Applications of machine learning to improve hydrological model regionalisation

Hydrological streamflow models are commonly used to warn populations of extreme events. The parameters of these models are typically calibrated to observed streamflow. The learned parameters can then be transferred to areas that are lacking streamflow observations, a process referred to as regionalisation. This project builds on previous work, which proposed using machine learning models to improve methods for regionalisation. Specifically, the project intents to expand previous work from local case studies to a large scale national model. Furthermore, the project aims to advance the machine learning component, exploring different machine learning algorithms and configurations.

Faculty Supervisor:

Usman Khan

Student:

Partner:

Aix-Marseille Université

Discipline:

Engineering

Sector:

Education

University:

York University

Program: