 We are discussing the fluctuations between active and weak spells within a monsoon season and particularly the intense weak spells which are called breaks. What we have looked at in the last lecture is how breaks have been defined traditionally and what are their attributes in terms of the patterns of pressure, wind, rainfall, etc. Now, we start with another slightly another approach in trying to define breaks in terms of rainfall over the monsoon zone. As I mentioned last time, the traditional definition of breaks is on the basis of winds, surface winds. Breaks are said to occur when there is no easterly winds in the surface wind charts over the Indian region. Now, what we would like to do is to define breaks in terms of rainfall because after all the original definition of active weak spells or as active weak spells of monsoon as we understand it. In fact, involve lot of rain in active spells which Blanford as you know called height of rains, spells at the height of rains and intervals of drought as he called them which involve dry spells or relatively dry spells during the monsoon. So, it would be good to define breaks also in terms of rainfall and this was done in a set of two studies. First was by Gadgil and Joseph in 2003 and followed by Rajiv and Atal in 2010. So, let us discuss to begin with how we define breaks in terms of rainfall and what are the implications in terms of morphology of the breaks. So, that was the first paper and the second is Rajiv and Atal in 2010 which also looked at the same thing. So, Gadgil and Joseph in 2003 and Rajiv and Atal in 2008 and 10 have defined breaks and active spells as well on the basis of rainfall over the monsoon zone. Using rainfall criteria which are chosen so as to ensure a large overlap with the traditional breaks documented by Ramurthy and Dayatal. Now why did they do that because there is a large body of literature pertaining to breaks defined in the traditional way and if there is a lot of overlap between the rain breaks as defined from rainfall and these breaks then our understanding of these breaks one can use the all the understanding gain from the large body of work that exists earlier and so in fact they chose the criteria so that there was maximum overlap with the traditional breaks. Now, the basis of Gadgil and Joseph's work was that they had data for 273 well distributed stations over the Indian region for the period 1901 to 1989 obtained from IMD. So, this was the basis for the work of Gadgil and Joseph. On the other hand Rajiv and Atal had prepared in an earlier set of papers and presented in an earlier set of papers grid data for the Indian region initially it was at one degree later on it was higher resolution at 0.5 degree so they use the grid data as the basis. Now let me begin with Gadgil and Joseph and Gadgil and Joseph which I will henceforth say G.J. define breaks and active spells of the Indian monsoon in terms of the rainfall over the western and eastern parts of the monsoon zone using thresholds for the maximum minimum rainfall for a spell to be considered for a as a break or active spell. Now this is the monsoon zone that you see and this is the network of stations that Gadgil and Joseph had at their disposal. What they did was to divide it into two zones western zone and eastern zone. Now this is you see this is the mean rainfall over the monsoon zone and you can see that it is quite large over the eastern zone but smaller much smaller over the western zone. So now why did they have two zones? See when we are looking at averages over a region it is a good idea to take a region over which variations of rainfall at individual stations within a region are coherent. Otherwise the aerial average is not representative to give a simple example. If half the fluid is boiling and the other half is resting over ice and we take the average temperature of that that will be 50 degrees. Now that average temperature is not a very meaningful average because part of the fluid is actually boiling and the other part is very very cold. So it is not representative of any part of the fluid but it is an average. So same way if we have a region over which some of the stations have above normal rainfall and other stations have large deficit in rainfall and we make an average of over all those stations that average may be above rainfall or below rainfall depending on how many stations were above normal or below normal but that average will not be a meaningful representation of all the stations. This is why it is important to look at regions over which variations of rainfall are coherent. So the temporal variation of the average rainfall of any region or zone can be considered to be a meaningful representation of the temporal variation at the stations within the zone. So we are saying is the average over the region a reasonable representation of what is happening at individual stations within a region. So it can be considered to be so only if the variations at the different stations are coherent that is to say they tend to occur in phase so that the zonal average rainfall is well correlated with that at the different stations. So it is important to ensure that and with the station data that we had it was not so for the entire monsoon zone. Now the mean July August rainfall as I pointed out varies considerably across the monsoon zone with that over the eastern part being much larger than that over the western part. Furthermore the normalized anomaly which is to say the percentage departure of the rainfall during breaks the picture from Ramamurti we saw earlier where it was seen that the percentage departure over northwest goes from 60 percent to 80 percent over parts whereas in the eastern zone over the eastern zone percentage wise it was smaller. So this means that since the normalized anomaly a percentage departure of the rainfall during breaks is much larger for the western part than the eastern part. So this suggests that the threshold for rainfall used for identifying break has to be