 Today, I am going to talk about the year to year or the inter-annual variation of the Indian summer monsoon rainfall, particular focus on links of this variation to events over the Pacific and equatorial Indian Ocean. We know that the Indian summer monsoon rainfall or ISMR, the long term average is about 85 centimeters and the inter-annual variation we is about inter-annual variation the standard deviation of that is about 10 percent of this average and what we see here is the inter-annual variation of the anomaly of ISMR which is to say the actual ISMR in any year minus the average ISMR long term mean of ISMR expressed as a percentage of the mean during it is 76 to 20 10 for which data are available. Now, this is something we have seen before and as you can recall we define extremes as those years in which the anomaly of ISMR has magnitude more than 10 percent or more than one standard deviation. So, if there is a deficit larger than one standard deviation or more than 10 percent it is a drought if the ISMR is above average and the anomaly is greater in magnitude than 10 percent then it is a excess rainfall season. So, there are extremes which are excess rainfall seasons and droughts and they are droughts are marked as red excess rainfall seasons as green and this is the variation we have seen. We have also seen that there is a decadal scale variation here with high frequency of droughts in this seaport and again here and relatively fewer droughts in this period here we have seen this as well. Now, as I mentioned before the standard deviation of the inter-annual variation of ISMR is about 10 percent and summer monsoon seasons with ISMR smaller than 90 percent of the mean are considered to be drought and those with ISMR larger than 110 percent of the mean as excess rainfall seasons. Now, Indian therefore, since the standard deviation is only 10 percent of the mean it is a very reliable phenomena because the maximum observed deficit is less than 30 percent of the mean. So, it comes year after year and there is a variation in the quantum of rainfall received, but it is not that great, but even though the amplitude of the inter-annual variation is not very large it has a very large impact on the food grain production and GDP and so the monsoon continues to govern the very pulse of life in our country. If it turns out to be a normal monsoon as in 2010 the nation heaves a sigh of relief and carries on with business as usual. If it turns out to be a drought as in 2009 major drought relief programs are launched thus the importance of prediction of ISMR cannot be overemphasized. So, although the inter-annual variation is not large the amplitude is not large it is very important to predict it. Now, a major advance in our understanding of the inter-annual variation of the monsoon occurred in the 80s with the discovery or rather rediscovery of a strong link with the major phenomena over the Pacific known as El Nino and Southern Oscillation Enso. This was in fact this advance came about by work done in our own country Sikha 1980 and Parsa 81 followed by Rasmussen and Carpenter from US have shown that there is an increased propensity of droughts during El Nino. El Nino is the warm phase of Enso and the cold phase of Enso there is a high propensity of excess rainfall cold phase being called La Nina. Now, we use an Enso index defined as Enso index which is negative of the SST anomaly of Nino 3.4 normalized by the standard deviation. Now, why the negative sign? So, that positive values of Enso index which means negative SST anomaly is over central Pacific indicate the occurrence of a La Nina that is a phase of Enso that is favorable for the monsoon. So, positive values of index of favorable for the monsoon. Now, what is the nature of impact of Enso on the monsoon? The easiest way to see it is actually by looking at the correlation of Enso index which we have defined with OLR over the entire tropical region here. So, what you find is over the Indo-Pacific region is shown here what you find is of course this is the signal of Enso and this means this says that the correlation of OLR is positive here which means this is representing if you wish a La Nina condition and this implies that everywhere here the sign of the OLR anomalies is opposite to what it is over the central Pacific and so should we just pause for a minute I will just switch off this mobile. So, what this is saying is that when there is an El Nino that is when the OLR anomalies here are negative in that case OLR anomalies everywhere here will be positive which means that when convection is enhanced here it is going to be suppressed over the Indian region and also suppressed over the entire Arabian Sea most of Bengal as well as the equatorial Indian Ocean. So, impact of El Nino which as you recall Sikha had shown that in fact there is a increased propensity of droughts during El Nino and so when you have an El Nino you do expect that if the convection flares up here it will be suppressed over India which is what Sikha and others had shown, but what this shows that in addition not only is the convection suppressed