By Dr. M M Ali — June 7 2023
Why we should observe the ocean to better predict the atmosphere?
I may be embroiled in controversy with the monsoon professionals and modelers by writing this article. However, I would like to trespass and put forth some of my ideas in meeting the challenges to predicting the mysterious Indian southwest monsoon rainfall by including ocean observations.
Monsoon rains in India are mystical and erratic, either with too little water causing drought or far too much flooding! Even worse, extreme rainfall events have been very common in recent years with the number of days with heavy rains increasing and longer dry spells stretching out in between. As a result, normal and steady rains that can reliably penetrate the soil are decreasing. Adding to this, increasing deforestation and urbanization continuously reduce the infiltration capacity of the soil. As a result, groundwater is withdrawn faster than rain can recharge it. This is an alarming situation for a country like India, which gets the maximum share of its water through the rain. In addition, [it is claimed – kh] climate change is now messing with the monsoon, making seasonal rains more intense and less predictable.
Adding to these problems is the predictability or the unpredictability of the monsoon rainfall!
A monsoon is a seasonal reversal in the prevailing wind direction, that is usually initiated by the land sea temperature contrast. The Indian summer monsoon, for example, is triggered when the land gets heated up more than the surrounding sea during the summer creating a pressure gradient between the land and the sea (Figure 1).
Figure 1: Climatology of the southwest monsoon circulation during Indian summer (Source: http://www.meteo.co.in/Monsoons/image004.jpg).
While efforts are underway to improve the understanding of the physics of the problem of monsoon rainfall prediction, it is worthwhile to look again at the efficiency of the input parameters presently used in models and to look for new approaches. Sea surface temperature (SST) is one such parameter that needs to be reconsidered for predicting the Indian summer monsoon rainfall (ISMR). SST is routinely used for predicting the weather phenomena such as monsoons or cyclones, while it is well established that the thermal energy required for atmospheric phenomena comes from the upper ocean, not from the thin layer of the ocean sometimes reflected in SST alone. SST is restricted to a few millimeters of the top ocean layer, particularly when it is estimated from the satellites and is largely influenced by strong winds, evaporation, or thick clouds. Hence, it does not reflect the thermal energy available in the upper ocean. Rapid (of the order of a day to a month) heating (such as strong solar heating) and cooling (such as more evaporation due to strong winds and/or clouds) events can quickly erase the thermal signature of subsurface warm or cold features (Pickard and Emery 1990), leading to SST misrepresenting the ocean thermal energy. In contrast, ocean mean temperature (OMT), which is measured up to a depth of 26o C isotherm or up to a fixed depth, is more stable and consistent, the spatial spread of which is also less compared to SST. This is even evident from the average coefficient of variation, defined as the relative magnitude of the standard deviation to the average (1993–2017) value (Figure 2) for monthly SST (0.02) being double that for OMT (0.01) for the North Indian Ocean.
Figure 2: Coefficient of variation of SST (a) and OMT (b) during 1993–2017. The rectangle represents the south Indian ocean area that has a major influence in ISMR. (courtesy: Venugopal et al. 2017).
The application of ocean thermal energy for cyclone studies was already demonstrated through several studies. For example, Mao et al. 2000 reported that the rate of intensification and final intensity of cyclones are sensitive to the initial spatial distribution of the mixed layer, a proxy for ocean thermal energy, rather than to SST alone. Similarly, Shay et al. (2000), Ali et al. (2007), Mainelli et al. (2008), Ali et al. (2013) and Lin et al. (2013) and Jaimes and Shay, (2015) demonstrated/suggested the importance of ocean thermal energy for cyclone studies.
Similarly, the role of the heat energy available in the sub-surface ocean layers in El Nino studies was confirmed by Smith et al. (1995), Ji et al. (1997) and Latif et al. (1998). They proved that even the El Niño forecast models could be improved by initializing the models with the observed ocean heat content (OHC). OHC is the amount of thermal energy available in the oceans from surface to a fixed depth, say, 100m or 200m given by equation (1).
where ρ is the density of the sea water, Cp is the specific heat capacity of the seawater at constant pressure, p; h1 the top depth, h2 the bottom depth, dz the thickness of the layer and T is the average temperature of the layer in oC. Although in situ temperature profiles are required to estimate this parameter, it can be indirectly inferred from the satellite-derived sea surface height anomaly (SSHA) and SST.
