Meteorological factors

Unkind winds

Print edition : September 14, 2018

Figure 1 (a): The temporal variation of daily rainfall intensity. There are four distinct peaks of rainfall: two in June (around June 14 and 20), one in July (around July 20) and one sustained long one in August (beginning August 8 until August 16).

Figure 1 (b): The cumulative rainfall over Kerala during the entire monsoon season.

Figure 2

Figure 3: The presence of the monsoon trough, whose eastern end was anchored lower than the normal position (towards the south) also facilitated the pressure lows to evolve into depressions, which then moved norht-north-westwards along the trough.

Table 1

Table 2

Over the last decade, Kerala has actually been showing a decreasing trend in rainfall. Somewhat unusual meteorological factors caused this year’s rainfall to buck the trend.

Floods are due to God. Disasters are man-made” is a remark attributed to the late Gilbert White, a floodplain management expert from the University of Colorado, and which one has heard increasingly in recent times in India. By God, of course, here one means the climatic and meteorological factors responsible for the heavy rains, and one has witnessed the truism of this saying in all the country’s recent flood-related disasters: Chennai (2015), Kashmir (2014) and Kedarnath (2013). It is just as true in the case of the current disaster in Kerala.

This year, the rainfall in August was the most devastating, with the total rainfall over Kerala until August 24 being 788.3 mm at the time of writing this article. While this is certainly the highest August rainfall in living memory, one may be surprised to learn that the 2018 August rainfall is only the sixth highest.

According to Pulak Guhathakurta of the India Meteorological Department (IMD), who looked at the data of the past 143 years (1875-2017), the August 1931 rainfall of 1,132.9 mm (175 per cent of the normal) was the highest in the recorded rainfall in the country. The other four high rainfall years were 1923 with 876.6 mm, 1907 with 850.8 mm, 1897 with 820.2 mm and 1878 with 800.1 mm. But, in terms of point rainfall, several rain gauge stations recorded higher rainfall extremes than in earlier years. “There are a few days where all rain gauge stations reported more than 10 cm rainfall,” said Guhathakurta in an email.

According to him, in Idukki, the most affected district, during August 1-20, Station Idukki received five days of heavy rainfall (7-11 cm), three days of very heavy rainfall (12-20 cm) and two days of extremely heavy rainfall (more than 20 cm). The rainfall of over 6 cm in these 10 days exceeds past records of maximum frequency of heavy rainfall in a month, he said. Similarly, two other stations of Idukki district, Munnar and Peermadu, received more than 6 cm of rainfall on eight days each, with the respective category-wise break-up being two high, one very high and three extremely high and two high, three very high and three extremely high.

On August 16, stations at Peermadu, Idukki and Munnar also recorded the highest ever rainfall on a single day, exceeding the previous highest with 34.9 cm, 29.5 cm and 29.2 cm respectively. The figures are somewhat reminiscent of figures preceding the Chennai deluge of 2015.

“Such heavy spells will only increase in the coming years,” pointed out M. Rajeevan, Secretary, Ministry of Earth Sciences. Indeed, in their study published in 2008, Rajeevan and Guhathakurta analysed 104 years of gridded rainfall data over the country and observed that the frequency of extreme rainfall events showed not only significant inter-annual and inter-decadal variability but also a statistically significant long-term trend of 6 per cent increase per decade. This seems to be strongly influenced by warm sea surface temperatures (SSTs) over the tropical Indian Ocean that in turn could be in response to global warming and climate change. “A 15 cm rainfall that used to occur over a day now occurs within eight to 10 hours,” Rajeevan added.

Only a year earlier, the duo had studied the long-term trend in rainfall over the country and concluded that rainfall over Kerala showed a decreasing trend. Kerala has, indeed, over the last decade and more been receiving less rainfall, and whenever there have been episodes of heavy rains, these have largely been in the coastal regions. This may be why the authorities did not expect floods of such a devastating magnitude this year and why there was less than adequate preparedness. But the very heavy rainfall of 2018 is, statistically speaking, but one datapoint in the long-term trend, which could still be decreasing, and so the two are not contradictory. The causes for this year’s rainfall bucking the trend are somewhat unusual meteorological factors, which will be discussed presently.

The model-based short-range forecasts (three to five days ahead) by the IMD and the National Centre for Medium Range Weather Forecasting had indeed warned of impending widespread very heavy to extremely heavy rainfall, and appropriate alerts were issued to the vulnerable districts. However, inadequate preparedness coupled with injudicious operations of dam reservoirs compounded people’s misery and resulted in a heavy toll of life and on property and infrastructure.