here, but it is also suppressed very much over the eastern equatorial Indian Ocean also suppressed over western equatorial Indian Ocean and Arabian Sea and much of Bay of Bengal. So, this is the impact of El Nino on the Indian region. So, what are the implications for the CTCZ and the Monsoon consider the implication of the suppression of the convection over the Arabian Sea Bay of Bengal and the equatorial Indian Ocean in association with El Nino. Now, we have already seen that the uncertain retreat phase of the Monsoon involve northward propagation. So, you remember this is an envelope of the low LR region or the convection region at 90 degrees east and from March till November and what you see is that there is a series of northward propagation all starting from the equatorial Indian Ocean and culminating somewhere near the red line which is the envelope and this is the onset phase of the Monsoon. Ultimately, this is the established phase July August and this is the retreat phase and we have noted that both the onset and the retreat involve genesis of bands over the equatorial Indian Ocean and northward propagation. So, spring to summer transition which is the northward shift of the TCC from its mean location near 5 degree north in April May to 20 degrees north in July August is accomplished by successive generation of northward moving epochs over the equatorial region. Summer to autumn transition again as we noticed here even during summer to autumn when on the monthly scale you see a southward movement of the envelope it is accomplished again by northward propagation of bands generated over the equatorial Indian Ocean. So, genesis of the equatorial Indian Ocean and propagation northward contributes a great deal to both the spring to summer transition as well as summer to autumn transition. So, summer to autumn transition also comprises northward moving epochs just like spring to summer transition and not southward moving epochs as we would have thought. So, there are the this is the onset this is 75 this is at 70 degrees over the Arabian Sea and a part of India 80 degrees central longitude of India 90 degrees is across the bay going partly through Calcutta. So, what we see is that very coherent northward propagations which occur during the transitions here, but which also are seen in July August. So, these northward propagations are a prominent feature throughout the monsoon season this is the same picture for 2007 this is again with OLR and what you see is very prominent northward propagation which are coherent across 70, 80 and 90 and several of them occurring through the monsoon season from April to November shown here. So, now we have seen that in the transition seasons these northward propagations are very important. Sikha Kathir also showed that the convective days associated with the CTCZ over the monsoon zone in July August are dominated by epochs generated over the equatorial Indian Ocean which move northward. So, they have quantified the contribution of these northward moving epochs to the maintenance of the CTCZ over the Indian region and show it to be a very large fraction. So, the contribution of northward moving epochs of the oceanic TCZ to the maintenance of the CTCZ is substantive. Now, northward propagations of the TCZ are generated over the equatorial Indian Ocean and generally involve northward propagations of the synoptic scale systems over the Arabian Sea and Bay of Bengal. Therefore, we expect the variability of the CTCZ on different scales to be linked to the variability of convection over the equatorial Indian Ocean Bay of Bengal as well as Arabian Sea. So, the surrounding seas are all going to be involved in determining the variability of the band over India the CTCZ. Now, it must be borne in mind. So far we are saying that these oceanic TCZ is very good because it contributes to the maintenance of CTCZ via northward propagations. But you know the relationship is complex as we have noted before and while the oceanic TCZ contributes towards maintaining the CTCZ with northward propagation it also competes with the CTCZ because you have two tropical convergence zone over the same longitudinal belt each can suppress the other and therefore, generally weak spells of one coincide with active spells of the other. So, there is not only contribution of oceanic TCZ to maintaining CTCZ, oceanic TCZ also competes with the CTCZ. Now, so we have seen that as far as the oceanic TCZ is concerned and as far as convection over Bay and Arabian Sea are concerned which are all involved in northward propagation certainly the variability of convection over these regions will be related to the variability of CTCZ will contribute to the variation of CTCZ. We have also seen that apart from these northward propagation CTCZ is maintained by synoptic scale disturbances generated over the Bay which move westward on to the Indian region. And an example is what we have seen before 1999 when we saw so many systems lows depressions and so on actually get generated over the Bay and move across here. So, this is a typical example of year in which you had several systems move