 So, today I will continue our discussion on the Indian Ocean Dipole which we began in the last class. We were discussing the processes that lead to the mean patterns of variation and then we will go to what leads to the formation of the IOD events. So, I continue the discussion of the present understanding of the processes involved in the evolution of the mean monthly SST and convection over the EEIO and WEIO. Now, we have seen that the SST of both the regions decreases from mid-May to mid-August and that again I show you here again that the SST of both the regions this blue is EEIO and red is WEIO both of the regions start the SST starts decreasing from mid-May right up to this point here mid-August or so and then both start increasing after mid-September. So, we have seen that SST of both the regions decreases from mid-May to mid-August and the EEIO SST decreases because cooling due to upwelling generated by southerly winds parallel to the Sumatra course dominates the warming by advection of the warm water from the west we have seen this last time. We have noted that the average OLR over WEIO decreases from mid-April up to end of May and then oscillates around 240 watts per meter square up to the end of August. So, now we are looking at what is happening to OLR over WEIO and that is in red here. So, it initially is decreasing here which means convection is building up and then up to mid-August it simply oscillates around 240 watts per meter square then it increases in September and remains high until the end of the season. So, it starts that means convection starts disappearing from here and then it just remains suppressed in this part of the year from about mid-September onwards the convection over WEIO suppressed and OLR of WEIO is high. Now the SST of WEIO we have said before that WEIO SST responds primarily to fluxes from the atmosphere and so SST of WEIO decreases rapidly with the increase of convection. So, here we have when the convection increases SST of WEIO which is here decreases very rapidly here up to about mid-June till when the convection builds up. So, the SST decreases rapidly because of the radiation cut off by the clouds and then increases until the end of November once the clouds go away. So, from mid-May until end of November the mean SST of WEIO is higher than that of WEIO. So, we have from about throughout from mid-May up to end of November the SST of EIO is higher than that of WEIO because WEIO decreases very much more rapidly in the first part of the season with the build up of the convection here. So, now what is the situation then we have also seen that from mid-May to mid-August the atmospheric conditions appear to be almost equally favorable over the two regions for supporting convection. So, what does that imply? This implies a decrease in the strength of the Westerly winds over CEIO from mid-May to onwards. So, now you have the Westerly wind is denoted by black here and you can see it decreases rapidly as the convection probabilities in the two regions become similar you have less of a gradient and therefore less of a Westerly wind and then from here to here which is to say from mid-June to about September it remains very low lower than 1 meter per second. See this is a reflection of the fact that the convection over the two regions is very close. So, we have low Westerly winds in that period. Now, let us consider the variation of convection over WEIO and EEIO. See the most commonly used measure of organized deep convection over a region is the outgoing long wave radiation OLR which is what we have considered so far. Now, the regional average of OLR is a reasonable measure of convection and its variability for regions over which it is uniformly low. So, if you have a region filled with grid points with low OLR then there is convection over all of them and then the average OLR makes sense average OLR is representative of convection in the region. However, when a substantial part of the region is cloud-free with large values of OLR the OLR of the cloud-free part also contributes to the regional average. So, the average values of OLR for a region with on the weekly or monthly and larger scales can be relatively high despite the occurrence of deep convection for some days and or over some regions. So, on a single day part of the region may be clouded will have less OLR, but it will not get reflected in the average OLR if the other part overwhelms this region the cloud-free part. So, but what we are interested in is really not the variation of OLR, but the variation of convection. So, it is important to focus on where the convection is and how large is the extent of the region with convection. This is why indices such as the frequency of highly reflective clouds you know so far we have been talking of long wave radiation. Another thing the satellite measures is the albedo and you know deep clouds are highly reflective clouds and this is measured in the visible in the albedo. So, there has been work done in actually consolidating data on frequency of highly reflective clouds and atlas has been generated and this has been used as a measure of convection. Now, instead