Based on this concept, Venugopal et al. (2018) analysed 25 years of OMT (ocean mean temperature) of the north Indian Ocean (NIO) from 1993 to 2017 spanning 30°S to 30°N and 40°E to 100°E, with a grid spacing of 0.25° × 0.25°. They computed OMT from the Tropical Cyclone Heat Potential (TCHP) and the depth of 260 C (D26) obtained from the National Oceanic and Atmospheric Administration, Atlantic Oceanographic and Meteorological Laboratory (ftp.aoml.noaa.gov). The 26 degree C isotherm is seen at depths varying from 50–100 meters. During January–March, the mean 26 degree C isotherm depth in the Southwestern Equatorial Indian Ocean (SEIO), the rectangular area shown in figure 2 is 59 meters. The researchers analysed 25-year OMT data from 1993 to 2017. They found that unlike SST, OMT of SEIO was able to correctly predict 20 out of 25 years (80% success rate) whether the amount of rainfall during the summer monsoon was more or less than the long-term mean, 887.5 mm. The prediction based solely on sea surface temperature was correct only for 15 out of 25 years (with a 60% success rate). Using this approach the monsoon was predicted to be more than the average during 2018-2022 with an 80% success rate.
In addition to better predictive score, the information on whether the amount of monsoon rainfall will be more or less than the long-term mean will be available by the beginning of April, two months before the southwest monsoon sets in. This is because OMT is analysed by measuring the ocean thermal energy during the period from January to March and the southwest monsoon sets in around June 1 each year in Kerala .
Out of the 10 years of observed rainfall, there are 6 below average and 4 above average rainfall years during 2013-2022 (Figure 3). The all-India monsoon rainfall in 2014, 2015, 2016, 2017, 2018, and 2021 was below average during June–September but above average in 2013, 2019, 2020, and 2022. Except in 2016, all the years were predicted correctly using OMT estimated during January-March. Venugopal et al. (2018) claimed an accuracy of 80% for their 25 years study. Since the predictions during 2018-2022 were also correct, the success rate has now increased from 80% to 83.3%.
With this success rate, our prediction for the all-India June-September 2023 is that total rainfall is likely to be less than 887.5 mm. We have to wait and watch what happens!
# # # # #
List of references:
Ali, M. M., P. S. V. Jagadeesh and S. Jain, Effects of eddies on Bay of Bengal cyclone intensity. Eos, Transactions American Geophysical Union 88(8): 93–95 (2007).
Ali, M. M., T. Kashyap and P. V. Nagamani, Use of sea surface temperature for cyclone intensity prediction needs a Relook. Eos, Transactions American Geophysical Union 94, 177–178 (2013).
Jaimes, B. and L.K. Shay, Enhanced wind-driven downwelling flow in warm oceanic eddy features during the intensification of tropical cyclone Isaac (2012): Observations and theory. J. Phys. Oceanogr. 45, 1667–1689. https://doi.org/10.1175/JPO-D-14-0176.1 (2015).
Ji, M. and A. Leetmaa, Impact of data assimilation on ocean initialization and El Nino prediction. Monthly Weather Review 125(5), 742–753 (1997).
Latif, M., D. Anderson, T. Barnett, M. Cane, R. Kleeman and A. Leetmaa, E. Schneider, A review of the predictability and prediction of ENSO. Journal of Geophysical Research: Oceans 103(C7), 14375–14393 (1998).
Lin, I. I., G. J. Goni, J. A. Knaff, C. Forbes and M. M. Ali, Ocean heat content for tropical cyclone intensity forecasting and its impact on storm surge. Natural Hazards 66(3), 1481–1500 (2013).
Mainelli, M., M. DeMaria, L. K. Shay and G. J. Goni, Application of oceanic heat content estimation to operational forecasting of recent Atlantic category 5 hurricanes. Weather and Forecasting 23(1), 3–16 (2008).
Mao, Q., S. W. Chang and R. L. Pfeffer, Influence of large- scale initial oceanic mixed layer depth on tropical cyclones, Mon. Weather Rev. 128, 4058–4070 (2000).
Pickard, G. L. and W. J. Emery, Descriptive physical oceanography: An introduction. Elsevier (1990).
Shay, L. K., G. J. Goni and P. G. Black, Effects of a warm oceanic feature on Hurricane Opal, Mon.Weather Rev. 128, 1366–1383 (2000).
Smith, T. M., A. G. Barnston, M. Ji and M. Chelliah, The impact of Pacific Ocean subsurface data on operational prediction of tropical Pacific SST at the NCEP. Weather and forecasting 10(4), 708–714 (1995).
# # # # #
About the Author:
Dr. M M Ali is an Indian meteorologist and author/co-author of many peer-reviewed studies on the relationships between the oceans and the atmosphere. He is currently a Senior Scientist (courtesy) at the Center for Ocean-Atmospheric Prediction Studies, Florida State University (2015 to present).
He is a co-author, with Venugopal Thandlam (as lead author) and others, of “Statistical Evidence for the Role of Southwestern Indian Ocean Heat Content in the Indian Summer Monsoon Rainfall”. T. Venugopal computed the values for 2018-2023.
Dr. Ali wrote this piece for WUWT after reading “The Southwest Monsoon — More Erratic?”.
This essay has been lightly edited by Kip Hansen ( any editing errors are mine – kh )
# # # # #
Using a constant specific heat capacity at a constant pressure is a good first approximation, but it is both temperature-dependent, and maybe (sorry, you brought in the constant pressure) pressure-dependent for a 0-300 m depth.
I agree that the Specific heat capacity at constant pressure changes with depth as the temperature changes with depth. But in the absence of that information, we used it as a constant. This is the limitation of our study.
Not sayin’ it’s not ‘valid’, but making a prediction to .5mm suggests some not to serious ‘science’ to me …
I’m not sure Dr Ali was making a prediction to within 0.5 mm.
…….
You are right Dr. Redge. Thank you.
Sorry. We are not predicting with an accuracy of 0.5mm. We can predict whether the monsoon would be above or below 887.5 mm. Please see the reply by Redge below.
Could you please quantify the rainfall expected to nearest 20 mm for this Southwest Monsoon inferring from your research based on OMT of SEIO.
The monsoons were once explained as reversal of land and sea breezes but later research suggest using larger atmospheric winds, namely the shift of the Himalayan subtropical jet stream. A quick search leads here:
https://everestweather.com/mt-everest-weather/jet-stream/
A few years ago, I found more complete explanations.
Thanks. I will refer the link provided by you for use in future.
You have to go back to 2019 to have the Arabian Sea so far above 30C at June 8. That was an above average rainfall year. Convection has kicked into overdrive and is now self-sustaining:
https://earth.nullschool.net/#current/wind/surface/level/orthographic=-282.73,7.60,1128/loc=66.545,14.214
The cyclone will pull about 3C off the surface temperature over a few days.
Convective instability preferentially favours the warmest air column that can develop a level of free convection. Cyclic convection needs TPW of at least 45mm to kick in. There is not much of the west coast of India currently showing that level of atmospheric water:
https://earth.nullschool.net/#current/wind/surface/level/overlay=total_precipitable_water/orthographic=-284.00,11.96,1441/loc=74.570,15.833
That means the cyclone will build over the ocean before heading to land. Once water gets dumped on the land the convective columns over land will cycle preferentially to the ocean columns because the afternoon land temperature will be warmer than 30C, which is the regulating temperature for ocean columns. The east coast near Chennai now has enough atmospheric water to support cyclic convection and there is sufficient mid level moisture to support instability:
https://earth.nullschool.net/#current/wind/isobaric/500hPa/overlay=relative_humidity/orthographic=-284.73,14.43,997/loc=78.517,14.949
Trees are vitally important for retaining atmospheric water over land that is needed for strong moisture convergence to the land from the ocean. The Amazon would experience similar conditions to the Sahara if all the trees were removed. Trees beget water, which begets trees. It is a virtuous cycle. Chop down trees and that cycle can reverse to give desertification.
Persian Gulf will get to 34C in August but not experience convective instability because mid level moisture is too low to enable a level of free convection to develop.
You are right. The all India June-September rainfall is likely to be less than 887.5 mm, which was the average rainfall at the time the analysis was done by Venugopal et al. (2018). This average rainfall is different than the normal rainfall which ranges between 96% and 104%.
Thank you Rick – you’ve exemplified something that dawned on me recently = the buoyancy of moist air.
That is something that should be apparent to everyone but especially to Climate Scientists.
It comes from the very basic observation that a water molecule is half the molecular weight of the average atmospheric molecule – so if you fill any volume of air with water vapour, its density will fall and thus it will become bouyant relative to surrounding dry air.