across and this is of course, the monsoon zone and this is where the CTCZ is. So, this maintenance of the CTCZ by these systems is very clear from this. In fact, the tracks of all cyclonic storms for July this is over a very large period almost 100 years and what it shows is that in fact, all the tracks are along the monsoon zone along where the rainfall is non-horographic rainfall is. So, these synoptic scale systems also contribute to the monsoon rainfall. So, the variation of monsoon rainfall will naturally depend also on convection over the head bay region particularly. So, convection over Bay of Bengal, Arabian Sea as well as equatorial Indian Ocean all plays a very important role in maintaining the monsoon rainfall over the Indian subcontinent. So, it is not surprising that the suppression of convection over these oceanic regions during El Nino is associated with deficit rainfall over the Indian region. So, now you see if we plot the normalized isomar anomaly this is isomar anomaly divided by the standard deviation and plot it against the ENSO index. Then when ENSO index is negative remember this is the El Nino condition and this is when there is a high propensity of droughts. So, negative isomar anomaly means deficit isomar larger than one in magnitude means these are all droughts all these red dots represent droughts. Then of course, ENSO index less than minus 1 represent means El Nino, but in general negative ENSO index means unfavorable ENSO. Positive ENSO index means favorable ENSO and you can see that many blue dots are in the positive ENSO domain. So, you have a higher propensity of excess rainfall years excess rainfall seasons for which the isomar anomaly is positive and larger than one in the positive ENSO phase and particularly during El Nino. In fact, one can using this graph make a one sided prediction because you know now it is becoming increasingly possible to predict ENSO because a lot of the ENSO physics have been understood in the 80s and 90s. So, models have become pretty good in predicting ENSO. So, if you could predict the ENSO index and if you knew that it was less than minus 0.6 that is this line here then there is no chance at all of excess monsoon years. You see there are no blue dots if ENSO is less than minus 6 and there are no red dots if ENSO is greater than plus 8. So, it is minus 8 sorry this is minus 8 and plus 6 and so there are many most of the droughts are in this region most of the excess are here, but there are several droughts in the intermediate region also. So, we can make one sided prediction saying if ENSO is in this range they will not be droughts and if ENSO is in this range here up to here they would not be excess rainfall years, but there will be a lot of years here in between where you will not be able to say anything. And just wanted to mention that 82 and 87 were very famous El Nino years and as expected their droughts 88 was a famous La Nina year and it is an excess rainfall year, but note 97 and 94 we will keep coming back to these. These are two years in which ENSO phase was favourable, but still 90 was unfavourable because remember it is a negative year, but 94 was still an excess monsoon year and 97 was normal although it was the strongest El Nino of the century. So, we will come back to these intriguing features now. So, I have said that it is seen that when the ENSO index is favourable there are no droughts and when it is unfavourable there are no excess monsoon seasons. The variation of ISMR for the favourable and unfavourable and other phases shown in the next table is consistent with this. So, this is something computed by Rup Kumar et al for 1871 to 2001. So, these are the number of years with El Nino, La Nina and other years. This is deficient monsoon, normal monsoon positive and excess. So, deficient monsoon is our droughts and you see that of the total number of 26 droughts that occur 11 were El Nino years, but of the total number of droughts that occurred which were 22 years only 11 were El Nino 11 were some other phenomena working. So, ENSO alone cannot explain everything La Nina again there are no droughts associated with La Nina and there are 8 excess monsoon years associated with La Nina, but there are other 11 years which were also excess monsoon years. So, half of a droughts are not associated with El Nino and more than half of La Nina's sorry than half of excess monsoon seasons are not associated with La Nina. So, there is definitely a relationship between Monsoon and ENSO, but it is by no means a one to one relationship. Now, since one of the major achievements of the past 2 decades as I mentioned before is the ability to predict ENSO, it is possible to use this relationship for inputs into the prediction of ISMR. Now, for this we consider 5 categories of ISMR such that they are all equally probable. So, we take the data on ISMR and divide it into 5 categories such that each is likely 20 percent of the time. Now, taking this observed variation between 58 and 2009. So, what are the 5 categories these limits are determined by the data itself data on ISMR the time series of ISMR. So, large