New Zealand’s braided rivers are among the most morphologically active river systems globally, continually adjusting their form in response to sediment supply, floods and wider landscape disturbance. These rivers support highly valued ecosystems and cultural landscapes, but their dynamism also creates major challenges for flood-risk management, infrastructure planning and river restoration. Living safely and sustainably alongside braided rivers requires the ability to understand and predict their response to change, including river engineering, climate-driven changes in flood frequency and magnitude, and major natural disturbances such as earthquakes.
Existing physics-based morphodynamic models can, in principle, support these predictions. Applying such models with sufficient resolution at landscape scale is costly and slow, constraining forecasts that can support decision-making and management.
This project will develop a new data-driven modelling strategy to predict the morphological evolution of braided rivers. The approach will use machine learning to emulate outputs from physics-based simulations, creating models that retain key predictive insights while running orders of magnitude faster than conventional approaches. By enabling rapid exploration of future scenarios across a range of conditions, these models will support timely, robust decisions about braided river management, flood hazard reduction and long-term adaptation.
Relevant model reduction approaches include POD, dynamic mode decomposition and Fourier analysis. Extensive numerical simulation, with supporting analytical work, is the core element of this project. Real data sets will be used to validate the tools developed in the thesis.
This project is linked to the UC Commitment: To Strengthen Societal & Planetary Resilience
This project will be cosupervised by Dr Gu Stecca, Earth Sciences NZ.
Start date: as soon as possible.
Supervisors
Primary Supervisor: Prof Rua Murray
Other Supervisor(s): Prof James Brasington
Key qualifications and skills
Mathematics or Science student with applied mathematics to at least third year level; graduate level coursework preferred. The student must have programming capability, including Python (or equivalent). An engineering or computer science background, containing a significant amount mathematics, may be suitable.
Does the project come with funding
Per annum: Stipend of $32,650 plus PhD tuition fees. The project also has up to $5,000 available for research expenses across the duration of the degree.
Funding duration: Three years (360 points)
Final date for receiving applications
27 October 2026
How to apply
This is a UC Commitments Doctoral Scholarship and applications must be made through the Scholarship Portal in your myUC account. Emailed applications will not be considered. Find out more here: University of Canterbury Scholarship Portal - UC Commitments - Developing data-driven models of braided river morphodynamics
Keywords
Applied mathematics; fluid dynamics; sediment transport; scientific computation; model reduction; morphodynamics; geospatial data analysis; braided rivers.