And just look what I found and calculated – how much the buoyancy changes and how it changes with Temp and Humidity.
My graph shows how the density of air, at 1,000mB constant pressure, changes with temperature at 35% Humidity and at 95% Humidity.
And just look at the divergence as temp rises – ain’t that soooo beautiful to explain Monsoons, Hurricanes/Tornados and even why hot-air balloons work so well when powered by Propane ##
Indian monsoon:
(This applies to all monsoons)
It starts with the ITCZ – which everyone tells us is a point/line around the globe where “warm air” rises.
And it moves North/South dependant upon the season = basically it stays directly under the sun at all times.
i.e The latitude of the ITCZ is = that of the sun
As Willis has told us myriad times, a line/assemblage of thunderstorms follows in the wake of the sun as it circles the globe, certainly when Sol is passing over large bodies of water.
As Eunice Foote discovered 3 years before ‘Tyndall’ – air containing water-vapour is heated more by the sun than is dry air under the same sun.
So, looking at my graph, moist air directly under the sun will become hotter and much more buoyant that air that’s even just a few degrees north/south of the overhead sun.
That air will rise and cool creating the rain and thunderstorms that follow directly behind El Sol as he moves around the globe every day
That is the ITCZ
Enter India, north of the equator at the start of June and India will find itself in the path of the ITCZ as is moves north under a very strong sun.
Enter what Rick said above:
Some of the offshore thunderstorms are bound to run inland. They won’t themselves get very far – fizzling out rapidly just like hurricanes do when they run aground.
But they will dump a lot of water, say 20, 30 50 miles inland.
But next day when Sol arrives, he will heat the resultant moist air created yesterday and because of the divergence in my graph, will trigger an inland T-storm that might drift another 50, 70 100 miles inland.
But the T-storm at 20 miles inland, will create Low Pressure which will suck more moist air off the ocean to sustain itself tomorrow.
And next day, the moisture at both 20 miles and 70 miles inland will create even greater low pressure pulling even more water off the ocean and again again again, moving (say) 50 or 100 miles inland each day.
Thus the monsoon will spread across the land (India) exactly as its seen to do.
It is that buoyancy effect of moist air and how it becomes ‘not straight line’ ever increasingly buoyant with temperature.
It works as a water pump – sucking water off the ocean – maybe as a bucket brigade and every day, the brigade gets longer.
But as you see in my graph, you need (maybe say) temps of at least 20°C to get sufficient divergence = sufficient buoyancy to trigger the T-storms.
(What it is doing is changing the Lapse rate)
Yet everyone, (childishly) says/assumes that temperature alone is driving.
Yes it is BUT, it needs water PLUS the effect of how air density of moist air plummets as temp rises.
The density of dry air behave in a straight line – hence the classic PV/T calculations and the basis for the most accurate thermometers ever made
It’s so perfect because it explains a recent rave of mine about ENSO.
Esp that if Australia had ‘some inland water’ – that water would act like a lightning conductor and it would de-fuse/drain the warm pool that builds during La Nina
IOW: Water powers Climate
## Hot Air Balloons
when they are powered by propane gas, burning that gas produces as much water vapour as it does heat.
And it is the massive drop in density of that very vapour that lifts the balloon – NOT the hot air
Thus, (lets say) an electrically powered hot air balloon would never get off the ground because the density decrease of dry air would be insufficient.
So ‘Hot Air Balloons’ are not= Hot Air Balloons…
They are actually=
Hot Water-Vapour Balloons
It also perfectly explains what we all see every single spring summer autumn day.
Thro the night = clear sky, no wind, no convection no clouds
Early morning = same as night.
But, through that clear sky, Sol will heat any water vapour there is in the air and as the lines on my graph rapidly diverge, a point will come when convection (wind and clouds) suddenly start happening.
I’m gonna put my neck on the line and assert that CO₂ does no such thing
That ‘science’ is utter garbage.
Then in the evening as the sun weakens, it all flops into reverse to give another clear night.
Summer = dayWinter = nightWe all see ‘a monsoon’ happening Every Single Day (apart from in winter) – so why can no-one explain the ‘big monsoons’
here’s the calculator – go have a play
edit: so as to remain On Topic..
If you swallowed the above, you’ll realise that ‘Predicting The Indian Monsoon‘ relies on predicting the arrival of the first large T-storm to run aground as the ITCZ moves relentlessly North under the strengthening sun.