deficit is less than 75 centimeter deficit is 75 to 83 normal is 83 to 86 above normal is 86 to 91 and above 91 is large excess or excess years. So, these are 5 categories which we have divided the ISMR into because climatologically they are equally likely each is likely 20 percent of the time. So, chance of occurrence of rainfall in each of these categories how does this probability distribution which is flat right each category chances 20 percent how does it change if we put in some information of ENSO index and the kind of information we have put in of ENSO index is rather crude will ENSO index be negative or positive that is all we are asking. So, blue corresponds to ENSO index negative I am sorry positive and negative ENSO index is red and what you find is remember 20 percent is the climatological probability that is to say if we had no information other than the historical data set then we would say all these categories are equally likely. Now, what happens when we put in the fact that LNO is likely to be negative then immediately the chance of drought has increased from 20 percent to 35 percent very substantive increase and chance of excess rain has decreased from 20 percent to less than 5 percent. So, the extremes have been very much affected although there will be some it is not as if there is a zero chance of excess there is some chance of excess even during LNO years it is very much smaller than climatological probability and the chance of drought is very much larger than climatological probability go to the other extreme. If the ENSO index is favorable then again the chance of excess has increased beyond 20 by more than 10 percentage points 12 or so. So, again there is an enhancement of probability of excess rain and there is a substantive decrease in the probability of drought from 20 percent. So, even if one had some reliable predictions of the sign of the ENSO index during June July August September we could actually give definite predictions on the probabilities of extremes, but note that as far as non extremes are concerned you cannot really say very much because in this case for example, you know the probability is just 20 percent of excess when LNO is positive and it is slightly reduced when LNO ENSO is negative and same thing here see the difference between positive and negative is not spectacular for the intermediate cases. So, important take home lesson is that if one can predict ENSO you will be able to say something about how the probabilities of the extremes are going to change from the climatological probability, but you will not be able to say much about what happens to years that are in between the three categories in between this is very interesting. So, note that the probability is substantially different from climatological one for the extremes and we just noted that the probability of a large deficit decreases from 20 to 7 percent and increases to over 35 percent for appropriate signs of the ENSO index. So, the probability of large deficit will decrease when ENSO is favorable to 7 percent and increase to 35 percent when ENSO is unfavorable and similarly for the excess that the probability of large excess will increase to 30 percent from 20 percent when ENSO is favorable and decrease to 4 percent when it is unfavorable. So, there is a very clear signal we can get from ENSO. Now, let us consider the inter annual variation of the monsoon season since 1980 and look at it more closely. So, this is the inter annual variation from 1980 and these were the 2 years I pointed out before 82 and 87 and they are in fact, El Nino years and 80 T 88 is a La Nina year, but I also pointed out the excess of 94 and the normal of 97 and what we see here is the link of the inter annual variation of the Indian monsoon with ENSO. So, the first one the first bar tells you what the ENSO index is like. So, negative would mean I am sorry the first bar tells you the ISMR the second the ENSO. So, you have in this case 82 and 87 this is 82 and 87 both are this is 82 and you have very unfavorable ENSO in 82 and 87 here, but and you also have in 88 very favorable ENSO here. So, this is the favorable ENSO giving rise to excess rainfall here, but then you have a case in which 1997 where the El Nino was the strongest El Nino of the century. So, ENSO was highly unfavorable we still got very near normal monsoon and in 94 El Nino was not that unfavorable, but it was unfavorable it was a weak El Nino and still we got an excess monsoon year. So, these are odd things that we have to understand they are years in which they do not seem to go with ENSO, but there are several years that do go with ENSO like 82 87 and 88. Now, what happened after 88 is that for 14 consecutive years beginning with 1988 there were no droughts and you can see it here. See beginning with 88 there were 14 years from here to here where there were no droughts at all and this despite the fact that there were weak El Nino occurring, but there were no droughts at all. So, further more during the strongest El Nino event of the century in 97 ISMR was higher than the long term mean which we have noted and