This year’s south-west monsoon rainfall over Kerala throughout the season has generally been above the long-period average (LPA) and widespread as well, and in August it has been exceptionally high, something unprecedented. Until August 19, up to when date data compiled by the IMD is available (Table 1), Kerala had received 2,346.6 mm of rain against the LPA of 1,649.5 mm (42 per cent more). With only September remaining in the monsoon season, rainfall has been high in all the three previous months. While in June and July, the percentage departures from the LPA were only 15 and 18, the departure in August was (until August 19) a whopping 164 per cent, which is when much of the flooding and associated devastation occurred.

Distinct peaks

Figure 1(a) shows the temporal variation of daily rainfall intensity. There are clearly four distinct peaks of rainfall: two in June (around June 14 and 20), one in July (around July 20) and one sustained long one in August (beginning August 8 until August 16), the month when the normal rainfall is usually the least of all the monsoon months and on the decline. Figure 1(b) shows the cumulative rainfall over Kerala for the entire monsoon season up to August 19. Figure 2 shows the district-wise distribution up to August 19. The highest excess over the LPA was recorded at Idukki (88 per cent), followed by Palakkad (70 per cent). It was actually a fresh spell of heavy rainfall that began on August 8 and was sustained without any let-up until August 16 that was responsible for the extensive flooding and associated damage. After which, as Figure 1(a) shows, there was a gradual decrease.

As a result of the heavy rainfall until the end of July, 35 reservoirs were close to the full reservoir level. When the fresh onslaught of rain began, reservoirs had no buffer storage capacity to accommodate the fresh inflows that began after August 8. “Flooding was mainly due to the opening of the dams. Reservoirs were nearly full in July itself, and still they had two months of rain,” Rajeevan said.

Given the heavy rainfall in June and July, with a prolonged spell in mid July, and the forecast-cum-warnings issued by the IMD for June and July (which were by and large either yellow (“stay updated”) or orange (“alert”) for most districts), the appropriate management strategy would have been to begin release of water from the reservoirs in end July itself according to a well-thought-out plan and not wait until faced with an unmanageable overflow situation. When the excessive rainfall spell continued into the catchment areas, the authorities were left with no option but to release the excess water into the rivers downstream. “Unlike the major rivers in basins elsewhere, Kerala’s rivers are minor, and unlike having a few days to take appropriate action elsewhere, here the banks began to be breached within a day of the discharges, and this continuing for a week caused widespread flooding in many parts of the State,” Rajeevan added. Also, Kerala does not figure in the Central Water Commission’s (CWC) flood-forecasting network of stations (Table 2), as was the case in Jammu and Kashmir during the 2014 floods there. Water being a State subject, the CWC does not establish a station unless a State makes a request to the Centre for the same and creates an appropriate State-supported protocol for flood monitoring of reservoirs. The CWC has only flood-monitoring sites in Kerala.

Stronger monsoon flows

The chief causative factor for the occurrence of heavy rainfall over Kerala was the strengthening of monsoon flows during July and August. The frequent formation of pressure lows and depressions over the north-west Bay of Bengal, their movement over the mainland and their further intensification caused an increased flow of westerly and south-westerly winds over the Arabian Sea along and off the Kerala coast. The presence of low-pressure systems over the mainland causes an increased flow of cross-equatorial monsoon winds towards the low-pressure region. “These strong winds interacted with the hilly topography of the Western Ghats, leading to orographic uplifting of the moist air,” said Mrutyunjay Mohapatra of the IMD in an email. “This results in the formation of clouds causing heavy rainfall.”

“Orographic effects cause small rain-bearing weather systems to form. It will also give rise to small eddies, which can spiral moisture, giving rise to heavy rains,” Rajeevan pointed out. “These systems are only 10-20 km in size, which form quickly, and we do not notice them as they can be seen only by [Doppler] weather radars,” he added. The Indian Doppler Weather Radar network, which comprises 25 DWRs, unfortunately, does not have any close to the Western Ghats; the only one in Kerala is in Thiruvananthapuram.

“But all our models clearly predicted such intense heavy—more than 20 cm—rainfall days. Normally, you see this kind of rainfall along the western coast. But this year it was not like that,” Rajeevan said. Indeed, because the rainfall was mainly due to the orographic effect of the Western Ghats, the extreme events occurred in Idukki and Palakkad districts, which are far removed from the coast, and the coastal regions themselves were not affected by such heavy rainfall.

According to Mohapatra, all except the June 20 peaking of the rainfall were associated with depressions over the north-west Bay of Bengal. The June peak was due to a trough off the west coast, running from Konkan to Kerala with an embedded cyclonic circulation that extended right up to the mid tropospheric level over the south Konkan neighbourhood, he said.