of using frequency of highly reflective clouds what we can do is we can use a convective index CI which is a measure of the intensity and the horizontal extent of deep convection over a region based on daily OLR 2.5 degree by 2.5 degree resolution. That is to say we do not go to smaller units of space than 2.5 by 2.5 we ask the question is the mean OLR of every grid point 2.5 by 2.5 is it low enough to have convection there and from such grid wise analysis generate a convective index. So, what do we do? We assume that on any day deep convection occurs only over the grid points for which OLR is below 200 watts per meter square and take the difference of the OLR value from 200 watts per meter square to represent the intensity of deep convection. That is to say how much lower is the OLR than 200 watts per meter square. So, if it is 180 the difference will be 20 if it is 160 the difference will be 40 and so on. So, that gives a measure of the intensity of clouding now the CI for a specific region such as EIO which comprises typically several grid points for a particular day is calculated as following. See what we see is over how many grids on that day was the OLR less than 200. We take only those grids for those grids compute the intensity as 200 minus OLR and sum over all those grids that is to say CI is the sum of 200 minus OLR over all the grid points for which OLR is less than 200 all the grid points which support deep convection. Now, the variation of the mean CI over EIO and WEIO is actually not surprisingly consistent with variation of OLR. So, we have here OLR variation which we had before and we have the variation of CI. Now, we had seen that OLR of EIO is lower than that of WEIO and indeed CI of WEIO is lower than that of EIO. So, the thing is consistent even the variation is generally consistent variation of CI is generally consistent with this. However, when we talk of 240 being a reasonable limit so that if OLR is below 240 we can say the region is convecting we have here regions or days on which WEIO is above 240. But even then you see it has non-zero values of CI this is exactly reflecting what I said before that there is a considerable cloud free region which is leading to high values of OLR. But actually it is not indicating it is not an absence of convection which is what we would think of if we looked at only the mean OLR. In fact, there is convection of the order of 4 convective index of about 4 in this case. So, in fact the convective index is never 0 which is saying that there is always some convection going on in the mean over some part of WEIO and some part of EIO. So, while the convection occurs intermittently over WEIO and EIO in any season during a season with a strong positive phase of equino such as 94, 97 and so on the propensity of convection over WEIO is higher than that over EIO this is because there is a strong IOD event there. On the other hand in the negative phase of equino which again we will see the propensity of convection is higher for EIO. So, what you see here is a somewhat complicated figure but I will explain to you what it is these towers represent convection over EIO and this is the convection over EIO. The convection over WEIO convection over EIO is red convection over WEIO we multiply that by minus 1 the CI value so that we can see it because otherwise there would be too much overlap of the towers. So, this is the convection over WEIO and imagine that this means lot of convection and this means less convection. So, we have simply inverted the axis so to speak for convection over WEIO. So, this is CI for EIO the red towers and blue towers are for WEIO. Now, what you see here is the SST of EIO, SST of WEIO and various other parameters there is also this is the SST of WEIO and WEIO and you can see that in this case these are dipole events. So, SST of the west actually is larger than SST of the east over a large part of the region this is a reflection of it being a dipole. So, what do we see for 94 and 97 that initially there is convection over EIO, but then it is largely suppressed here you see the red towers are very few here and same thing here it is the blue towers that dominate the red when you have positive IOD events and exactly the opposite occurs in the negative equino phase where you have the EIO convection dominating the WEIO convection, but note that it is not as if convection is totally 0 rather the propensity of convection is less over WEIO in negative phase and less over EIO in the positive phase of equino this is to be born in mind. So, a transition in the phase of equino involves a change in the probability or chance of convection occurring in over WEIO or EIO it does not involve a total shift of the convection from EIO to WEIO or vice versa. This is something very important to remember that it is not as if an entirely new regime takes has come into place when we have an IOD event rather the probabilities have shifted from what they were. So, now it is important to understand what are the factors that lead to variation of convection various. So, let us consider one of the factors suggested how does the convection vary with local SST in particular we need to understand the relationship between EIO convection and SST. Because we need to understand the factors that lead to the anomalous suppression or enhancement of the convection over EIO and WEIO in any season. Now, important factors could be remotely driven descent or ascent over the equatorial Indian ocean. For example, we know that phenomena like El Nino do drive descent or ascent over the Indian ocean and that could that could be the so the suppression or enhancement could be a manifestation of remote factors operating on the region. However, the SST of the equatorial Indian ocean could also play a role. For example, it is believed that the suppression of convection over EIO during the positive phase of equino in 1994 is associated with the cold SST anomaly of EIO. So, because even in the original paper of Saji et al and in all the succeeding literature it is very clearly stated that the suppression of convection over EIO and the cold SST anomalies go hand in hand for positive IOD events. So, let us now look at the relationship between OLR over EIO and the average SST of EIO. The variation of the monthly OLR over EIO with the SST of EIO for all the months in the period 82 to 2009 we see here. So, what is this each of these points represents one month and for example, for this point this is a specific month in that period and for this period the average EIO SST is about 27.5 and the average OLR anomaly is about 15 watts per meter square. So, this is how all the data for all the months between 82 and 2010 are plotted because each month is characterized by a pair of values one for the SST and the other for the OLR anomaly. Now, when we plot this what do we see 27.5 is a threshold here for convection and what we see is once the SST is above the threshold it is a big mess you know you could either have suppressed convection or you could have enhanced convection. There is a large range of variation particularly above 28 degree, but once it is below 27.5 you see all the points you have positive OLR anomaly that means once it is below the threshold of 27.5 convection is suppressed. Now, to what extent varies it could be suppressed with an OLR anomaly of about 20 watts per meter square or it could be very much more suppressed with OLR anomaly is between 40 and 50 watts per meter square. So, the extent does vary here, but there is no month in which the OLR anomaly is even within 20 or negative. So, this is a very clear Lakshman ratio if you wish or a threshold of 27.5 which seems to be operating in determining what is the kind of convection that is possible for a given SST range of EIO. So, we have noted that the most intense suppression with OLR anomaly is greater than 45 watts per meter square occurs when SST here this is the most intense suppression. This occurs when SST is less than 27.5. However, while all the points with SST less than 27.5 are characterized by OLR anomaly is greater than 20 watts per meter square which we saw comparable anomalies also occur for other SSTs. So, you have here you know points like this and comparable anomalies here occur also for much warmer SSTs here you see that. So, it is difficult to given an OLR anomaly to say what the SST range would be unless of course, the OLR anomaly is so large as plus 45 or so watts per meter square in that case we can say the SST is definitely below 27.5. So, now all the colored dots here represent IOD events purple is 94 all the IOD events in that period red is 97 and blue is 2006. So, you can see that certainly for some months in each of the IOD years SST goes below the threshold and that is where you have sustained suppression of convection over EIO. But you can also see that in the other months before the IOD has attained a mature stage in fact, the SST is higher than the threshold for these IODs, but with cooling eventually it goes below the threshold. So, when the SST is below the threshold for several weeks what happens we expect sustained suppression of convection over the EIO and you see that here this is the case of 97. Again this black line corresponds to SST of EIO red towers are the EIO convection and blue towers are the WEIO convection and what you see here is actually that the SST of EIO decreases rather rapidly here and then crosses the threshold around here and once it crosses the threshold then there is hardly any days with convective activity over EIO. So, you can see the kind of suppression sustained suppression that occurs when SST is below the threshold. So, cooling of EIO to SST below the threshold appears to be a very important attribute of strong positive IOD events. Now, you know ever since the papers of Sahajie et al and Webster et al in 1999 which first talked about the Indian Ocean Dipole mode there have been lots of studies of the relationship of IOD and ENSO. Now, there is one school of thought which feels that the Indian Ocean is passive and merely responds to the atmosphere and inter annual variability of the Indian Ocean SST are forced mainly by ENSO through an atmospheric bridge. So, what they believe is that there are lots of activities over the Pacific namely ENSO. This ENSO through the atmospheric bridge has an impact on the atmosphere over the Indian region and equatorial Indian Ocean and through that influences the Indian Ocean SST. So, it believes that most of the events of the Indian Ocean are actually remotely forced Indian Ocean is passively responding to that and the bridge is through walker circulation which as we know climatologically we have ascending motion over west specific and descending over ocean motion over east specific which is the walker circulation, but this circulation changes during El Nino and what we get is descending motion over west specific as well as eastern part of equatorial Indian Ocean. Now, about 4 months after large scale SST anomalies are observed in the equatorial Pacific the whole of the Indian Ocean is covered with SST anomalies of the same sign. This we have seen in fact that when we looked at SST anomalies even of WEIO and EIO we could see in the plus plus or minus minus quadrants that is to say in the quadrant in which both the anomalies were positive that quadrant was dominated by El Nino years because there is general warming of the Indian Ocean during El Nino and cooling during La Nina. Now, however if that is the case then why do they think that IOD is forced by El Nino because the Indian Ocean warming associated with an El Nino is not based in wide at all stages of ENSO. In the early stages of development of an El Nino there are cold SST anomalies in the eastern equatorial Indian Ocean, but they generally disappear as El Nino matures. So, as the El Nino is evolving the one important feature is suppression of convection over eastern equatorial Indian Ocean this is what they talk of the Walker circulation being anomalous and also cold SST anomalies of EIO. Now, people after the IOD papers came out people looked at simultaneous correlations between SST of the eastern equatorial Indian Ocean and ENSO indices and these correlations tend to suggest that IOD events are forced by ENSO that there is a correlation between the two. And in fact, the strong IOD event of the recent types that occurred during 1997 coincided with a very strong El Nino. However, it is important to note that IOD events have occurred together with ENSO as well as independently and in fact, according to the second school an important role is played by the coupled interactions of the atmosphere Indian Ocean system in the IOD mode. So, the fact that the Pacific will could and does have an influence on the Indian Ocean is not denied, but the differences of the two schools are in what how much importance to give to the role of the coupled Indian Ocean atmosphere system and the second school believes that that plays a very important role. So, its Indian Ocean is not merely passive, but is an active player in this in the generation of an IOD mode. This is what the second school believes and in fact, the roots of this second school and in the very first study of the inter-annual variations of convection and SSD in the equatorial Indian Ocean which revealed the IOD mode. However, it was then not called an IOD mode by the authors. This was done by Reverend and Etel from France and they are the ones who actually revealed the structure of the IOD mode and this they did by studying on the basis of ship reports available the IOD event of 1961. This is the picture of DMI now that you have seen before, DMI positive means positive IOD events and you can see here this is a major IOD event of 61 after which one occurred here and another occurred in 82 and then bigger occurred in 94, but the one in 61 is comparable to that of 94 and then 97 which was a very large IOD event. So, this IOD event of 61 is what revered in another study and they suggested that coupled air sea dynamics over the Indian Ocean should be considered in order to understand the inter-annual variability. And from limited ship observations, they concluded that the SSD anomaly affects cloudiness rainfall and consequently causes westward wind anomalies along the equator. See, this is the key element of positive equino and positive phase of the dipole inverse the SSD anomalies. So, they actually talked about the coupled Indian Ocean atmosphere processes being important to the IOD event of 61. Now subsequently there has been a lot of debate on this and so discovery of the IOD and studies that followed demonstrate that the Indian Ocean can sustain its own intrinsic coupled ocean atmosphere processes. So, now it appears that we can say that it is true that the Indian Ocean is influenced by the Pacific, but it can also sustain its own coupled ocean atmosphere processes and about 50 percent of the IOD events in the past 100 years have occurred with ENSO and the other half independently. So, it is not as if all the events are associated with ENSO. Actually coupled models have been able to reproduce IOD events and process experiments by such models like switching ENSO on and off support the hypothesis based on observations that IOD events develop either in the presence or absence of ENSO. So, it is not necessary to have an El Nino developing in the Pacific for a positive IOD event to develop over the Indian Ocean. Now, there is a general consensus among different coupled models as well as analysis of data that the P IOD events that is to say positive IOD events co-occurring during El Nino are forced by a zonal shift in the descending branch over the eastern Indian Ocean. This is what I mentioned earlier that the major distinguishing attribute of El Nino is that the ascending limb of the Hadley cell shifts towards the central Pacific and the anomalous vertical velocity over west Pacific is actually descending and this is it descends not only over west Pacific, but up to the eastern equatorial Indian Ocean. So, this is now the after a lot of debate it appears that if one looks objectively at what are the data available and the modeling done so far then one has to concede that El Nino does have an impact La Nina does have an impact all this is very well known since the 80s, but Indian Ocean by itself also can generate IOD events or important variability of the coupled ocean atmosphere system this is the consensus now. Now, let us look at evolution of P IOD events so consider the P IOD events in the satellite era and which is here and what we see here again are the major events in the satellite era are 82 and then we had some here 87, 92, 94, 97 and another in 2006. Another interesting case is actually 2003 where we thought an IOD event is going to develop, but it got aborted. So, we have the major IOD events are May to November 82, March to December 94 and July to December 97. So, May to November 82, then 94 and 97 these are the major IOD events where DMI is consistently larger than one for several months these are the major IOD events. Now, during the strong P IOD events of 94 and 97 very large values of DMI occurred for several months. In 2003 and I mentioned this before a P IOD event was triggered DMI was positive and large from June to August, but midway through the season the event was terminated that is the case of 2003 and you see here midway through the season this event simply got terminated. A P IOD event occurred during the latter period part of the summer monsoon season and autumn of 2006 with large values of DMI from September 2006 onwards and that is this event here it sort of developed a little late. Now, positive IOD events are relatively rare frequency of occurrence is less than 1 in 5 years and it is believed that they are triggered by some event during April May. Important facets of the P IOD are now when we talk of triggering we have to say see what are the important facets of these events and the major facet of the IOD event is of course, negative SST anomaly of EE IO which we know now occurs due to upwelling in response to winds parallel to the Sumatra coast and positive oil or anomaly that is to say suppression of convection over EE IO. So, these are two very important features and we have to see how these features evolve. So, the important facets of triggering are suppression of convection over EE IO and strengthening of the winds parallel to the Sumatra coast which will lead to upwelling. Note that since in late spring W E I O is also favorable for convection this is what we have shown in the early part of today's lecture that from May onwards up to August both EE IO and W E I O are almost equally favorable for convection. So, what happens when we have suppression of convection over EE IO immediately you will get enhancement of convection over W E I O because there is a negative correlation between convection over EE IO and convection over W E I O. So, suppression over one will lead to enhancement of convection over the other thus triggering of a P IOD event necessarily involves triggering of a positive phase of equino. So, if we say we say of an event is triggered when the convection over EE IO is suppressed and those winds parallel to Sumatra are strengthened then as soon as the convection over EE IO suppressed necessarily convection over W E I O will get enhanced. This means that we get a positive phase of equino because positive phase of equino is characterized by having enhanced convection over the west and suppressed convection over the east. So, triggering of a P IOD event necessarily involves triggering of a positive phase of equino. Should be noted that positive phase of equino is triggered whenever the convection over EE IO suppressed irrespective of whether it is accompanied by enhancement of upwelling favorable winds that is irrespective of whether a P IOD is triggered or not. So, this is something to remember when we look at variation of convection over east and west that irrespective of what happens to the ocean if by some means we ensure that the convection over EE IO suppressed then convection over W E I O will be enhanced this is during May to August and this means that a positive phase of equino will be triggered. Now important phases of triggering which is suppression of convection over EE IO and enhancement of the upwelling favorable winds parallel to the coast of Sumatra have been attributed to different factors by different scientists. The among the many theories proposed the leading ones are change in the