Krishna Kumar and others suggested that the link between Indian monsoon and ENSO has weakened in the recent decade. So, they just suggested that look monsoon is no longer responding to El Nino as it was before and people took their results seriously, but then came a relatively mild El Nino of 2002 which turned out to be a very severe drought and we can see that here see this is the El Nino of 97 and you can see here how unfavorable the ENSO signal was compared to that the yellow line here is not so unfavorable and yet while this monsoon was above normal little bit this monsoon turned out to be a severe drought more than deficit was more than to standard deviation more than 20 percent of the mean. So, it was the most severe drought for the period that we have plotted here and so these were the till 97 people thought the link between El Nino and the monsoon seems to have weakened, but then came this drought of 2002 which occurred in association with a much weaker El Nino than that of 97 and neither the statistical nor the dynamical models could predicted. So, this suggested the experience of 97 which was the strongest El Nino of the century, but which had above normal rainfall and 2002 which with a relatively weak El Nino turned out to be a very severe drought one began to wonder whether one understands completely the link between ENSO and monsoon given 2002 we could not say that the monsoon ENSO link has weakened. So, it was an intriguing point what is happening to El Nino what is happening to the monsoon ENSO link. In fact what was happening becomes very clear because while ENSO is important it was found that another mode is also important this mode is equino and so the intriguing monsoon seasons of 97 and 2002 triggered studies which suggested a link to events over the equatorial Indian ocean. So, not only is the monsoon related to events over the Pacific it is also related to events over the equatorial Indian ocean. And this Gardgill showed et al in 2003 and 2004 showed that in addition to ENSO the phase of the equatorial Indian ocean oscillation or equino makes a significant contribution to the inter annual variation of ISMR. Now, how did we find this mode our first encounter with this mode of convection over the equatorial Indian ocean was while anxiously observing the satellite pictures in 2003 after the intriguing season of 2002. You know the 2002 season was so intriguing that we had a meeting at the request of the DJ then director general of a metrology at Bangalore to try and understand what was the special feature of 2002 that led to this severe drought and why could no model predicted. So, when 2003 monsoon commenced we were all worried about what is going to happen in 2003 and we were watching the satellite pictures as we saw the satellite pictures then this mode became very evident and I will show you why. Now, it turned out that whereas July 2002 was a huge deficit almost 50 percent see in particular whereas there was an unprecedented deficit 49 percent in all India rainfall in July 2002 in July 2003 there was excess of 7 percent and if we compare the two OLR anomaly pattern this is July 2002. This means convection is highly suppressed over Indian region as well as here and convection is active here it is enhanced here whereas July 2003 what you see is the opposite over here convection is enhanced over the western equatorial Indian ocean suppressed over the eastern equatorial Indian ocean. Here convection is enhanced over the eastern equatorial Indian ocean and suppressed here and notice also and we will come to this point again that the OLR anomalies over the Indian region here these purple ones tend to be of the same sign as the OLR anomalies over the western equatorial Indian ocean whereas they tend to be of opposite sign to OLR anomalies over the eastern equatorial Indian ocean. This is interesting because you may recall I had said that the relationship between CTCG and the tropical convergence zone in the equatorial region is bit complex on the one hand it supports the CTCG on the other hand it competes with it now it is very clear from this kind of pictures that convection over western equatorial Indian ocean is the one that supports the CTCG it is favorable for CTCG whereas convection over the eastern equatorial Indian ocean for some reason is the one that takes charge of the competition with CTCG. So, the complexity is resolved in a way with division of labor the convection over the western equatorial Indian ocean taking care of maintaining the CTCG it is positively correlated with the Indian region rainfall or convection and the one over the eastern equatorial Indian ocean is negatively correlated with the one over India and so this is the this convection over this region is the one who has taken charge of the competition this is the one that promotes the monsoon. So, what is the equatorial Indian ocean oscillation now we have seen here that when convection is intense here it tends to be suppressed here when it is intense here it tends to be suppressed here. So, there is a seesaw in convection between the western equatorial Indian ocean and