lower over the western part than the eastern part because the mean is also lower and the percentage departure is higher for breaks. So we need to have separate thresholds for these two regions and GJ therefore divided the monsoon zone into two sub zones the eastern and the western zones which are delineated so as to maximize the cross correlation between the time series of daily and five day running mean rainfall at different stations within the zones. So I am not going to get into details of that but the idea was to determine zones which are coherent with respect to variation of rainfall on daily and five day running mean scale which is the scale of interest here for inter seasonal variation. Now so with this kind of a constraint then they chose these two regions so this is the western region and this is the eastern zone western zone and eastern zone and these are chosen such that the cross correlation between the time series at different stations within a zone are maximum. So the boundary is determined so that each zone is coherent and within the zone the cross correlation between any pair of stations time series at any pair of stations is significant. So this is how the zones were identified and then one had to choose thresholds. This for western zone and eastern zone so that one can define break days and active days. Now as I mentioned before these thresholds were chosen so as to get maximum overlap with breaks identified by Ram Murthy for the period 1888 to 1967 and the later breaks up to 1997 identified by Dey et al and these were identified using the IMD definition of breaks. Now we have seen that the correlations of the zonal average rainfall for eastern and western zone with the rainfall at different stations within each zone are found to be significant at 1 percent level and most of the stations in these zones are well correlated with the zonal average and the correlation coefficient of over 80 percent or 75 percent of the stations in eastern and western zone with the zonal average is above 0.4 or 0.3. So all of this is simply saying that the zones we have determined are coherent and so the zonal average rainfall can be considered to be a meaningful representation of the rainfall at the stations within each zone. Now how do we define a break day? If a break day is defined as a day on which the rainfall over the western zone is less than a particular threshold and that over the eastern zone is also below another threshold then it is a break day. Then we determine these two thresholds so as to get maximum overlap with Ram Murthy breaks and Dey breaks then we find that if we choose the threshold for western zone to be 2.5 millimeters and for eastern zone to be 7.5 millimeters a day then the breaks we identify are by and large consistent with the breaks from the large body of literature on the subject such as Ram Murthy, Sikha and of course Dey et al and so on and Raghavan. So these were determined these thresholds were determined primarily from the overlap thing. Now I should mention that one may think oh threshold of 2.5 millimeters a day that means there is still some rain on an average over the western zone during the break but actually when we determine break days using these two thresholds the actual rainfall in long intense breaks is very much smaller than 2 millimeters per day. In fact very often close to 0 particularly on the western zone and I show you an example here this is 1972 and this is the break of 1972 and you can see that this is on the scale of 0 to 10 millimeters and you can see that actually there is hardly any rain over the western region western zone for a long period including few days preceding the break and in the eastern zone there is some rain in the initial part and actually almost 0 rain towards the end of the break this is when it has become most intense. So these are the thresholds here. So these are the thresholds that were imposed for breaking a for defining a break but in actual fact the rainfall on those days break during the breaks is very much smaller in fact very close to 0 this has to be born in mind. Now how do we define active spells and intense spells? So thresholds of 8 millimeter per day for western zone and 12.5 millimeters per day for the eastern zone were chosen for identifying active spells. Now an active day is one on which the rainfall over the western and eastern zones is higher than the respective threshold. So western zone has to be higher than 8 millimeters per day and eastern zone has to be higher than 12.5 millimeters per day for it to be an active spell. So active spell then we envisage as one in which it is raining all over the monsoon zone both over the western zone as well as the eastern zone both the sub regions of the monsoon zone it rains and provided the rain is high enough to be above the threshold for the averages then we call it an active day and days continuous days with active days are called active spells. Now there are in addition to these active spells in which the rainfall is occurring all over the monsoon zone there are some days on which rainfall is associated with relatively intense systems such as depression, cyclonic storms and so on. For these intense systems the rainfall is not well distributed over the monsoon zone rather heavy rainfall occurs only or in the region of influence of that system which is smaller than the monsoon zone on any day. So GJ take the intense spells to comprise of days on which the average rainfall for the western zone exceeds 20 millimeters or that for the eastern zone exceeds 30 millimeters. So again they have two thresholds for it to make it as an intense category over the eastern zone it has to be more than 3 centimeters for the western zone it has to be more than 2 centimeters but these two do not have to be simultaneously satisfied for intense rainfall criteria to be met intense spell can occur either over one or the other zone. So this is how intense spells are defined. Now the composite patterns of rainfall anomaly for GJ breaks and active spells are shown in the next slide and what you see here primarily is that for the break you know you have entirely negative anomalies here and active spells you have entirely positive anomalies here you see also that over southeast part of peninsula there is a reversal of signs and over the foothill of Himalayas also there is a reversal of signs. So in fact it is very interesting that given this definition that we have of the rain breaks what we get is actually two mirror images the active spell rainfall over the monsoon zone the anomalies is a mirror image of this one with all the signs of the anomalies change to the opposite sign. So negative anomaly here changes to positive anomaly here and you have positive anomaly here which changes to negative anomaly here and similar for the foothills. So this is an interesting point now the break and active composites of OLR and OLR anomaly are seen here. Now first of all let us see the composite of the OLR anomaly let me remind you that outgoing long wave radiation is when it is low it means that the emitting surface is very high the cloud tops are very high and we have deep convection and rainfall. So if we have a negative anomaly of OLR this means there is going to be more rain and positive anomaly means rain is suppressed. So for the break anomaly naturally you see over the Indian region a huge suppression of rainfall or a big positive anomaly notice that the west coast also has a similar sign here of the anomaly and let me just go back and point out here that even in for the break and active by and large the west coast anomalies are of the same sign as those on the monsoon zone that is to say during active spells on the of the monsoon it the west coast also tends to be active and during breaks the west coast also tends to have suppress rainfall but our criteria was based only on monsoon zone. So it so happens that the correlation is such that most of the time the west coast rain is in phase with the rain of the monsoon zone this is something to remember and now we look at the OLR anomalies and OLR anomaly show that very clearly that during breaks the west coast also seems to have a break or suppression of rainfall but the biggest signal is here over the eastern equatorial Indian Ocean here there is a huge negative anomaly this means that the convection is flaring up here there is more rain here during breaks and let me remind you that actually this was pointed out by Koteshwaram long ago that there is a formation of lows in the equatorial region and Simpson also suggested that in fact the equatorial Indian Ocean becomes rainier during breaks. So it is a manifestation of this that you see notice that in fact corresponding to this positive OLR there is also a positive OLR here which means convection is suppressed also over the west pacific but to its north it is enhanced here and this has been called a quadrupole because you can see this and this are opposite sign this and this are opposite sign and this and this are opposite sign this is a quadrupole characterizing the break now if we go to the active spell what we find is that the major feature namely out of phase variation with the eastern equatorial Indian Ocean is there so that when it is active here it is actually suppressed over the eastern equatorial Indian Ocean as you see here but you do not quite see an equivalent of a quadrupole here and that is to say you do not see a mirror image of active and break signatures over the west pacific it is somewhat more complicated than that. So the dominant signal is the out of phase oscillation between the OLR of the monsoon zone and that over the eastern equatorial Indian Ocean. The convection over the east pacific is also out of phase with that over the Indian region and this I should point out to you see here this is the break anomaly it is positive over the Indian region and notice that over east pacific here it is negative and for the active phase exactly opposite holds when it is positive over the Indian region there is a signal here as well. So it is interesting that when we wrote this paper here we did talk about the east pacific but we did not emphasize this seesaw in convection between Indian region and the eastern equatorial Indian Ocean. The importance of this eastern equatorial Indian Ocean dawned on us only later on and then when we went back and looked at the composite lower and behold it was very much there I will come to this aspect of the link of monsoon to convection on the equatorial Indian Ocean in later lectures but it turns out that the link is very strong. So as I mentioned note that the break anomaly is not simply a mirror image of the active anomaly particularly over west pacific. Now these are the OLR composite patterns so these are not anomalies this is the actual average OLR for break and for active. What you see for active what is done is only 240 onwards is plotted and within 200 regions are shaded. So 240 this is 240 contour here in fact 255 has also been plotted but 240 is this contour here this is 240 watts per meter squared is considered as the limiting OLR that is to say OLR below 240 generally is associated with deep convection and rainfall. And of course OLR below 200 is very much so and what you see is that during breaks we have rain over this region and over the equatorial region but the eastern part not over the western part and of course Indian region is dry whereas during active case we have suppression of rainfall over here you see very high OLR here little bit of actually enhancement of rain over the western part. So there is a convection seems to be in phase with west and out of phase with the eastern equatorial Indian Ocean as I said we will come back to this point. Another interesting thing to see is that if you look at the active composite it appears as if this TCG is in fact coherent right across all the way from India up to the east Pacific here this is a very interesting thing. So active phenomena seems to be have organization of