That one T-storm triggers a cascade of storms that work as a bucket-brigade water-pump and are = The Indian Monsoon
Peta – there’s another reason an electrically-powered hot-air balloon would never get off the ground: the weight of the bloody batteries!
and you do understand and ‘get’ the Soil Erosion angle???
It is that, if India had soils with high organic content, they would by definition then have a high moisture content.
Thus there would be ‘mini-monsoons’ happening 365 days of the year and that would work as a constant slow-acting water pump, pulling water off the ocean and so maintaining the soil moisture.
(Have I just described ‘English Weather’?)
It is The Very Fact that India = desert is why it gets Monsoons.
Monsoons are = The Climate Of deserts
Monsoons are the climate of deserts
monsoons are the climate of nice places
(A Desert being defined as= A Place With Low Soil Organic Content)
Don’t you think it would be a good idea not to make any more deserts?
That is:
Stop chopping and burning treesStop ploughing and tillingStop using Nitrogen fertiliserStop using GlyphosateStop draining wetlands, swamps and aquifersGrow, eat, survive upon: Perennial Plants rather than annuals
It is because, once a desert has formed, they are utter pigs to get rid of, they don’t just ‘go away’
i.e. & e.g. Australia; The Natives created a desert out of the whole place 30,000+ years ago and it is still there.
And everything the present inhabitants do only reinforces it, by growing wheat, with grazing sheep and with ‘protective’ burnings
edit: for Big Comment/Rave as above
The divergence in the temp/humidity graph nicely explains Willis’ ‘Emergent Phenomena‘
That they emerge when the difference between the two lines is large enough
Cutting down trees makes a desert. Riiiiightttt tuh!
Wait, so England’s a desert? You told us that ‘Enry VIII cut down all the trees and caused the LIA.
The convective instability does not only rely on the buoyancy of moisture. The instability arises from the fact that moist air in thermal equilibrium with the air around it will not rise above the level of free convection. For an ocean surface at 30C, that level is 6900m. So the region below the LFC is communicating with the surface through normal convection but the region above the LFC is losing heat to space causing it dehumidify. So there is dry, dense air above the LFC and moist buoyant air below; all just waiting for ensuing instability.
Convective potential works like a heat engine where the air above the LFC dehumidifies. That creates convective potential that results in the instability causing cloudburst. that is gaining heat will not rise above the level of free convection
The slope of the lapse rate has a discontinuity at the LFC. I explain it in detail here:
https://wattsupwiththat.com/2022/07/23/ocean-atmosphere-response-to-solar-emr-at-top-of-the-atmosphere/
Nullschool attempts to show convective potential but the vertical resolution is not good enough to guarantee that it is actual CAPE. If it is over open ocean then it will be close to actual but near land and on land so not so good.
Convective potential is building in the Gulf of Mexico right now:
https://earth.nullschool.net/#current/wind/surface/level/overlay=cape/orthographic=-69.82,15.32,864/loc=-76.322,15.258
Where I have the pointer, the water is far enough away from land for it to be real CAPE but the surface temperature is still under 30C and it usually needs to get above 30C for the first cloudburst. IT takes about 10 days for the atmosphere to equilibrate in the Atlantic ocean. The Pacific takes about 24 days.
Without convective instability to pump up the atmosphere, there would be persistent cloud right down to freezing temperature and Earth would become a snowball. Convective instability begins at around 15C surface temperature. It means that a column over water at 15C can never get to 100% humidity unless there is cloudburst. It is a cyclic process. That process creates persistent cloud at 30C surface temperature that just lets enough sunlight to the surface to keep the convective engine firing. Once any region over land gets above 45mm water column it can support cyclic cloudburst so it spreads as you point out. Convective columns over land have greater tendency toward instability and usually in the afternoon. Columns over open ocean fire less frequently and often at night as the air column cools.
Very interesting discussion by all concerned.
Peta, the mass of water molecule of 18 times its specific heat, Cp, of about 4 make it a coolant. The same is true for CO2 its mass of 44 times a Cp of .846 depending on temperature makes it a coolant too.
We should not neglect the role of solar activity in the Indian monsoon. There are a gazillion articles on the subject. These are just the ones that come first in a search for “Indian monsoon solar” in Google Scholar.
Here’s ChatGPT’s take on it: 😉
How much effect does solar radiation have on the Indian Southwest Monsoon?