eastern equatorial Indian ocean these are the boxes by which we denote denote these and I will call them W E I O western equatorial Indian ocean and E E I O hence forth and these are the longitudes and latitudes of the boxes we have drawn. So, what we are saying is that there is a tendency of seesaw between convection over these two regions if this is enhanced this is suppressed if this is enhanced this is suppressed that is what seems to be happening. So, suppression of convection over the eastern equatorial Indian ocean tends to be associated with enhancement over the western equatorial Indian ocean and vice versa. The oscillation of a state with enhanced convection over the western part and reduced convection over E E I O and this is one state with enhanced over W E I O and suppressed over E E I O and the opposite anomalies which means suppressed over W E I O and enhanced over E E I O these are two states and the oscillation between these two states is what we call the equatorial Indian ocean oscillation positive phase of the oscillation we have taken corresponds with enhanced convection on W E I O and you will see why because if you have enhanced convection on W E I O then that is favorable for the monsoon. So, this we consider as positive phase of equino and this is negative phase of equino. Now, associated with these changes in convection we will get changes in the sea level pressure gradient and in the zonal component of the surface wind over the central equatorial Indian ocean. So, these are in fact two again examples of opposite extremes and if you recall 94 was another intriguing case like 97 94 also although and so was unfavorable you can see that equino is highly favorable. So, this is the positive phase and this is our 2002 which is a negative phase of equino. So, as I mentioned OLR anomalies over the Indian region are of the same sign as those over W E I O and opposite to those over E I O thus a positive phase of the equino favours the monsoon while the negative phase is unfavorable. Now, positive phase of the equino is associated with easterly that is from east to west anomalies in the equatorial zonal wind whereas negative phase is associated with westerly anomalies. Let me just explain why it is very simple to see this is a positive phase. So, there is a lot of heating in this column. So, you will get lower pressure here than here. So, higher pressure here and lower pressure here and remember we are at the equator. So, at the equator like in non rotating fluids air will tend to move from high to low pressure there is no coriolis force acting on the equator. Therefore, we will get winds for the positive phase which are coming from the east to the west. So, these are easterly component whereas in the opposite phase they will be westerly component they will come from west to east. Now, we have used as an index of this equatorial Indian ocean oscillation a wind index and this is called equine and we take it as a negative of the anomaly of the surface zonal wind averaged over 60 to 90. So, our central equatorial Indian ocean if we take the zonal wind anomaly and then take the negative of that negative. So, that positive values of equine correspond to positive phase as we have defined. Then for this situation which we have seen before where the positive phase of the equine is very strong equine in fact turns out to be 1.47. So, this is a large positive value of equine corresponding to a strong positive phase of equino. That a positive phase of equino with enhanced convection over W E I O is favorable for the monsoon is clearly seen from the pattern of the correlation of ISMR with OLR and we have seen this before this is a correlation of ISMR with OLR everywhere. So, you can see here more convection means less OLR. So, the negative correlation represents positive correlation between rainfall and ISMR here or convection and ISMR and you can see that this and this are the same sign. So, if you have more convection here more rain here you will have more convection and rain here. So, it is clear that the positive phase of equino is favorable for the monsoon and this is the negative correlation corresponding to the negative phase. Note that this correlation is as strong as the negative correlation between here and here which corresponds to the Enso link. This is the central Pacific and the Indian region. So, it is as strong as the negative correlation. So, we have seen that convection over W E I O is clearly favorable for the monsoon and magnitude of the correlation of ISMR with the convection over W E I O is comparable to that with the convection over central Pacific corresponding to link with Enso. So, this link is equally important and we should not ignore the equino link. This is what this is telling us. So, we have to look at both Enso and equino. In fact, it turns out that for June to September the correlation between the Enso index and equino. Enso index is the index for the El Nino Southern Oscillation and equino is the index for equino which is the equatorial Indian Ocean Oscillation. So, the correlation between these two indices is not significant