convection over a very very large scale compared to breaks. So this is an interesting point we have to look at. Now as I mentioned OLR for the break composite is higher than 240 watts per meter squared which is considered to be the limit for deep convection and rainfall over most of the Indian region. Now these are the main results that I mentioned from the Gadge-Joseph study. Rajiv and Nital studies somewhat similar because in fact it is rather similar because it again tries to define thresholds such that there is maximum overlap with traditional breaks and it is also based on rainfall over India and particularly rainfall over monsoon zone. The difference is in the basic data set. So in this paper as I mentioned here the criteria for the identification of active and break spells on the basis of recently derived daily gridded rainfall data set. So the basic data set they have used is the data set derived and in fact presented by Rajiv and Nital in 2006 Rajiv was at IMD then derived by IMD which is a daily gridded rainfall data set which we used to be one degree grid and which is actually routinely updated by the India Meteorological Department. So using these data criteria for identification of active and break spells are suggested and the criteria were carefully chosen so that they can be used on real time applications during the monsoon using operational daily rainfall analysis. So this is the value of this work that in fact one can operationalize the definitions of break and active spells by using the criteria they have suggested. Now the criteria adopted were derived from the rainfall over the region over which significant rainfall fluctuations between active and break spells are observed namely the core monsoon zone and this core monsoon zone is roughly from 18 to 28 and 65 to 88 this is the core monsoon zone as you see plotted on top of July to July and August mean rainfall. So this is where the rainfall is very high rainfall decreases as you go higher towards the northwest and this is the core monsoon zone within which the CTCG fluctuates primarily in the peak monsoon months of July and August. So data over this core monsoon zone has been used for defining active and break spells and while choosing this zone care was taken not to include foothills of Himalayas where substantial rainfall is received during the monsoon breaks. Now that we have seen earlier even in the break and active composites of Gargir and Joseph. Now this core monsoon zone is very similar to the geographical area considered by Gargir and Joseph for identifying active and break spells. The major advantage of using graded data instead of station data is that because of the smoothing the entire monsoon zone turns out to be coherent. So all the exercise that GJ had to do in determining coherent zones and determining separate thresholds for the two coherent zones is now not required because the entire monsoon zone becomes coherent. Now and this is seen here this is the correlation coefficient of the 5 day average rainfall during July and August over the monsoon zone with rainfall at all the grid points. And so you see that the average rainfall for the monsoon zone is significantly correlated with all the grids in the monsoon zone and in fact is negatively correlated with rainfall over Himalayan foothills and over the southeastern part of the peninsula. Notice also that it is positively correlated with rainfall over the west coast. This is something we have noticed in the complexes as well in the composites of breaks and active spells that Gargir and Joseph had derived also. So this correlation coefficient then shows that the average zonal rainfall is significantly correlated with the rainfall over different grids showing that the intracesional variation is coherent over this zone and the average rainfall over this zone is indeed representative of rainfall within sub regions of the zone. Now it is seen that the rainfall over northeast India and southeast peninsula is negatively correlated with the rainfall over the core monsoon zone as I had pointed out. Now active and break events were identified by averaging the daily rainfall over this core monsoon zone and standardizing the daily rainfall time series by subtracting from its long term normal and by dividing with its daily standard deviation. So, first of all what one did was to identify the core monsoon zone, make a time series of the rainfall over the core monsoon zone daily rainfall, then make a time series of the anomalies by subtracting the mean and make it now standardized anomalies by dividing by the standard deviation. Now having got this, how do we identify the break spell? It turns out that a very very logical identification leads to very good results here. The break spell has been identified as the period during which the standardized rainfall anomaly is less than minus 1 consecutively for 3 days or more. That is to say that the rainfall anomaly is negative, it is deficit rainfall and larger in magnitude than the standard deviation. That is when we call it a break day and if it occurs continuously for 3 days or more that is a break spell. Remember that Ramurthy also had a lower limit of 3 days for the duration of the break. Now similarly the active periods are identified as the periods during which the rainfall anomaly is more than plus 1 standard deviation. In other words in active spells monsoon zone average rainfall anomaly is actually greater is positive, so that it is rainfall is more than normal and larger in magnitude than one standard deviation. That is when we call it an active day and it has to remain consecutively for 3 days or more for it to be called an active spell. Now the break spells identified in this study using the above method are comparable with those defined by Ramurthy and Day et al and there is an extremely large overlap with those identified by Gargill and Joseph which is not surprising because after all the basic criteria used are very similar it is the rainfall over the monsoon zone