 Thanks. It's my pleasure to welcome Dr. Yuhei Takaya. Yuhei is a researcher at the Atmosphere Ocean and Earth System Modeling Research Department for the Meteorological Research Institute, MRI in Japan. Yuhei is an expert on climate dynamics, climate modeling, S2S prediction, studying all aspects of S2S prediction, including forming ensembles and how best to set up ensemble prediction for S2S. Yuhei was one of the founding members of the WMO S2S panel since the beginning of the international S2S panel, and is also currently a member of the WIGSEP, which studies sub-seasonal to interdicatal prediction. Thank you, Yuhei, and look forward to your lecture. Thanks, Anish, for your kind introduction. I'm sharing my screen. Yeah, it's not full screen yet. Do you see my screen? Yeah, I see your screen. It's not full screen yet, Yuhei. Okay, swap displays, and then so that the arrows on top. Yeah. Okay, perfect. Thanks to the organizers in particular, Anish and Judith, for making this happen. In this talk, I'm Yuhei Takaya with the Meteorological Research Institute of Japan Meteorological Agency. In this talk, I was the two important S2S related topics, monsoons and extreme events, and S2S predictability. So I decided to give the essence of the weather and the climate extremes, the statistical aspect and dynamical aspect of the extremes and the monsoon. And I will give an example of monsoon extremes, which is the Mayu Bayu rainfall in 2020. So let's begin with the monsoons. The monsoons originally refer to the change of seasonal winds. And sometimes the monsoon is defined based on the seasonal precipitation. And here we see the global monsoons. We know that the three dominant regions, regional monsoons, African monsoon and Australian monsoon and American monsoons. Here you see the differences of the 850 hectopascal wind and precipitation between the vulnerable winter and the summer. And the vector shows the wind difference and the color shows the precipitation difference. And we see the salient, prominent seasonal change of winds and precipitation in these regions. And these monsoons, these monsoons are preliminary due to the land-sea thermal contrast. And in the lower panel, we can see a large portion of the total annual precipitation is brought about in the rainy seasons. This figure shows the rainfall percent in summer with respect to the annual total precipitation. For instance, the Asian summer monsoon has roughly more than two-thirds of total precipitation in the rainy season. And once the extreme rainfall or both in the annual and inter-season time scales occurs, it results in the devastating the consequences, the disasters. So predicting the monsoon in the S2S time scale is quite important for disaster mitigation. So here I'm showing the yearly occurrence of extreme events in the U.S. whose impacts cost greater than 1 billion U.S. dollars. And we see the magent color and the orange colors. They mean the number of the events by the flooding severe storms. So every year we had a lot of many severe weather and climate events. For example, so studying these severe storms are important. So the American Meteorological Society issues the supplement, a special supplement to the building of the American Meteorological Society every year. And you can see the detailed analysis of the severe event in particular years. So first we'd like to define the extremes. I took the quote from the AMS Grocery of Meteorology. And it defines the extreme as in climatology the highest and in some cases the lowest value of a climatic element observed during a given time interval or during a given month and season of that period. So here I'm showing the May you rainfall. This is in June, the Chinese rainy season. And we see the time series here. And if we have the 30 years monthly observation, the monthly season extremes occurs only a couple of times. So the extreme is really event is really a rare events. So it's difficult to estimate the return period and the return level. And the return period is the extremes, how often the extreme occurs. And the return level means the extreme, the level of the extremes. And usually we define the extreme events above or below about five to 10% levels based on the observations. So if we have the 30 years observations here, for example, we separate the 10 years periods in these chunks we can find, identify the extreme events in these 10 years. So offering the level of these values we can identify and define the level of the extremes for the 10 year period. And the defining the threshold of extremes are very important for the presentation analysis. So I'll introduce the extreme value analysis theory. So if we have the N samples and we define the maximum of the data X max. And if we had an infinite number of the data samples we could have the estimate of the probability distribution function of extremes. So the probability function gives the distribution function that the maximum value follows. And now we know that the three types of the distributions can be applied for the extremes. And now these three forms are formulated in a unified function of the so-called generalized extreme value distributions shown here. And once we can estimate these distributions the beauty thing is