BENGALURU: India is the second-most flood-prone country in the world, according to the National Remote Sensing Centre, ISRO. To better understand the scale of this risk, researchers at the Indian Institute of Science (IISc) are using statistical modelling to map entire river basins, not just individual rivers, in Peninsular India, to find out whether floods in different rivers may be connected, even if the rivers themselves are far apart. Floods currently account for nearly half of all natural disaster deaths in India. Normally, the Central Water Commission (CWC) assesses flood risk at the level of individual river basins. But IISc researchers say this approach is outdated.
Climate change is making rains shorter in duration. This means water often runs off the surface instead of soaking into the soil, increasing the chance of simultaneous floods. To capture this, the IISc team studied records from 137 streamgauge stations. They found that floods are not confined within one river basin. Even rivers that are not connected geographically showed flood events at the same time.
“Flood risk assessment studies should not be restricted to individual basins. Spatial dependencies of different rivers flooding concurrently should be investigated,” says Shailza Sharma, researcher at ICAR-CRIDA and former postdoctoral fellow at IISc’s Department of Civil Engineering. The researchers identified five groups of rivers that showed similar flood behaviour. They also found that this connectedness of floods has increased in recent years.
Their findings, published in Scientific Reports, suggest that India must look at regional flood patterns. The study also builds on earlier work by IISc and IIT Roorkee, which together created CAMELS-IND, a massive dataset covering 40 years of data (1980-2020). This database highlights water flow, rainfall, temperature, humidity, wind, soil type, dams, and human activity across catchments in India.
“Inconsistent data is hindering work in Indian catchment hydrology,” says Kanneganti Bhargav Kumar, PhD student at IISc and first author of the study. Missing values, sudden changes in flow, or uncertainty in measurements make it difficult to create reliable models. The underlying causes have not been studied in detail until now, he added.
Reservoirs add another layer of complexity. Their operations are decided by humans, which means they can either reduce or worsen floods depending on how they are managed.