Pacific walker circulation which I talked about the intensification of Hadley cell in the western Pacific between the South China Sea and Philippine Sea. So, people have proposed that this intensification of the Hadley cell or intensification if you wish of the tropical conversion zone over South China Sea and Philippines will lead to triggering of positive I O D event this is one of the theories proposed and there is also a theory which is somewhat different from the rest in which it is suggested that severe cyclones over the Bay of Bengal during April May can trigger an I O D event. Now the proponents of the first school Annamalai et al incidentally they are also proponents of the theory that by and large Indian ocean is a passive entity and obeys whatever orders are given from the Pacific. Now, so Annamalai et al proposed that an El Nino can trigger a positive I O D since El Nino does lead to a suppression of convection over EE IO this I have mentioned already. Now, but then what about strengthening of the winds? So, they suggest that this suppression of convection forces an anti cyclonic circulation over the southeastern Indian ocean. Now, one has to remember that in the southern hemisphere anti cyclonic is also anticlockwise. So, if you have an anticlockwise circulation which is westward of Sumatra then the the circulation along Sumatra coast which corresponds to the limb from 6 o clock to through 3 o clock to 12 o clock in this anticlockwise direction is actually in the right direction for causing upwelling. So, we have winds which are south coming from the southeast or coming from the south parallel to the Sumatra coast. So, Annamalai et al proposed that actually this suppression of convection over EE IO associated with an El Nino leads to a circulation anti cyclonic circulation in near EE IO and to its west such that we have strengthening of the winds which lead to upwelling for the strengthening of the upwelling of Sumatra coast. However, there are some problems with this theory their experiments with an atmospheric general circulation model showed rather weak response of the EE IO rainfall to west central Pacific SST anomaly. See they say that the El Nino which is characterized by positive SST anomaly over west central Pacific is the one that leads to suppression of convection and rainfall over EE IO, but their atmospheric model did not show a strong response over EE IO at all to west central Pacific SST anomaly. So, no attempt was made to demonstrate that SST and thermocol and depth anomalies evolve into conditions favorable for the IOD. So, that part also was not convincingly demonstrated. One should also remember that mechanisms which involve and so in the initiation of IOD events obviously do not operate in cases like 61 remember 61 is the case that revered in studied and which first revealed the IOD and 94 in neither of the cases did we have an El Nino over the Pacific. Secondly the mode in 1967 IOD event in 1967 occurred in the presence of a La Nina. So, La Nina will have the opposite sign of anomalies in the walker circulation. So, it is not possible to explain all the observed IOD events if we insist that they are generated by and so in the Pacific. Now, there was another theory proposed by Japanese Kazikawa Etan which suggested that somehow the convection over south China Sea which corresponds to intensification of Hadley cell over the western Pacific can enhance the south easterly in EEO, SST cooling and suppression of convection. So, they are suggesting that if you have a lot of convection over side China Sea then one can get this kind of a response and which remember these are the factors that lead to triggering of an IOD. Suppression of convection over EEO and enhancement of the south easterly winds of the coast of Sumatra. So, they claim that this can then lead to the triggering of an IOD. But the intensification that they are talking of takes place during the summer this is during June and July and can be either due to Enso or Monsoon this is what they say. However, now it is well established that the wind anomalies associated with the IODs appear during spring which the above mechanism fails to explain. See they have suggested a mechanism which could have worked in terms of suppressing the convection over EEO, but the timing is wrong if we compare with the observations. Now, so far people were talking about Enso and its relationship to IOD and what happens to Hadley cell over west Pacific and its relationship to IOD in quite a deviation from all other theories Francis et al in 2007 proposed a very different theory. They said severe cyclones over the Bay of Bengal during spring season can trigger IOD events. In fact, they showed that all positive IOD dipole events during 58 to 2003 are preceded by at least one such severe cyclone. Now, severe cyclones over the Bay of Bengal strengthen the meridional pressure gradient. Why is that? Because severe cyclones imply that the pressure is very low over the Bay of Bengal relative to the equatorial Indian Ocean. This means that the north-south pressure gradient has enhanced and that would lead to an intensification of