Adjusting to climate How Kerala got it right in 2022

Entry type: Single project

Country/area: India

Publishing organisation: The Hindu

Organisation size: Big

Publication date: 2022-08-29

Language: English

Authors: Vignesh Radhakrishnan,
Sonikka Loganathan ,
Pratap Vardhan.


The members of the The Hindu Data Team which did this project takes care of the data needs of The Hindu, India’s leading national daily.

Project description:

In Kerala, India, it rarely rains in August. Rainfall usually peaks between June and July before tapering off by August—but patterns have changed due to climate change. Unpredictable rainfall is a headache for people who manage dams. In 2018, Kerala faced one such August, with copious amounts of rain coming down in a short span. Dams overflowed and close to 500 people died. But in 2022, despite facing a similar August, dams did not overflow, and there were hardly any deaths. What changed? Using data we explain how Kerala managed its dams better this time around and escaped another tragedy.

Impact reached:

To understand better how Kerala did it, we interacted with the manager of one of the biggest dams in the State, in the city called Idukki. We converted the conversation into a podcast for our data based podcast channel: Data Point (https://open.spotify.com/episode/2R3OfndcFDa6oQPwEovHUT). The “Data and dams: How Kerala cracked flood management” remains one of the most listened to podcasts of last year.

Data about dam maintenance, with all the technical jargons broken down, is a rarity in Indian journalism. There are various factors involved: amount of rainfall, inflows into the dam, storage capacity of the reservoir, outflows and spills from the dam. If there are two years which recorded very high amounts of rainfall in a month when it was least expected, and only one of them led to a disaster, it is important to understand how it was managed.

Given that climate change is a clear and present danger, the world is moving towards adaptation. One of the primary goals of the COP26 summit was adaptation. The goal describes adaptation as follows: “building defenses, warning systems and resilient infrastructure and agriculture to avoid loss of homes, livelihoods and even lives.” Kerala’s successful dam management falls under this category given that climate change has shifted the rainfall patterns in the State.

This piece, using a data-based approach, has given a new way to analyze and understand dam management instead of relying on information supplied by those involved in the work. It gives a new tool for journalists and researchers to come up with bias-free conclusions about a subject which is hard to understand and impacts thousands of lives during excessive rainfall — a phenomenon which is only going to increase in the coming days.

Techniques/technologies used:

Kerala’s statistics about dam management were available in https://sldckerala.com/index.php?id=7
It gives the data in a daily format. So we used python to scrape the data out. The program was designed in such a way that it will load all the dates and scrape the data out and give the final numbers in a CSV file.

Now more than scraping, the bigger challenge was to understand what the column headers stood for and the unit as like most Indian datasets, this one too did not have any meta data.

For instance, the website’s table has a column named “spill”. Nowhere it is explained, what spill means. So most of the data was in dam related jargon, if we are not able to decipher it, it is at the end of the day a bunch of meaningless numbers.

So we started calling dam managers in Kerala to understand what spill means, what is minimum drawdown level, how are inflows and outflows measured and was able to make sense of the numbers which led to the story.

Context about the project:

The biggest problem with the project is the jargon that is used by those who are responsible for managing a dam. Hundreds of people die in India due to poorly managed dams that leads to floods in times of excessive rainfall. So, there is a need to fix accountability.

However, the experts themselves will not explain the issues behind, given that they are involved in managing them in the first place. Because if mistakes emerge, they may have to face the consequences.

So, mostly we need to learn the jargons involved and the dam management techniques used, by ourselves. And so, we did that over many years. Yes, this is not the first time we attempted to understand the workings of a dam.

In 2015, when parts of Chennai city in Tamil Nadu sank, we experimented with dam related data for the first time. Using parameters such as inflows and outflows, we explained how the tragedy was a result of a “wrong call” made by persons incharge. Please refer to attachment “Chennai 2015” for details .

In 2018, when Kerala sank, we again dabbled with dam data and this time we added more details such as spills and outflows to enhance the understanding of how dams were mismanaged. Refer to “Kerala 2018” attachment for details.

In 2022, it came a full circle when we were able to compare two years — 2018 and 2022 — and able to come to a conclusion regarding what worked and what did not. Refer to “Kerala 2022” attachment for details.

What can other journalists learn from this project?

Following a topic with persistance, however tough it may be, will yield results. And if we crack it once, multiple stories will follow. Also, having a complete understanding of what all the columns represent is important before we analyse a dataset. Reaching out to experts and having a detailed understading of the underlying data is important.

Project links: