How GIS helped Kerala: Flood Response and Disaster Management

This August, Kerala experienced one of its worst floods in nearly a century, causing unprecedented damage to life and property. Statistically, ~12% of the total geographical area of India is flood prone, and flooding in India accounts for ~20% of global deaths with a death toll of over 100,000 people in the past 64 years. The recent floods have scarred Kerala with more than 400 deaths and infrastructure damages of over Rs. 20,000 crores (200000 million), displacing the lives of over 700,000 people; and it will take time to reach a state of normalcy. In this time of need, everyone must play a part in helping those that were affected, in whatever way possible. We at GeoSpoc, collaborated with our partner SatSure to bring about the power of geospatial intelligence to assist rescue efforts in Kerala by analysing the extent of flood damage.

How did we try to help?

We used satellite data to visualise the extent, magnitude, and height of flooding. These floods were a result of heavy rainfall (~2,377 mm) which meant that the area was under a dense cloud cover. We needed satellite imagery that could look past the cloud cover giving us a clear picture of the conditions on the land. Sentinel 1A and 1B satellites carry a synthetic-aperture radar (SAR) instrument that can collect data in any weather, and in the day or night. In other words, data from Sentinel, which is freely available and could penetrate through cloud cover, aligned with our requirements. SatSure processed the Sentinel data to generate Digital Elevation Models (DEMs) of the area, using which we created an interactive web map equipped to intelligently display such a heavy dataset. You can access the map here.

This map was presented to the Government of Kerala. They wanted us to go further and perform simple spatial analysis to divide Kerala into zones of high, medium, and low flood risk. Combining our spatial data with weather and administrative data, we delineated Kerala for a greater understanding of the floods. On top of this, we shared a series of reports to further aid the government’s disaster relief efforts. Our spatial intelligence not only helped the government, but also a private insurance company. The insurance company used our data and analysis to figure out which of their insured properties were affected by floods, where these were located, and who they belonged to. With this, they had an advanced information about where a potential insurance claim could come from. This allowed field offices to direct their agents to affected customers instead of waiting for a claim to be filed.

It was important for this data and analysis to be open to the public, and not just the authorities. Considering the disturbed telecommunications in a disaster-stricken area, we wanted to make the information as easily accessible and quickly loadable as possible, in a short span of time. To this end, open source tools were used to analyse open source data as these were readily available for our requirements. Everything was prepared to be visualized on a web map so that important information such as the location of hospitals, rescue efforts, flood extent, could be accessed from anywhere.

kerala flood map

Not Just Kerala…

We also applied the same methodology in the wake of Cyclone Titli on the eastern coast this October. The tropical cyclone from the Bay of Bengal is categorised as a Category 2 hurricane on the Saffir-Simpson scale (SSHWS) and is a Very Severe Cyclonic Storm. Titli has primarily affected the states of Odisha and Andhra Pradesh, causing landslides and heavy flooding. We mapped the Ganjam district of Odisha and Srikakulum district of Andhra Pradesh. Using Sentinel SAR imagery of October, we mapped the flood extent for 10th and 17th October and visualised it on our web map. Along with this we also mapped the potential crop damage to rice against the actual acreage. The difference from mapping Kerala was that the cyclone and consequent floods are ongoing and our focus was to figure out where the flood was and wasn’t. We did not map the flood height as we did not have data during the maximum flood, but from a few days before and after. You can see the map here.

How do we fare compared to the UK?

The UK is also currently experiencing what has been called its worst flooding in 30 years. However, Storm Callum of has caused three deaths so far; strikingly less than the estimated death toll caused by the Kerala floods which is somewhere above 400. With a loss of such magnitude, we need to think how we can reduce the damages caused by such disasters. An advantage the UK has over us, is its comparatively monotonous geography, which makes it easier to plan and implement flood management practices. The geographical diversity of India coupled with its vast extent adds a complication to implementing robust and widespread flood management practices. An administrative complexity is that flood management schemes are state-controlled and funding for these are dependent on the priority of the particular state. However, flooding extent of a river will not be limited to a state and its management practices.

Most of the havoc caused by flooding can be averted by smart town planning and creation of flood defence structures based on hydraulic modelling. A river (hydraulic) model is a mathematical model of a water/sewer/storm system which represents the behaviour of the river channel and floodplain. An understanding of the hydraulic behaviour of a river system aids in planning and creating infrastructure for its management. This model requires data on the river flow, catchment topography, and calibration data to extrapolate and predict flood frequency and magnitude that can be used for flood management practices. Hydraulic modelling (2D, or 3D) is where we come in. Geographic Information Systems can help mitigate the effect of floods. However, owing to complexities outlined before, availability of such extensive data for modelling is difficult in India.

UK has an extensive network of gauging stations – currently comprising around 1500 flow-measurement stations augmented by a substantial number of secondary and temporary monitoring sites – that routinely measure flow data. Similarly, the Central Water Commission (CWC) of India has set up ~ 901 stations responsible for the collection of hydrological, meteorological, and water quality parameters data across India. For a country that is 13 times larger than the UK, there is a lot of untapped fluvial data waiting to be captured; data which could significantly alter (for the better) our flood management practices. CWC launched the National Hydrology Project (NHP) in 2016-17 with the aim of increasing availability, accuracy, and accessibility of hydrological data across India. This could potentially push us forward in disaster management.

Flood analysis and consequent policies can only be as accurate as the raw, ground data. Public participation in gathering and collecting data in a large country like ours could have the potential of creating an extensive hydrological network. Disaster management is not just the responsibility of the administrative bodies and data on our natural systems (such as rivers) should also be available to the common public and not limited to the experts, scientists, or the administration. This data transparency can foster greater public awareness and maybe even responsibility. However, capturing and accessing the data is not enough. What is even more important is making sense out of such a big data. This can be achieved through the use geospatial analysis.

Floods are not something out of our control. A general rule of thumb with natural disasters is that higher magnitude events occur less frequently, and vice versa. This helps in estimating and predicting, based on flooding history, when a flood event of a particular magnitude is expected to recur. The key cause for urban flooding is being unable to manage excess surface runoff due to impervious surfaces and insufficient drainage. This can easily be prevented with smarter town planning keeping flood management in mind. Why leave something that can affect us so drastically to the administrative bodies? Even we, as individuals, can significantly contribute to reducing the impact of urban flooding. All it takes is awareness and small changes to our practices of water management. Follow this link to see how you can help –


Nishil Agrawal | GIS Analyst, GeoSpoc

Leave a Reply

Your email address will not be published. Required fields are marked *