upwelling favourable south easterlies along the Sumatra coast. They also showed that severe cyclones can lead to a decrease in the integrated water vapor content over EEO and suppress the convection there. So, they actually generate conditions which are favourable for triggering. Now, I do not have the time to get into details about this, but I would just like to mention the jury is still out on what triggers the IOD events, but what would be interesting to look at is a mechanism which Francis et al. proposed to explain how once the convection over EEO is suppressed, the positive equino phase can be sustained for several days even after the end of the event that led to the original suppression of such as the severe cyclone in the Bay. So, this is an interesting problem in itself that you have an event on the synoptic scale. It could be intensification of convection over South China Sea as proposed by Kazikawa and others or it could be a severe cyclone in the Bay. Now, if this event has the kind of impact that has been suggested that is suppressing convection over EEO and strengthening the south easterlies parallel to Sumatra coast, it should all end once the event ends because these events are only few days. But for the triggering to take place for the IOD event to be triggered, this has to be sustained for much longer than the synoptic scale at least for timescale of the order of weeks. Question is how does that happen? Now, for this Francis et al. proposed the mechanism which is shown in this slide here schematically. So, initially state is the following. Initially we have convection over EEO and where is this is all this is the entire warm sea surface remember west is also warm east is also warm and convection is over EEO because that is where the propensity of convection is high. So, we have strong westerly winds here and what happens? We get a severe cyclone here over the Bay. Once we get a severe cyclone over the Bay, then it will enhance south easterly flow over the Sumatra parallel to the Sumatra coast and it will also suppress convection over here. Once it suppresses convection here the pressure gradient here decreases. So, the flow becomes weaker westerly. So, the equatorial winds become weaker westerlies equine is reduced the magnitude. So, once we have a severe cyclone this convection disappears and we have enhancement of upwelling favorable winds and weakening of westerly winds. Now, once this convection is suppressed because this region is equally favorable and there is a seesaw in convection between these two regions which you have seen. Once this cyclone forces this convection over EEO to be suppressed immediately convection will build up over WEIO. So, now, we get a situation in which the convection builds up over WEIO and this means that you have now the pressure gradient is switched it is the low pressure and the equator is here. So, the winds also switch if it is in terms of anomalies we get easterly anomalies here and eventually actually in the development of an IOD you get actually easterly is here when there is sustained convection here. So, the picture switches here and what happens when you get convection here? What happens when you get convection here is that a positive feedback is set up because now if we talk of anomalies this anomalous convection drives winds towards it which are winds which are easterly anomalies. Now, these are this means there is more convergence into this region more convergence means more convection and more convection means more convergence. So, once the system gets set up here it is a self propagating it is a it is a system which with a positive feedback which can lead to sustained increase in convergence and sustained increase in convection here and which is what happens when this happens actually what happens is instead of just having easterly anomalies you get easterly winds here. So, strong convergence contributed by easterly winds from here. Now, once the winds become easterly here the SST of EEO drops even further because not only do we have upwelling favorable winds parallel to Sumatra we also have like the Pacific you know surface water being driven away from this region which will lead to equatorial upwelling as well. So, this is very analogous to Pacific. So, if this process continues then we can get triggering of an IOD event and remember that all this mechanism asked for is some event it could be of a synoptic scale event also need not be of a larger scale than synoptic intracisional scale some event which lead to suppression of the convection over EEO is all that is required and then all these feedbacks will take place and we can go over to a system in which this develops a cold anomaly of SST along with the suppressed convection and if the process goes on actually the SST can go below the threshold and once it goes below the threshold there is no way the convection can revive here no matter what and we get a IOD event. So, they suggest that since WEIO and EEO are almost equally favorable for convection in April and May once an event such as the cyclone over the bay leads to the suppression of convection over EEO convection