we can calculate the return period from these distributions. So after some computations we can have the return period for certain extreme levels, return level. So this expression is quite useful if you can obtain these shape parameters here using sufficient number of samples. So with these three forms we can compute the how often and how high the return level and return period. And here I'm showing the extremes of the monsoon's example of extremes. So in this talk I used the dimensions of the monsoon. In this talk I used the May you buy you rainfall. This is the rainy season in the East Asian summer monsoon. And here we have many dots here. This is the monthly averages of rainfall with respect to calendar months. And we see the red dots this denotes the value of 2020. So the 2020 rainy season is extraordinary high present the extraordinary high precipitation. The impact of hydrological extremes are substantial in rainy season because the rainy season is basically we had more rainfall than the other seasons. On top of that, on top of the seasonal change if we have the extreme event then that cause disasters flooding. So the investigating these events are very important. And in June and July data the most of the years in upper 10% were suffered from flooding actually in China. So using the GV distributions we can fit the distribution and we can compute the return level and return period. For example, this figure shows the extreme events for given a 10 year period the level of the extreme is about the eight millimeters per day. So we can know that the severeness of the extreme event and the return period using this statistical method. The other way to define the threshold of extremes is just counting the observed precipitation. For example, if we have the 30 years later and we took the highest three years then we can define the threshold 10% of the threshold's highest extremes. And we can also compute the exceeding probability and we can fit the exceeding probability using so-called the Pareto distribution. So these are the basic mathematical or the analytical theory that can be used to analyze the extremes. Now I'd like to discuss the prediction. Why extremes are difficult to predict? Please think about it because it is rare. The rareness maybe may not be a complete correct answer. Actually the extremes are difficult to predict because the rare event often occurred due to the chaotic variability which is likely unpredictable with a long lead time, for example, the sub-seasonal times scale. So I will show an example for explaining this. So if it's unpredictable, what we can do? We just keep up. Actually the seamless prediction will help. So the seamless prediction is very important to predict the extremes. So to explain this, I'll take an example of early summer 2020, the Meiyu Bayou rainfall. And here I'm showing the time series of Meiyu Bayou frontal zone rainfall. And as I said, the 2020 is the record high, has the record high precipitation. And here I'm showing the spatial distribution of anomalous rainfall during these two months. And we see anomalous precipitation over China and Japan. And here I'm showing, this is just the zoom up of this box area. And the vector shows the anomaly of moisture flux. And we see a lot of moisture transport in these regions. And the enhanced rainfall accompanied with the southward extension of western north Pacific subtropical high, which is typical of in the western Pacific Capacitor Mall. I will explain this in the next slide. And the in the western Pacific Capacitor Mall accompanies with the warm Indian ocean condition. And this condition typically occurs after the El Nino event. And with this warm Indian ocean SST, we have enhanced precipitation over the Indian ocean and suppressed convection over the western north Pacific. And over the Yangtze river basin and the Meiyu Bayou rainfall, Meiyu Bayou frontal zone, we have more rainfall. And this is very typical conditions. We can observe the Asian summer months soon. And the other impact is with this mode, we tend to have the high temperature over the Indian subcontinent and the mainland of Southeast Asia. So this model has a pervasive influence on the Asian summer climate. And the JMA analyzed the causes of the enhanced Meiyu Bayou rainfall in 2020. And the JMA gives the some eligible factors. For example, as I said, the southwestward extension of western north Pacific subtropical high. And moisture transport to the Meiyu Bayou frontal zone and upper level 12 over the ELC and south world shift of jet under the warm Indian ocean. But this expression does not give the primary cause. And the attribution study is important to understand the causality and the S2S predictability. So attribution study of extreme events is essential to understand the predictability and its underlying mechanisms. So if you can identify the key factors with the high potential predictability or high predictability for a particular extreme event, maybe you may be able to predict this event. So understanding the predictability and understanding the causality is very important for the subsaison prediction. Subsaison to season prediction. But in general, it is not easy to separate the contribution of many factors, as I introduced