Research project title
Physically interpretable AI emulator for hydrological extremes
Education level
Doctorate
Director/co-director
Director: Julie Carreau
Co-director(s): Ali Améli
End of display
August 31, 2025
Areas of expertise
Primary sphere of excellence in research
Energy, Water and, Resources
Unit(s) and department(s)
Department of Mathematical and Industrial Engineering
Department of Earth, Ocean and Atmospheric Sciences
Conditions
-
Master’s degree in Climate Science, Hydrology, Applied Mathematics, Computer Science, or a related field.
-
Strong background in deep learning, particularly probabilistic models and recent deep learning architectures.
-
Knowledge in hydrological modeling and climate projection data.
-
Experience with AI/machine learning techniques applied to environmental data.
-
Proficiency in programming (Python) and the use of high-performance computing (HPC) infrastructures.
-
Excellent written and oral communication skills for collaboration and reporting.
-
Familiarity with extreme events prediction, physical system constraints, and ensemble simulation methodologies is a plus.
Detailed description
This PhD project offers a unique opportunity to contribute either to the advancement of deep learning methodologies or to hydrological impact studies, depending on the candidate's expertise and interests. The focus is on developing physically-coherent deep learning (DL) emulators that can downscale low-resolution climate projections to high-resolution outputs. These emulators will ensure physical consistency between key meteorological variables (e.g., precipitation, temperature) and improve their interpretability for practical applications.
From a deep learning perspective, this project aims to address challenges in uncertainty quantification and the integration of physical constraints into DL emulators, offering the potential to work on cutting-edge techniques in AI applied to environmental systems. Alternatively, from a hydrological impact studies perspective, the project aims to assess climate change's impacts on small watersheds using emulated meteorological variables, with a particular focus on streamflow prediction and extreme events such as flooding.
This interdisciplinary project has far-reaching implications for both fields, contributing to better climate adaptation strategies and enhanced hydrological risk assessments.
Financing possibility
The successful candidate will work under the joint supervision of Professors Julie Carreau from Polytechnique Montreal (Montreal, Canada) and Ali Ameli from the University of British Columbia (Vancouver, Canada). The mathematical unit within the Mathematics and Industrial Engineering Dept. of Polytechnique Montreal, is very active in AI and optimization and works on applications of mathematical modelling in climate and environment. The Dept. of Earth, Ocean, and Atmospheric Sciences at the University of British Columbia formed a committee to address the climate emergency, with A. Ameli’s research group leading efforts on watersheds’ role in regulating climate change impacts. The PhD student will benefit from the research environments of both Professors.
We are committed to fostering equality, diversity, and inclusion within our team. We strongly encourage applications from all underrepresented groups, including visible minorities, women, Indigenous peoples, persons with disabilities, and individuals of any gender identity.
Start date and duration: The PhD student is expected to begin in Fall 2025. The candidate has the choice to be located either in Montreal or Vancouver.
To apply, please send an email to either julie.carreau@polymtl.ca or aameli@eoas.ubc.ca with a motivation letter, CV and official transcripts.
