. There's no reason why we like to do other modes. Probably if you pick up the blocking and other European mode, you can also get some signal. But yeah, I'm not clear now but it's worth doing for sure. Thank you. Thanks. Jacqueline, please go ahead and mute and ask your question. Well, thank you. I was typing my question. Okay, so we know that the MGO occurs on like different synaptic patterns, like different seasons, different ends of states. And so it's effects on the extra tropics, for instance in your case NAO can vary considerably. Have you looked at how does the NAO influence the MGO at a specific season, for instance, or ends of state? Like, are you confident that your correlations for phase two and three or six and seven, do they happen or occur more frequently during a particular season during a particular end of state? Can you comment on that? Thank you. Yeah, that's an interesting question also. The analysis we did, that's only for the winter season. Because we know that in winter the westerly wind is strong in the extra tropics and that's easier for recipe waves to propagate. In summer, probably the influence is not as clear. But for other, like the modulation of ENSO and other internal variability, I think that some years when the background flow influenced by the ENSO can be more favorable for the propagation. And that can have a lot of different like influence. But that is, I think, further investigation to look at those interesting questions. It's quite useful. Yeah, thank you. Thanks very much. Thank you so much. Hi, Lynn, for your presentation. That's very interesting, a new perspective. We are moving to the next talk by Charlie.