IIT-B to deploy urban flood monitoring system in city this monsoon
To this end, at least nine ‘reflection-based flood sensors’ will be installed at key locations in the city, such as Bandra-Kurla Complex and Hiranandani Complex in Powai, to collect empirical data on waterlogging. Based on the success of the project in its first year, the network of sensors can be expanded to include dozens more across Mumbai
Mumbai: Researchers at IIT-Bombay’s Interdisciplinary Programme in Climate Studies (IDPCS) will, during the upcoming monsoon season, pilot a real-time flood monitoring and forecasting system for the city. Titled ‘Urban Flood Risk Map: Monitoring and Modelling’, the project aims to provide citizens with a granular, locality-wise picture of waterlogging, similar to instantaneously checking the air quality index (AQI) of a particular neighbourhood at the touch of a button.
“In 2005 and 2017, the city saw debilitating floods. While the civic administration’s overall flood management has improved, Mumbaiites still do not have any actionable information which they can use to make important, or even life-saving decisions, like leaving their vehicles at home or shutting down schools. We see that people tend to rely a lot on FM radio as a way to know which parts of the city are waterlogged,” said Subimal Ghosh, institute chair professor, dept. of civil engineering, IIT-B and convener, IDPCS. “Our aim is to generate more reliable data in real-time, which citizens will be able to access via a dedicated web-portal.”
To this end, at least nine ‘reflection-based flood sensors’ will be installed at key locations in the city, such as Bandra-Kurla Complex and Hiranandani Complex in Powai, to collect empirical data on waterlogging. Based on the success of the project in its first year, the network of sensors can be expanded to include dozens more across Mumbai. Citizens will also be asked to share anecdotal information on social media using the hashtag #MUMBAIFLOODDATA, along with geo-tagged pictures.
“For example, if someone says there is ‘waist-high’ or ‘ankle-deep’ water in a particular area, we can ascribe a numerical value to these general terms, and using rainfall information and a digital elevation map of the city, tell people with reasonable certainty how risky it will be to venture there. We can identify which other nearby areas may be waterlogged. Using data on drainage and hydrological models, we can estimate how soon the waters will recede. There are many possibilities we are trying to explore,” added Ghosh, who is also a lead author (Working Group 1) of the recently concluded Sixth Assessment Report of the Intergovernmental Panel on Climate Change.
A second facet of this project deals with higher-resolution forecasting, in that it will use macro-level rainfall warnings for Mumbai, issued by the India Meteorological Department (IMD) and translate those into area-level warnings using artificial intelligence and machine learning models, so people will know ahead of time which areas to avoid during periods of heavy rain. Considering the spatial variability of rainfall across the city, which has been well documented in scientific literature, researchers said that such ‘downscaled’, hyper-local warnings are the need of the hour.
“It doesn’t have to be a particularly wet monsoon for flooding to occur. Even during a deficit season, there is a likelihood of isolated extreme weather events, which have been increasing in frequency. When the IMD issues a ‘red alert’ for the city, it doesn’t tell people living alongside the banks of the Mithi River whether or not they need to evacuate their homes. It doesn’t tell a hospital located in a flood-prone area whether patients need to be evacuated,” said Raghu Murtugudde, visiting professor, IIT-B and professor of earth system science at the University of Maryland.
Murtugudde added, “We hope to downscale the official forecasts to a street-level resolution so that such decisions can be taken in real-time. For someone living along the banks of the Mithi River, ankle-deep water has a very different meaning than it does for someone living in a high-rise in BKC.
Researchers revealed they are still scoping out suitable locations for the deployment of flood sensors. “We are hoping to install them at places like hospitals, schools, arterial traffic junctions, and so on. They are small devices which can be affixed to a lamppost and require minimal power. Our larger concern is vandalism, which is why we are looking for places where the devices will be safe. We are hoping to start with placing them in the Mithi River’s catchment, which will give us good data to extrapolate from,” Ghosh explained.