Research Projects

This project aims to develop a scalable, interpretable, and physics-informed artificial intelligence (PI-AI) framework for hydrological forecasting in the river basins of Peninsular India. By embedding governing hydrological laws—such as mass conservation and Darcy’s law—into deep learning architectures, the model will overcome the limitations of conventional approaches that rely solely on either physical simulation or data-driven methods. The framework will utilise multi-source datasets, including satellite observations (GRACE, CHIRPS), ground-based hydrometeorological measurements, and geospatial basin characteristics, to simulate streamflow and groundwater dynamics under monsoon-driven variability and data scarcity. Innovative components encompass spatio-temporal neural network architectures, uncertainty quantification through Bayesian inference, and domain adaptation strategies to facilitate model transferability across diverse basins. The forecasts will be implemented via a collaboratively developed Decision Support System (DSS) comprising interactive dashboards and scenario simulation tools, customised to meet the requirements of water managers, planners, and farmer cooperatives. By integrating scientific rigour with operational applicability, the project advances national priorities concerning water security and climate resilience, aligning with initiatives such as Jal Shakti Abhiyan and Viksit Bharat 2047. The ultimate deliverables comprise open-access forecasting tools, reproducible model codebases, and regionally adaptive insights to support sustainable water resource management throughout the Peninsular India region.