in the previous slide. But we can use the model experiment to untangle the cause and effect. And maybe it was noted that the extremes are not necessarily predictably in the S2S time scale. Meanwhile, if there are preferred conditions or precursors of extremes, then the extreme event may more likely occur in such conditions. So I did the sensitivity experiment to assess the tropical influence in the ocean aesthetic impacts. We conducted the two experiments, EXP control and EXP in the ocean climatology with these aesthetic conditions. Then we find the clear impact of the warm in the ocean. So using this sensitivity experiment we can attribute the... we can identify the cause of the impacts made by rainfall. And here I'm showing the circulation maps. And the one thing we didn't know is the why, what caused the warm in the ocean. As I said, the warm in the ocean condition in early summer 2020 can be... the warm in the ocean condition in early summer 2020 did not accompany with the preceding El Nino. But actually we can trace back to the record strong positive phase of Indian ocean dipole for 2019. And here I'm showing the subsurface conditions in the ocean. And basically we see the dynamic ocean variability in the ocean. And here I'm showing the ocean heat content or the upper surface temperature anomalies for the seasons. And we see a very strong downweighing loss of the web in the south Indian ocean. And we also see the Kelvin web in the crater shown here. So we can identify, we can know that the Indian ocean dynamics actually can explain the extreme event in the Mayu Bayou rainfall. And let's look at the the sub-season, season predictions with a different lead time. And here I'm showing the predictions of the Mayu Bayou rainfall starting from the initial date of 26th of April, the end of April. And this resource is based on the predictions from the 16th of May. So we see a larger shift of the PDF. By the way, the red color bars shows, pink color shows the PDF of the hind cast. And the right blue bars shows the forecast. And we see a larger shift in the closer prediction from closer initial date to the event. And the causes of the enhanced Mayu Bayou rainfall is not only the tropical SST but also the exotropical circulations. Here I'm showing the intermember correlations between Mayu Bayou rainfall and the atmospheric fields, the 500 hectopascal height, 200 hectopascal zone of wind and the sea level pressure. And basically, for example, this figure shows that if we have the deeper trough here over the EOC, we tend to have more precipitation over the Mayu Bayou frontal zone. And then, for example, in this figure, if we have the more stronger the subtropical height here, then we tend to have the more precipitation of the Mayu Bayou rainfall. So these conditions are precursor for the enhanced Mayu Bayou rainfall event. We heard a lot about the MGA in France. In the summer, we know that there is so-called boreal summer in seasonal oscillations. Actually, in this year, we have more active boreal summer in seasonal oscillations, more active convection in the ocean. That favors the active convection of the China. Here, I'm showing the MGO diagram, and we see the convection was active was active over the Indian Ocean in June and July. And this is the precipitation anomaly prediction, prediction of precipitation anomaly at the end of April and the middle of May. And we see more precipitation, a larger anomaly of the precipitation over the Indian Ocean. This indicates that the the active convection over the Indian Ocean is better predicted in the application with a shorter time. I will show some other impacts. For example, in 2020, we had extreme high temperature over the southeastern Asia. This is typical of the Indian in the western Pacific mode. We see the similar characteristics in the observation. The other impact is the record low western north Pacific cycle activity. In fact, there was no topical cycle during the June to July 2020. So in the pre and early TCCs. This condition is often observed in the summer after the new years. Here I'm showing the composite of the TCC count for in the western north Pacific region. So the condition in the last year actually was similar to the typical condition of the IPOC mode. So take-home message. I gave brief introduction of analytical and physical aspects of the extremes. The extreme value analysis is useful, but its ability is subject to the sample size and the extremes of monsoons often cause devastating hydrological disasters. In general, extremes in tropics are caused by multiple causes of climate interactions. So the entangling, the causes are very important research topic for the subsistence of prediction research. And predicting extremes would be useful due to the inherent chaotic nature of the climate system, even if we had a perfect model. So this topic actually is very challenging. The challenging body important research topic. To predict the extremes, the seamless prediction approach by changing the lead time. We have a continuous prediction we tend to detect the risks of the extremes. So it's a key by bringing the gap between the extreme range of predictions. I think that's all I have today. Thank you, Yuhay for comprehensive talk.