Noting down my quick analysis on what really happened during the 2018 Kerala floods.....
Immediately after the 2018 floods in Kerala, there were lot of arguments and discussions on what triggered the calamity and whether the government and the associated machinery could have done anything to avoid/ minimize the same. The major questions that were discussed are the following:
The following analysis is my attempt (based on data i could gather) to understand what really happened....
Data used : http://sldckerala.com - Daily water levels, rainfall, inflow and spill details of Idukki dam from July1st - August 28th 2018. I was not able to get the daily rainfall forecast data, the best approximation i could get was the weekly rainfall departure data (departure from expected/normal). I also tried to create a time series forecast of rainfall from 9th - 18th Aug based on data from 1st July through 8th August.
During the first week of July, the average rainfall received in Idukki was around 10mm/day and the water level in the dam was increasing at a very gradual rate of around 2.5feet/week (around 2353 feet). The second and third weeks of July saw heavy rainfall with an average of around 75mm; Idukki received highest rainfall in 17 years on July 15th (156mm). There was a sudden spike in the rate of water level increase in the dam. By the end of 3rd week, water level had touched 2384 feet; almost 2feet/day increase!! The last week of July again saw average rainfall with 1-2 days of sporadic high rainfall (average 48cm). Water level reached 2395 feet in the last week of July (1 feet/day increase) raising blue alert.....
[2395 - Blue alert, 2397 - Orange Alert , 2399 - Red alert]
This was the first time the question popped up - Should we open the dam???
At the current rate, the water level will reach the red alert level within 4-5 days if the rainfall sustains. The rainfall forecast was normal (around 15-20mm). Based on the available data, it was decided not to open the dam.
The first week of August saw very low rainfall at an average 8mm/day. Water level rose only by 1 feet in the whole week thus making it 2396 feet. The decision not to open the dam got some thumbs-up. But the question was still lingering, should we or should we not??
Things changed on 8th August. Idukki received very heavy rainfall of 130mm making the water level 2398 feet!! The question again popped up - Should we open the dam???
Let us evaluate the available data points on 8th August basis which a decision was to be made....
And it was decided to open one of the shutters of the dam on 8th August. As feared, 9th August saw very heavy rainfall of around 120mm making the water level rise to 2401 feet. If we hadn't opened on 8th August, situation would have been worse based on the 9th rainfall.
Now let us see how the spill was managed for the next 5 days to bring the water levels down. The below graph shows the same. Spill is the amount of water let out of the shutters and the inflow is the amount of water coming into the dam... From 10th Aug - 13th Aug, water was released from the dam at a controlled manner starting at 52mcm peaking to 64.8mcm on 12th Aug and then coming down to 50mcm on 13th Aug. The rainfall was medium - heavy as expected with an average of 70mm/day. By 13th Aug, the water level came down to 2396.8 feet, much below the red alert level. If this trend had continued, it would have been even decided to stop the spillage by 14th or 15th August!!
If you see the amount of spillage on 11th - 13th August, it is more than the amount of water inflow. That means the dam was releasing more water than what it was receiving. This is an important metric to consider while deciding whether the water release from dam caused flooding. During these days, there was minor floods along the Periyar stretch; but still the situation was under control.
Things took a completely different turn on 14th August. Idukki received a rainfall of 235mm; highest ever; way above any predictions ever made.... Followed by 15th August - 295mm!! The inflow into the dam quadrupled... Adding to this, Mullaperiyar dam also opened on 15th August increasing the inflow to Idukki...
Now let us look at the spill rates vs inflow rate during 14th and 15th Aug. The spill rate is never above the inflow rate for any of these dates. Meaning, the dam never threw out additional water than what it was receiving during these days.
The data given above shows how the spill was managed. The water levels touched 2402 on 18th Aug, but still the spill was in a very controlled manner; making sure it is never above the inflow.
Another interesting observation is that the dam water level on 29th Aug is 2396.6 which is the same level as on 13th Aug!!
And today, 7th Sep, the water level is 2391 feet.. same as on July 26th :) Should we or should we not?????????????????
So to conclude, let me give MY answers to the questions posed initially
Immediately after the 2018 floods in Kerala, there were lot of arguments and discussions on what triggered the calamity and whether the government and the associated machinery could have done anything to avoid/ minimize the same. The major questions that were discussed are the following:
- Whether opening of the dams triggered the floods?
- Whether the dams were opened at the right time?
The following analysis is my attempt (based on data i could gather) to understand what really happened....