it is some 0.052 suggesting that they can be considered as independent modes. So, as far as this season is concerned June to September they can be considered as independent modes and therefore, we can use them as x axis and y axis ok. For any year we know the value of Enso index and we know the value of equino right. So, we can use one of them as the abscissa and the other is the ordinate x axis and y axis and on using those the location of any year is determined and we can actually plot it as a phase plane. But before we do that let us see how these look. So, the ISMR anomaly the Enso index and equino for all the extremes that occurred during 79 and 2010 and also the special season of 97 is what we will see here ok. So, the first bar is ISMR the second bar is equino and the third bar is Enso index ok. So, this is 2002 the most severe drought this is the rainfall and you can see that both of these were unfavorable both equino and Enso were unfavorable. Then comes 87 for which also they were unfavorable and so on and so forth 85 one was opposing the other it was not as severe a drought as the rest. Then comes 97 and you can see here that it is a tug of war there is a very strong negative unfavorable Enso, but a strong positive equino and in the tug of war we are getting a very near normal very near 0. In fact the ISMR anomaly is and we get to 88 where again both are favorable and then 2 years in between 94 a weak Elnino, but it was an excess and 83 was a case of a complex Elnino which I will come to later and again a favorable equino and we get an excess. So, it appears that each drought or excess rainfall season is associated with unfavorable. So, let us take it one at a time each drought is associated with unfavorable phase of either Enso or equino or both each excess rainfall season is associated with favorable phase of either Enso or equino or both. Now, we come back to this intriguing years of July of 97 and 2002 this is the climatology for July. So, this shows where the convection generally is and this is the part of a central pacific where during Elnino year you get enhanced convection and these are the OLR anomalies and you see that here you see enhanced convection here in 2002 enhanced convection here in 97 much stronger because this was a stronger Elnino, but look at a equatorial Indian ocean what has happened in 97 it was a positive phase of equino you see the convection of a west is flared up convection over east is very much suppressed on the other end you see 2002 exactly opposite convection over east is enhanced convection over west is suppressed. So, what is happening is that although the Enso is unfavorable for both these in 97 equino was favorable and this is why I talked of the tug of war, but in 2002 equino was also unfavorable. So, it colluded with Enso in making it a severe drought. So, what we have seen is that while the phase of Enso in 97 is the same as that in 2002 the phase of equino is positive in 97 and negative in 2002. Hence, while Enso and equino had opposite phases in 97 with equino being favorable they reinforce one another in 2002 with both being unfavorable. Hence, while 2002 turned out to be a severe drought in 1997 the tug of war led to a normal monsoon. Now, we find actually we have been looking at extremes here in the last slide also and we find that with these two modes Enso and equino we can quote unquote explain or understand all the extremes of ISMR. And how do we do it as I mentioned before remember Enso index and equino are not correlated. So, they are perpendicular to one another and so we can use them as axis and plot in the phase plane of these indices each year because for each year the value of each of these indices is known. So, what we get is what I will show you here this is equino x axis Enso index. Now, you take any year such as 75. 75 it so happened that equino was unfavorable minus 0.5, but Enso was highly favorable you see here this is the case. So, for 75 we know what is a Enso index we know what is equino. So, we can plot a point here as 75 now I have colored it blue why because it happens to be an excess monsoon season. So, the color shows the location shows what are the values of the indices for that season 75 and the color then indicates whether it is a excess or a drought. And remember we are looking at only excess or drought except for this special year of 97. So, we are looking only at all the extremes and you can see what we find it is amazing it is that in fact there is a separation in this phase plane of equino and Enso between droughts which are all red and excess monsoon years which are all blue. So, this says that if you are below this line L that we have drawn then there are no excess monsoon seasons if you are above this line L then there are no droughts. So, this is a very clean separation and we talked of one sided prediction if Enso index was given remember that if magnitude of the Enso index was greater than some 0.6 or 0.8 or something then we could say something about whether if Enso index was favorable and its magnitude greater than some quantity then you cannot have droughts and