being deficit. So in fact there is a very large overlap and the details are in the paper of Rajeevan et al about how much overlap there is with Ramurthy and Day breaks and so on. Now these are the break spells and active spell composite figures and yellow and orange is negative anomalies and blues are all positive anomalies. You can see that in the break spells you have a huge anomaly right across the monsoon zone and in fact a pretty large anomaly over the west coast as well whereas active spells you have exactly the opposite pattern and notice the opposite pattern over southeast peninsula and the foothills of the Himalayas. These are really very much mirror images of one another the break anomaly and the active anomaly from Rajeevan et al. Again this is similar to what Gargill Joseph had found. Now these are the OLR composites for break spells and active spells and these are the anomalies. So what you see is this is the break spell on top and you see suppressed convection and rainfall over the Indian region and enhanced convection or rainfall over the eastern equator in Indian Ocean. Exactly opposite in active spells active here and suppressed here and this is the quadrupole that you can see plus minus plus and minus that you can see in the break composite it is somewhat mixed up here in the active case. So here again active is not a mirror image of the break composite over the west pacific. Now if one looks at what is the correlation of OLR over the monsoon zone see so far we have been looking at specific events breaks and active spells but if you ask the general question that on the 5 day scale or whatever what is the correlation of the OLR over the monsoon zone to OLR everywhere else then what you find is of course the OLR of the OLR of the monsoon zone is highest here where the monsoon zone is and positive but eastern equatorial Indian Ocean has a negative OLR here and you also see a negative OLR here. So the part of the quadrupole is very much there even in the OLR correlation and notice this outer phase oscillation with east pacific is also there. So it is in the general nature of things that on the inter seasonal scale now in fact you have outer phase oscillations between the Indian region and the eastern equatorial Indian Ocean between the Indian region and the western pacific not western pacific and also eastern pacific. So this is part nature of the beast if you wish and later on when we look at links of monsoon to convection over ocean all these factors will come in. So the OLR anomaly patterns for breaks and active spells are similar to those of GJ with the largest signal being the anomaly over the eastern equatorial Indian Ocean being opposite sign to that over the monsoon zone. Note that as in GJ the break anomaly is not simply a mirror image of active anomaly particularly over the west pacific. Now how does the monsoon revive from breaks? There have been two mechanisms have been suggested either by northward propagation of the TCG from the equatorial Indian Ocean and we saw in fact in the earlier lecture that northward propagation of TCG from the equatorially Indian Ocean onto the heated subcontinent is the dominant feature of the TCG as observed from satellites. So these northward propagations are certainly there and they play a role in the revival from breaks. Alternately revival from breaks can also occur by genesis of a system over the head bay of Bengal and its propagation across the monsoon zone. In fact when Ram Murthy did his study only the latter this synoptic scale system being generated in the head bay and moving across the monsoon zone that was the only hypothesis proposed and Ram Murthy meticulously analyzed all the data at hand and came to the conclusion that only about 45 to 50 percent of the breaks are terminated by this process. This suggests that almost half the breaks or even more are terminated by northward propagation of the TCG from the equatorial Indian Ocean. But you can also see that the role of convection over oceans is very very important in monsoon. Once you get a break revival can only occur with genesis of convection over ocean and propagation onto a region either from the head bay as in the second mechanism or from the equatorial Indian Ocean as suggested in the poleward propagation mode. Now I would like to illustrate this revival of the monsoon from breaks or transition to an active phase occurs with either a northward propagation of the equatorial TCG or with westward propagation of synoptic scale system generated over the warm waters of the Bay of Bengal. However in recent literature people have got so carried away with northward propagations that often this other method which was originally the only hypothesis for revival is ignored the process by which you get genesis in the bay over the head bay and revival of the monsoon. Now I would like to illustrate this by looking at rainfall over the Indian region. So this is 10 north this is 30 north and these are dates and this is 1979 where revival occurred with northward propagation okay. So what you see here is latitude north going north these are the kind of images we had earlier at 70 80 and 90 degrees but instead of clouds now this is rainfall and orange and red mean very heavy rainfall and lighter greens is somewhat lower okay. So now if you look at 79 you can see this is the onset and then the system is sort of stable here this is for central longitudes of India 77.5 to 82.5 where there is considerable data available and then a break occurred here and during the break you see in the equatorial region systems are forming slowly moving northward and the revival occurs through this northward propagation here. So this is a case of northward propagation you can ask the question was there any propagation from headway on to the monsoon zone and the answer is no the revival is entirely by northward propagation because if we look at the band average between 20 and 25 north now we are looking at to see if there is any propagation either towards the