Govt deploys AI models for flood forecasting to strengthen disaster preparedness

New Delhi, July 24 In a major technological leap toward disaster mitigation, the Ministry of Jal Shakti has revealed the deployment of Artificial Intelligence (AI) and Machine Learning (ML) models for short-range flood forecasting across India.

This initiative is spearheaded by the Smart Water Resources Modelling Organisation - Centre of Excellence (SWRMO-CoE), set up under the Central Water Commission (CWC) in September 2024.

The update was provided by Minister of State for Jal Shakti, Raj Bhushan Choudhary, in a written reply to a question in the Lok Sabha on Thursday.

The initiative is being hailed as a step forward in blending technology with environmental risk management under the government’s larger vision for water and disaster resilience. The AI-based models, developed in-house, focus on short-range flood forecasting at Level Flood Forecasting Stations through time-series forecasting techniques.

The models are supported through funding under the Development of Water Resources Information System Scheme by the Department of Water Resources, River Development and Ganga Rejuvenation. Currently, the CWC issues flood forecasts with up to 24-hour lead time at 350 stations nationwide - including 200 level and 150 inflow forecast stations - in coordination with state governments and project authorities.

To enhance preparedness and provide more time for evacuation and response, CWC has also developed basin-wise flood forecasting models capable of generating 7-day advance advisories. These are based on rainfall-runoff mathematical modelling and are disseminated through a separate platform.

The flood information is further integrated with the Common Alert Protocol (CAP), which alerts respective State Disaster Management Authorities (SDMAs).

Widening its outreach, the flood forecasts are also shared via social media platforms such as Facebook, X (formerly Twitter), and the Flood Watch India mobile app.

Disclaimer: This post has been auto-published from an agency feed without any modifications to the text and has not been reviewed by an editor