On the other hand, the exceptionally heavy rainfall in August, Mohapatra said, was due to the consecutive formation (within seven days) of low-pressure systems, first on August 6 and then on August 13. The presence of the monsoon trough, whose eastern end was anchored lower than the normal position (towards the south), also facilitated the pressure lows to evolve into depressions, which then moved north-north-westwards along the trough (Figure 3). (The monsoon trough is the horizontal east-west region along the locations of relatively minimum sea-level pressure in a monsoon region. Most of the active transient disturbances producing rain, such as depressions, develop and move along the monsoon trough.) This angular tilt of the trough is the reason why there were rains in the north and north-western regions around the same time as the heavy rains in Kerala, while the north-east had deficient rainfall.

An unusual feature

“The total number of low-pressure system days over the Indian region during August 1-18 was 10,” Mohapatra said. “The most unusual feature of this year’s monsoon was the consecutive formation of two depressions over the north-west Bay of Bengal and their nearly westward movement in favourable regions during August 6-17, which led to cumulative rainfall of such intensity,” he added. “August behaviour is very typical of a positive Indian Ocean Dipole [p-IOD] year,” observed Rajeevan. “You get very frequent depressions forming over the Bay of Bengal. By the time one reaches Gujarat or thereabouts, another one will come in two to three days. The year 2006 is a very good example,” he said.

(The IOD refers to the oscillatory SST difference between the western equatorial Indian Ocean (off the eastern coast of Africa, from the northern half of Madagascar to the northern edge of Somalia) and the south-eastern Indian Ocean (off the northern coast of Australia and throughout Indonesia). A p-IOD phase is when the SST in the former increases and drops in the latter.)

But is it a p-IOD year? The long-range forecast and the subsequent short-term forecast releases as well as weekly weather bulletins issued by the IMD to date all mention neutral IOD conditions. The p-IOD phase is believed to be correlated to a good south-west monsoon. In May and June, the IMD had, in fact, talked of its own Monsoon Mission Climate Forecasting System and global models predicting the neutral IOD condition turning to a negative IOD (n-IOD). Now, however, the predictions, the IMD says, are that they will stay neutral during the remaining monsoon period.

Interestingly, however, the Japan Agency for Marine-Earth Science and Technology, which was the originator of the notion of the IOD, claimed in early August itself that p-IOD conditions had emerged at the end of July itself. The African agency the IGAD Climate Prediction and Applications Centre too said that the IOD turned positive in early August. But the Australian Bureau of Meteorology still maintains that neutral IOD conditions continue to prevail. So what is going on?

“There has been an indication of positive IOD during late July and then August. Dr R. Krishnan from the Indian Institute of Tropical Meteorology, Pune [who has extensively studied the correlations between IOD phases and occurrences of pressure lows and depressions over the Bay of Bengal], also confirmed using ocean data analysis. It may not further grow. But the August rainfall pattern is very typical of a positive IOD,” Rajeevan reiterated.

But Sulochana Gadgil, formerly of the Indian Institute of Science, Bengaluru—who evolved a slightly modified index called the Equatorial Indian Ocean Oscillation (EQUINOO), which, she feels, has a better correlation with the Indian monsoon than the IOD—has a different take on the issue. “The EQUINOO phase that was unfavourable in June has been favourable from July onwards, and it continues to be positive,” she said. “The eastern equatorial Indian Ocean has already cooled, and I expect it will be so into September as well. Good monsoon conditions should continue into September as well unless El Nino offsets the positive effect of EQUINOO. There are signs of positive El Nino developing, but atmospheric teleconnections, which are what actually affect the monsoon, are yet to respond to this positive SST anomaly developing over the Pacific. So I am keeping my fingers crossed for monsoon behaviour in September,” she added.

Ground reality

So while the IOD conundrum can remain with the meteorologists and the debate on it can continue, the ground reality in the context of the Kerala rainfall disaster is that there were pressure lows and depressions that caused widespread heavy rainfall over the State, which was unprecedented. The human tragedy of huge magnitude and the infrastructural havoc and the damage caused have been extensive and it will take quite some time for the affected regions and people to return to normalcy.

Balaji Rajagopalan, an expert in environmental engineering from the University of Colorado, put it poignantly: “Catching up on the Kerala floods both from the meteorological and hydrological aspects is quite fascinating and depressing. Depressing because it is almost the same movie we saw with the Chennai floods in 2015! Similar meteorological pattern, similar forecasts made, not taken into account in the operations of reservoirs. Thereby, having to release the reservoirs at the last minute without warning to folks living downstream, leading to disastrous flooding much more than what would have resulted from just the rainfall event. Of course, like in Chennai, human causes—haphazard urban growth, destruction of the ecology of the Western Ghats that added to this by landslides. In all this the question is, how many times should we be hit on the head by nature before we do what is sensible and obvious?”

table 1

This article is closed for comments.
Please Email the Editor