east or the west and we are beginning here with the longitude of the Bay of Bengal 87.5 that is near the Bay of Bengal and going westward and westward propagation would look somewhat like this on this map here and you see there is no westward propagation here during the break on the other hand there is one case here in which you see that the propagation occurred primarily from the west to the east and that is this break here 72 average between 20 and 25 and you see that system this is the break here and the system formed here and moved over the headway. So both these can be illustrated now I will not go into great detail on this but since Rajeevan et al had actually derived composites for breaks they could also derive lag composites so this is lag of 12 days ahead of the break day the center of the break spell being taken as 0 so this is minus 12 minus 8 minus 4 and you can see that about 4 days before the break already you see signals of negative anomalies over the monsoon zone here and now they are intensifying interestingly here when you get this you also get the typical pattern of more rain over the foothills and more rain over southeast peninsula again lag of minus 2 it is more intense lag 0 of course is the most intense and then this is plus 2 2 days after the height of the break and you can see the whole system is weakening both the anomalies here as well as here are weakening and then finally what happens is here you see a spreading of the positive anomaly of rainfall and that anomaly seems to move northward here and move further northward here and eventually it actually covers the monsoon zone so this seems to show revival by northward propagation and lagged similarly one can look at lagged correlation with lag composites for active spells active spell again what you find is that it goes towards the break mode again beginning with the anomaly in the southeast so in this scenario then the negative anomaly also propagates from north from south to north another important thing to look at is we are concerned with dry spells in the monsoon but do they actually matter in terms of the seasonal rainfall as a whole all in there is seasonal rainfall June to September rainfall because these are breaks only during July and August that we have been talking about now in fact the two are related because as I have shown you several examples for example the break of 72 which occurred in the drought year of 72 and Ramdas' example of 1918 in which also there was a long spell and it was a drought year so phenomenon of breaks has been of great interest because it is been known for a long time that several droughts are associated with long breaks in the season so Ramdas' comparison let me remind you of the so-called weekly rainfall over the so-called central provinces and central India which I again bring here you see this is the break and there was a drought in 1918 so this is the long break there now Krishnamurti and Balmay also showed that the major difference between the rainfall variation in good and some poor monsoon season is the occurrence of a long dry spell or break in the latter and this you see in the picture of 72 in which the seasonal rainfall this is the all India seasonal rainfall was only 73 centimeter it was a very major drought and a very good monsoon year of 75 in which the seasonal rainfall was 101 centimeter remember the mean is around 85 centimeter so this was very much higher than normal and you can see that the difference between these two is not so much in the active spells in fact the active spells seem even higher in 72 but it is in the presence of this big break which occurred in 72 now Sikha also showed that the average number of break days in poor monsoon years is much larger than that in good monsoon years so since ISMR is highly correlated with rainfall over monsoon zone it is not surprising that this happens so this is the all India summer monsoon rainfall and this is the monsoon zone rainfall okay and the correlation over I think 90 years is extremely high you can see how well the two are related the correlation coefficient is 0.91 so it is not surprising that a major deficit over the monsoon zone will be associated with a drought of the all India monsoon rainfall so now let us see the relationship between ISMR and active break days identified by using Ram Murthy criteria and on the basis of the rainfall over the monsoon zone as determined by Gargil Joseph and Rajiv Natal okay now what is the result that ISMR is significantly negatively correlated with the number of rain break days as determined by both G.J. and Rajiv and also by Ram Murthy but the correlation with Ram Murthy breaks is much smaller so here you have these are the rain break days of Gargil Joseph and this is the all India monsoon rainfall and as the number of rain break days increase in fact the rainfall does decrease very much but when the break days are within a certain limit you can see there is a huge spread here okay so it is true that long breaks are generally associated with droughts but you see here there are some long breaks in which there were no droughts okay the rainfall was reasonable so it isn't a one to one correspondence in fact very few things in meteorology are one to one correspondence it's a complex system but the correlation is significant it's minus 0.4 so we can say that the inter seasonal variation is related to inter annual variation with a significant correlation between all India summer monsoon rainfall and the number of rain break days same thing holds also for Ram a break days but you see the correlation is less than half the magnitude of the correlation coefficient is less than half of what you get with rain break days so this is a better indicator of what will happen in the year as well now if we took the Rajiv and et al thing also again we have very good correlation in fact it's higher it's minus 0.61 but notice that even here in a year like 1991 there was a huge deficit of all India monsoon rainfall it was all 8% or so and