Official title: Modélisation spatio-temporelle de la recharge des eaux souterraines et des étiages sous l’effet des changements climatiques à l’aide d’un réseau de neurones graphiques informé par la physique
Objectives: The general objective of this project is to develop a physics-informed graph neural network for modelling the spatial and temporal dynamics of groundwater recharge and river low flows under climate change. The proposed framework will represent connections among climate, groundwater systems, watersheds, and river networks while respecting fundamental hydrological principles. Historical observations, hydrogeological information, and climate projections will be integrated to identify the principal controls on recharge and low-flow variability. The project will assess how the magnitude, timing, duration, and spatial distribution of these processes may change under future climate conditions. It will also identify regions that may become increasingly vulnerable to groundwater-recharge deficits and reduced streamflow. The resulting modelling framework and maps will support climate-adaptation planning and sustainable water-resource management in Québec.
- Project PI(s): Rahim Barzegar (PI, UQAT), Eric Rosa (Co-PI, UQAT), Vincent Cloutier (Co-PI, UQAT), Jan Adamowski (Co-PI, McGill University)
- Project partner: Ouranos
- Funding: Ouranos - MELCCFP QClim’Eau Program
- Period: 2026/01 to 2027/12