Data used : http://sldckerala.com - Daily water levels, rainfall, inflow and spill details of Idukki dam from July1st - August 28th 2018. I was not able to get the daily rainfall forecast data, the best approximation i could get was the weekly rainfall departure data (departure from expected/normal). I also tried to create a time series forecast of rainfall from 9th - 18th Aug based on data from 1st July through 8th August.
During the first week of July, the average rainfall received in Idukki was around 10mm/day and the water level in the dam was increasing at a very gradual rate of around 2.5feet/week (around 2353 feet). The second and third weeks of July saw heavy rainfall with an average of around 75mm; Idukki received highest rainfall in 17 years on July 15th (156mm). There was a sudden spike in the rate of water level increase in the dam. By the end of 3rd week, water level had touched 2384 feet; almost 2feet/day increase!! The last week of July again saw average rainfall with 1-2 days of sporadic high rainfall (average 48cm). Water level reached 2395 feet in the last week of July (1 feet/day increase) raising blue alert.....
[2395 - Blue alert, 2397 - Orange Alert , 2399 - Red alert]
This was the first time the question popped up - Should we open the dam???
At the current rate, the water level will reach the red alert level within 4-5 days if the rainfall sustains. The rainfall forecast was normal (around 15-20mm). Based on the available data, it was decided not to open the dam.
The first week of August saw very low rainfall at an average 8mm/day. Water level rose only by 1 feet in the whole week thus making it 2396 feet. The decision not to open the dam got some thumbs-up. But the question was still lingering, should we or should we not??
Things changed on 8th August. Idukki received very heavy rainfall of 130mm making the water level 2398 feet!! The question again popped up - Should we open the dam???
Let us evaluate the available data points on 8th August basis which a decision was to be made....
- The current water level is 2398 feet; just 5 feet below the FRL of 2403 feet. Maximum Water level (MWL) of the dam is 2408 feet
- With heavy rainfall, water level increases at the rate of 1-2 feet/day even after full generation utilization
- There was moderate-heavy rainfall forecast for the coming week
- 1 MCM of water is equivalent to roughly 1.47 MU of power - So for every 1 MCM water lost there is an associated monetary loss
- One more day of heavy rainfall can make the water level above 2401 which is above the red alert point and way close to FRL
Now let us see how the spill was managed for the next 5 days to bring the water levels down. The below graph shows the same. Spill is the amount of water let out of the shutters and the inflow is the amount of water coming into the dam... From 10th Aug - 13th Aug, water was released from the dam at a controlled manner starting at 52mcm peaking to 64.8mcm on 12th Aug and then coming down to 50mcm on 13th Aug. The rainfall was medium - heavy as expected with an average of 70mm/day. By 13th Aug, the water level came down to 2396.8 feet, much below the red alert level. If this trend had continued, it would have been even decided to stop the spillage by 14th or 15th August!!
If you see the amount of spillage on 11th - 13th August, it is more than the amount of water inflow. That means the dam was releasing more water than what it was receiving. This is an important metric to consider while deciding whether the water release from dam caused flooding. During these days, there was minor floods along the Periyar stretch; but still the situation was under control.
Now let us look at the spill rates vs inflow rate during 14th and 15th Aug. The spill rate is never above the inflow rate for any of these dates. Meaning, the dam never threw out additional water than what it was receiving during these days.
The data given above shows how the spill was managed. The water levels touched 2402 on 18th Aug, but still the spill was in a very controlled manner; making sure it is never above the inflow.
Another interesting observation is that the dam water level on 29th Aug is 2396.6 which is the same level as on 13th Aug!!
And today, 7th Sep, the water level is 2391 feet.. same as on July 26th :) Should we or should we not?????????????????
So to conclude, let me give MY answers to the questions posed initially
- Whether opening of Idukki dam triggered the floods in Periyar? - Not really.. The spill from the dam contributed to the flood that happened from 11th - 13th Aug because the spill was more than the inflow. The dam was throwing out more water. But on 14th and 15th, when the major flooding happened, the spill was always less than the inflow. So the dam hasn't added any 'additional' water outflow to cause flooding. But one thing to note is that the dam also didn't do any flood-control which is one of the use-case of a dam. It threw out whatever water inflow it was getting on 14th - 18th Aug. If it had stored this water instead of spilling it, it would have performed the flood control role.... But Idukki dam was built for power generation and irrigation as the main purpose; flood control is an add-on. The dams in Periyar basin store only about 25% of the river's annual flow with just Idukki and Idamalayar storing 21%.
- Whether the dams were opened at the right time? - In retrospect - NO; In Real time decision making - YES! "Everything is unprecedented until it happens for the first time" "No one ever trained for an incident like that; no one warned us" - Courtesy Sully Movie