ChatGPT
Solar radiation plays a significant role in driving the Indian Southwest Monsoon, although it is not the sole factor responsible for its onset and intensity. The monsoon is a complex weather phenomenon influenced by various atmospheric and oceanic factors.
Solar radiation, particularly the differential heating of land and sea, plays a crucial role in establishing the temperature gradient that drives the monsoon circulation. During summer, when the Northern Hemisphere receives more solar radiation, the landmass of India heats up more quickly than the surrounding oceans. This contrast in heating creates a low-pressure area over the Indian subcontinent, which leads to the onset of the monsoon.
Solar radiation also plays a role in determining the distribution of temperature and pressure gradients across the Indian Ocean and surrounding regions. The temperature differences between the warm Indian Ocean and cooler landmasses influence the establishment and strength of the monsoon circulation.
Furthermore, solar radiation affects the water cycle and the formation of convective systems that contribute to monsoon rainfall. Solar energy provides the necessary heat to evaporate water from the oceans, forming water vapor that subsequently condenses and precipitates as rainfall over the Indian subcontinent.
It’s worth noting that while solar radiation is a crucial driver, the monsoon is a complex interplay of various factors, including sea surface temperatures, atmospheric circulation patterns, moisture content, and topography. El Niño-Southern Oscillation (ENSO) events in the Pacific Ocean, for example, can modulate the monsoon by altering the atmospheric circulation patterns and sea surface temperatures.
In summary, solar radiation plays a significant role in initiating and modulating the Indian Southwest Monsoon. However, it is just one among many factors that contribute to the complexity of the monsoon system.
Solar radiation powers the entire climate system. The interesting question is the effect of the changes in solar activity that take place with the 11-year solar cycle and secular solar cycles. Tons of research on this matter.
We never claim that the heat content of the equatorial Indian Ocean is the only parameter that influences the ISMR. This is just one of the important parameters that is being neglected! Nothing more.
Times ago, there existed a Near East Monsun
Mediterranean Moisture Source for an Early-Holocene Humid Period in the Northern Red Sea
Monsoon will return to the Mediterranean. It is already close to hitting the 30C limit which will pull in water from the Atlantic off Morocco, Spain and Southern France. Then feed water into Northern Africa once it gets wet enough to revegetate.
The traces of erosion on the great Sphinx in Egypt have been identified by Schoch and West as produced by heavy rains around that timescale.
Ayers Rock has comparable traces of rain erosion.
Kip, I am not grasping this “..,measured up to a depth of 26o C…”.
Does this mean you look to find how far down the probe (or whatever) goes until you go below 26 C?
mkelly ==> This is an essay by M M Ali, PhD, a scientist currently at Florida State University.
This page has a bit of explanation. Basically, it is the depth at which the water temperature is as high as 26 degrees C. More from NOAA here.
Thanks.
Integrated temperature from surface to a depth of 26 deg. isotherm multiplied by the density of water and the specific heat capacity, known as the tropical cyclone heat potential (TCHP) primarily represents the amount of heat energy available for atmospheric phenomena such as cyclones. Venugopal et al. (2018) studied the impact of this parameter on the south equatorial Indian Ocean and its relationship with the Indian summer monsoon season rainfall. This parameter was estimated with satellite observation after a proper validation with the in situ temperature and density profiles. For more clarity, you may see the paper referred above.
Thanks.
You don’t say a lot about air pressure, though Fig 1 hints at it.
I recall reading, some decades ago, that the breaking of the Indian Monsoon was associated with the changes of the local jet stream (whatever it’s called). In the space of a few days it breaks down from flowing south of the Himalaya to reappear north of the Himalaya, when everything starts to happen at lower altitudes.
You are right. This was just a representative figure – not much to do with the analysis.
Dr Ali,
Thanks, you have raised a very interesting topic. Yes it is true that OMT has more correlation with the atmospheric phenomena than mere SST and it has been demonstrated also in Cyclone cases (Your article also says that). As Venugopal paper says that it has 80% predictability than with SST (60%) for SMR, still the question is of 20% One need to critically analyse using longer datasets, why 20% could not be predicted, whether those years were much below/above normal years or there were any other (local/remote) phenomena controlling that. It may be good research topic.
However, for the practical purpose, as you have also mentioned that whether total ISMR prediction over whole India is useful for the agriculture/disaster monitoring/planning or rainfall prediction over a region for better planning of water resources/disaster mitigation/change in agriculture practices or the need of prediction of intense rainfall.
Raj Kumar