also equivalent statement about excess rainfall season. But we noted that many years did not meet this criteria of Enso index either being above a certain value or below a certain value there were many years which had Enso index between the two values. But in this case on the other hand what we find is that every value is accounted for any point that you have in this phase plane we can say whether which of the extremes will be ruled out. So, if the point is here then no excess rainfall season if the point happens to be here no droughts. So, this is true for all the years now and this is an amazing thing and in fact subsequent to our paper Iharah et al looked at a much longer period than 58 2003 and they just asked the question how does one linearly reconstruct ISMR on the basis of multiple regression from Enso index which they took as Nino 3 and Equino index which they took as Equin and they said that multiple regression from Nino 3 and Equin better specifies ISMR than Nino 3 alone. So, they also showed that it is important to incorporate information on Equin if you want to understand the inter annual variation of ISMR. Now, this is something that we have seen last time with the Enso. Suppose we could actually predict the composite index now what is the composite index see since there is a straight line which determines the separation between these two here. Since this line determines the separation between the droughts and the excess rainfall seasons we can determine a composite index which is a distance from this line right perpendicular distance from this line and if we say distance from this line it is 0 here it is negative everywhere here and it is positive everywhere here right. So, in terms of a composite index which is simply the distance from this line we can say if that composite index is negative then you would not get excess here if the composite index is positive then you would not get droughts. So, composite index being positive is a favorable situation from the point of view of Enso and Equin together and composite index negative is an unfavorable situation from the point of these two together. So, now we are talking in terms of composite index values and negative composite index and positive composite index there are two sets here and remember we have chosen the categories of ISMR such that the climatological probability is actually 20 percent. And what happens now if we had information on composite index a prediction of what the Enso index and Equin would be for next season and therefore, we could compute what the composite index is then if the composite index is positive then no droughts if the composite index is negative then no excess rainfall season. So, again we have chosen we are getting a very nice prediction for only for the extreme categories category one which is the lowest rainfall drought and category five which is the highest rainfall excess rainfall season. So, category five then has no droughts at all remember in when we looked at Enso alone it had a little percentage of droughts and this one has no excess rainfall years at all. So, we have got a very clean probability distribution given the composite index which is better than what we would have gotten with Enso alone. Now, so the whole it is very clear now that if we have to predict what the All India summer monsoon rainfall is going to be for next season then it would be very very useful to have prediction of both Enso index and Equin. We have already made some progress towards getting Enso predictions these are phenomenal progress in this direction was made in the 80s and 90s globally and now the question is suppose we knew some at some point in the season what Equin and Enso were can we say something about the what will happen in the rest of the season. And interestingly enough we find that if we took in fact I do not have it here I will show it in the next class this is simply showing that July plus August is also well separated even better separated than JJAS and in the next class I will show that if you plotted July plus August in the phase plane of June indices again we have a separation. So, again we will have a one sided prediction possible. So, what we have found then is that what we have found that there are two modes which are important for understanding the inter annual variation of the All India rainfall they are Enso and Equino and if we can predict these modes we will be in a very good position to say something about the non occurrence of extremes. Remember if the composite index is negative this says that there will be no excess rainfall, but you could still have either droughts or a normal monsoon. Similarly, the composite index is positive you can say there will be no droughts. So, one sided prediction of this kind will certainly be possible if we can predict both Equino and Enso. And in the following lectures we will try and see to what extent we can understand how this Equino evolves and to what extent we will be able to generate the prediction. And what is our prediction for when we will be able to predict this this is what we will cover in the next lecture.