Kerala 2018: Rainfall, dams, and flood
August 2018, saw the greatest flood in the south Indian state of Kerala, due to unusually high rainfall in monsoon season. Over 483 people died, 14 are missing and at least a million people were evacuated. The event was declared a level 3 calamity by the Indian government.
In the evening of 8th August, Kerala received heavy monsoon rainfall which was 256% more than the usual rainfall it receives. 35 out of 54 dams were opened as the water went close to the overflow level over a short period of next few days.
With media flooded by different opinions and claims, there was a necessity to answer some of the basic questions that arose but was left unanswered. Questions which intrigued me were:
Why were floodgates not opened any sooner, when the first forecast released by India Meteorological Department was on 2nd August? What could have been the situation of the dams before they were released as the rain intensified and the conditions before and after the warning?
Process
It started with the data collection form State Load Dispatch Center. There in the System Statistics>Storage the data for each dam, their storage level, inflows, spills, daily rainfall, etc are being kept track of.
The dimensions I started off with was Dam names, location, storage level, and rainfall pattern.
I condensed the data into different sheets with each containing information about different dams separately.
Explorations
With all the data in hand, it was time for explorations. So think think and sketch sketch!
The x-axis contains the amount of rainfall received, with the bubbles bearing the name of the dams. On the left side, the dams would have their storage level depending on which date you are on. Could have worked, but finding anybody who would have such a long attention span would be difficult.
Similar to the first, but the rainfall was represented using bar charts. This visualization seemed very cluttered to work with.
Coding
I learned D3. It’s not hard but might need some time(depending on how lazy you are!).
Unfortunately, while working with .csv and .json files, unexpected errors occurred. By no means was my javascript reading either of the formats. So I took the other path, probably never taken. Converted the csv to json, and made a js file out of it. Improvise. Adapt. Overcome. (here is the code)
Finally
The right part contains the graph to show the rainfall level. X-axis contains the dates from 21 August to 17 July. When you hover over it you can check the respective water level of the selected dam, on the left side.
(link)
Observations
• The red lines mark the dates on which the warnings of heavy monsoon rainfall were issued by the IMD.
• If you hover onto the date 22 July for the Idukki dam, you will notice the dam already had a water level of 80%.
• In fact, most of the dams in Kerala had high water levels before the heavy rains hit Kerala in the month of August.
Could the sudden release of huge quantities of water from different dams, be the reason behind the catastrophe? In the article “Kerala floods: Poor dam management behind disasters, experts say” published on Hindustan times, Chetan Chauhan points out the different opinions expressed.
Is there a need for reforms within the dam management system of India?