yet there was not a single break day zero break days and similarly 74 was a very very high deficit very intense drought and only 5 break days so there are problems in looking for high correlations when the number of break days are small there is a spread but when you go beyond say 18 days or so then a drought is guaranteed if the broke break prolongs beyond that another example of the 99 kind of case where you have almost no breaks but deficit rain in fact occurred in this monsoon now what you see here when I mentioned Rajiv and et al I said they have defined things so that they can be operationally determined and indeed operational values of the monsoon zone rainfall actual as well as the normal are available at the IMD website in Pune and this is what you see for 2012 this is the average and this is what happened the green bars or what happened in 2012 and this is the anomaly and whenever the anomaly magnitude is larger then one if it is deficit it is red and if it is excess it is green so what happens is in June there were a lot of deficit towards the end of June very large negative anomalies occurred over the monsoon zone but in July August if you see there are only 4 or 5 days of rain breaks and they are all relatively they are all scattered there is no breaks fell as such so this is an interesting result but the point is that with this definition criteria defined by Rajiv and others one can actually monitor the occurrence of breaks on a year to year basis with the IMD data so given the relationship of inter-seasonal variation with inter-annual variation understanding the break monsoon is not only important for prediction of this important event on the inter-seasonal scale but also may pave the way for understanding the poor monsoon seasons or droughts now there is a controversy about whether inter-annual variation of the monsoon rainfall is merely a manifestation of inter-seasonal rainfall I will discuss this at a later point in the lecture series now the relationship okay now let us look at the relationship between active days and ismrc so far we have said that the relationship of ismr with break days is strong what about with active days the relationship of ismr with active days is not as strong as with break days and in fact this is what is shown here you see here this is from Godgid and Joseph these are active rainfall and notice also the relationship to intense rainfall days see they had determined also intense rainfall days and in fact the correlation with intense rainfall days is almost zero okay this ties in with the fact that Sikha had shown that the all India monsoon rainfall really does not depend on the number of depressions or these strong events which give these intense rainfall days it actually depends on lows which gives much more distributed rainfall over the monsoon zone now active of course it is there is a positive correlation but it is not as large as the breaks and this is the correlation for Rajiv and data for some reason correlation with is active is somewhat less than that for Godgid Joseph but notice here 2006 which had actually maximum number of active days is just a normal monsoon zero ismr departure so the relation is not so strong so then we have actually looked at all the major features of the breaks and we find that the rain breaks defined by Godgid Joseph which are very similar to the rain breaks defined by Rajeevan et al and rain breaks by Rajeevan et al are in fact such that one can operationally monitor them with the data available at IMD and that these breaks have a very large overlap with breaks defined in the traditional way so basically traditionally of course the breaks were defined with easterly disappearance of the easterly winds on the surface wind charts and so on now when we define it on the basis of rain breaks that is to say rain over monsoon zone we are getting very very similar results except a better correlation with ismr now what has happened is that the term break monsoon has captured the imagination of a lot of scientists because well I do not know precisely why and so much so that different scientists have used the same term to denote different features of convection and or circulation over different regions so while traditionally the term break refers to very weak spells of rainfall over the Indian monsoon zone in July and August breaks have been identified for the entire summer monsoon season by some scientists and there are many examples of this which I could give for example we are concerned only with the monsoons over India but Webster et al look at the work done by Magana and Webster in which they take a very large region 65 to 95 and 10 to 20 north and use two criteria one is weak spells of convection so OLR is involved as well as 850 millibar zonal winds over this region and both those are taken to indicate breaks because it is believed that breaks in the Indian monsoon are on a scale which is much larger than India or even South Asia now on the other hand Goswami and Mohan define breaks on the basis of the strength of 850 HPA wind at a single grid point 15 north and 90 east and Krishna et al define breaks as death with positive OLR anomalies over northwest and central India provided that the average OLR anomaly exceeds 10 watts so each one has their own definition of breaks and even when the so-called breaks are identified in terms of rainfall or convection over Indian region a variety of definitions are used so Rodwell and Anna Malai use the term break to denote weak spells of all India average rainfall calculated operationally by IMD whereas Cadet and Daniel have some 15 stations over which they take the average rainfall and denote weak spells of that as breaks the point is they keep all of them use the word break monsoon at the same time what they refer to are different criteria for definition of breaks now Krishnamurti and Shukla have defined active spells and breaks on the basis of all India daily rainfall with break days associated with negative anomaly of magnitude