come and says over WEIO this leads to the weakening of westerly surface wind over CEIO and hence enhancement of convergence over to the WEIO this positive feedback can lead to further enhancement of convection over the WEIO this is what we just explained. So, this continues until the wind over the CEIO becomes easterly and the convection over the EEO remains suppressed for periods much larger than a synoptic time scale characterizing the cyclone. The strong upwelling caused by the easterlies along the equator and the south easterlies along the Sumatra coast can decrease the SST of EEO very rapidly provided the thermocline is shallow and a positive dipole event is triggered. So, we now have a story of how one could get trigger positive IOD events and why is a positive IOD event of greater interest because a negative IOD event is intensification of climatology. Whereas a positive IOD event involves reversal of the east-west gradients of sea surface temperature and convection over the Indian Ocean. Now, it is important to note that all these processes of triggering can lead to cooling of EEO only when the thermocline there is shallow because if the thermocline is not shallow the up welled water is not cold enough. So, you do not get as much cooling as you would when the thermocline is shallow. So, like the Pacific this is a constraint the thermocline has to be shallow for all this ocean dynamics mechanism to operate and lead to in fact an IOD event. Now, what you know we have seen in earlier lecture that the thermocline over the eastern part is deeper than the thermocline over western part in the equatorial Indian Ocean. Now, what leads to thermocline being shallow in April May over the eastern part is something I do not have time to get into in this lecture series. And you know this is what sets the stage for the possible development of IOD and Kelvin waves and Rossby waves and so on play a very important role up welling and down welling waves play a very important role in the changes of the thermocline along the east-west direction in the equatorial Indian Ocean. As I said I will not have time to go into it, but let me just show you how important preconditioning for the ocean is. You remember that we had two cases 97 and 2003. In fact, Francis et al have shown that the initial evolution of the OLR, SST and so on of EEO were very similar in the two cases, but in the case of 97 we got a full fledged IOD event developing with SST going below the threshold, but in the case of 2003 SST never went below the threshold and the event got aborted halfway through the season. And we can see here how it happened. This is in fact the thermocline depths in 97 and 2003 as well as climatology. This is from soda data which is analogous in the ocean to the encephry analysis data which we have been using. See this is climatology now and these are the months here you can see April to November so this is only for EEO. We begin with a very deep thermocline this is climatology. We begin with a very deep and warm thermocline in April and slowly as we have seen the SST cools and we have shallowing of the thermocline so that it becomes rather shallow in September, October and then starts deepening again. So, this is climatology the mean of all the years. Now what happened in 97 is here and it is a huge contrast. What happened in 97 is to begin with you can see that it was shallower even in April and June onwards you see it has become extremely shallow. So, that the thermocline here has become very very shallow now remember these are temperature contours this is 30 sorry this is 30 this is 29.5 29 and so on and so forth. So, the warm water is over a much smaller region in 97 even to begin with and then this 27.5 which is a threshold it actually surfaces around August. So, after August EEO is colder than the threshold. So, this is what is happening in 97 and you can see that initially itself the ocean was preconditioned with less heat content and a shallower thermocline. Look at the case of 2003. 2003 actually the thermocline was shallower to begin with all the processes of triggering and so on occurred in 2003 as well and the thermocline began to go shallow just like this. But the point is it started from a much deeper level and it never really got too shallow. So, that the water never became colder than the threshold and when for some reason the west convection disappeared the east revived and IOD event was about it. Now, I have mentioned at the end here several of the important references there are these are the revered in is the one who first described it Saaji et al and Webster et al 99 the ones who first talked about the Indian Ocean Dipole mode and Indian Ocean Zonal mode and then there are nice review articles one by Vinay Chandran et al one by short et al and some critical papers like the one on 94 by Vinay Chandran then that triggering papers that I mentioned at the end. So, with this we come to the end of a discussion on IOD it is a very hot topic and by no means have I been able to cover all the issues involved. But I think this should provide some guidance as to where to go from here. Thank you.