greater than 0.5 now this is very peculiar that they themselves have used all India rainfall we have seen that the monsoon rain zone rainfall is negatively correlated with rainfall over the foothill of Himalayas as well as the southeastern region so average all India average on a daily scale does not make much sense because the variances are not coherent and this actually a support for this came from their own study earlier see their use of this is because this Krishnamurti and Shukla had shown that the dominant mode in the daily rainfall as anomalies of one sign over central India and anomalies of the opposite sign over the foothills of the Himalayas and over southeastern peninsula and yet they insist on taking an average over all India so although the intracisional variations are not coherent anyway so many such criteria are used and since different criteria are used for definitions of breaks in different studies there are differences in the breaks identified hence in the duration their frequency of occurrence as well as the associated circulation and convection patterns so G.J. showed that there is hardly any overlap between breaks identified by them and those of Webster et al which means there is hardly any overlap between the traditional breaks also because there is an and Webster et al so Webster et al's breaks are something that are different from the breaks that are recognized as breaks of the Indian monsoon by the meteorologists over a large number of years see the second thing the state is that within each season there are three or four active break sequences so they are really talking of weak spells so that you have active weak fluctuations and we have seen every year there are many active weak fluctuations beginning from June ending in September so Ramurti breaks however do not occur every year they occur only in some years for example during 1901 to 67 there were no breaks in 10 years so it is not a phenomena that you see every year we have seen in the plot that for several years there were zero break days so break is a very special weak spell but what Webster and others are referring to are general weak spells and the morphology of those weak spells may be very different so while the duration of the breaks identified by Ramurti varies from 3 days to over 15 days with over 30 percent of the breaks of duration 7 days and longer so there are lot of long breaks in Ramurti breaks the duration of breaks identified by Webster et al is generally short varying from 1 to 7 days with 90 percent of the breaks of duration 3 to 5 days so these are really weak spells from which the monsoon over the large region that they call monsoonal in fact revives so there are then in fact they are looking at a different beast when they talk of breaks and one has to be aware of that now we can look at the OLR composites this is the one you have seen earlier this is from Rajivan et al this is the break and this is the active spell and if we look at the composites from different definitions this is Krishnamurti and Shukla but let us look at Webster to begin with this is his active spells and this is his weak spell you find that during the break in fact he does get enhancement over the equatorial region but it is over the entire equatorial region not just over the eastern see particularly in the active spell is when he is getting active thing over the eastern equatorial Indian region so you know this is the same sign of anomaly for both the cases this is very different from the kind of things that we had seen earlier now if we look at Goswami break you find this is the Goswami active one and this is Goswami break and you find Goswami break involves actually lot of rain all over here over the eastern Indian ocean it is realistic but also over the west coast and so on but we had seen that west coast tends to go along with monsoon zone so what happens is basic features of the breaks identified and studied over the years are not present here so they are looking at different facet of the system when they talk of breaks and this is very much seen in the this is the precipitable water content anomalies from Webster and you can see this is the break situation where you have very much less water here and very much more water here so the entire equatorial ITCG is enhanced convection is enhanced whereas what we had found was eastern part is enhanced and the western part is also slightly suppressed so you are getting very different anomalies and I will not dwell on this or I may just mention that while we have talked of breaks and active spells as events you know there is an approach in which they look at modes so they will filter for a 30 to 50 day mode and so on and they look for active spell as a crest and a weak spell as a trough of the amplitude of that mode so that may not translate into actual days. Now so there are many differences and one has to be very careful in seeing how they define break before using whatever they have learnt about breaks their processes and so on and so forth. So I think it is extremely important to predict this feature of intracesional variation of the monsoon and for that a deep understanding of the nature of the two phases active spells weak spells and breaks and the transition between them is essential only then models capable of simulating and predicting them can be developed. In this endeavor it would be good to adopt the definition of breaks proposed by Rajiv and Atal based on rainfall of the monsoon zone which is calculated operationally by IMD. So this is the suggestion then if you go on redefining the term break then it will be a full chase because we do not know what we are trying to understand or predict it is better to use this definition which is also can be used operationally and try and then understand how the breaks occur how the transition occurs from breaks to active spells and vice versa and try and see if we can incorporate that understanding into models so that they can predict these very important events. Thank you.