PhD opportunity: Towards a machine learning driven framework for quantifying regional scale rock glacier change in mountain regions
We are looking for a PhD student in to join our team at the Waterways Centre, based at the University of Canterbury, New Zealand.
The successful candidate for this PhD position (three years, full-time position) may be funded either by a GRI scholarship or a University of Canterbury scholarship, and will work on a project entitled “Towards a machine learning driven framework for quantifying regional scale rock glacier change in mountain regions”. The successful candidate will work on further developing a robust and transferable workflow that is capable of mapping rock glaciers from globally applicable satellite data. The successful candidate will work with the Waterways Centre team, colleagues in the School of Earth and Environment at the University of Canterbury, as well as with Associate Professor Ben Robson at the University of Bergen, Norway. Research stay(s) in Norway is possible to facilitate collaboration.
Main supervisor: Shelley MacDonell
Email: shelley.macdonell [at] canterbury.ac.nz
Project description: Snow, ice and permafrost are in a state of rapid change, directly impacting water resources, natural hazard occurrence and habitat availability in mountainous regions. While glacial and nival variability has been relatively well constrained using Earth Observation (EO) data, large uncertainties still remain in the state of mountain permafrost, including features such as rock glaciers. Recent advances in machine learning and computer vision offer new opportunities to automate detection and monitoring of rock glaciers over larger scales. The aim of this PhD thesis will be to develop a machine learning framework to reliably map rock glaciers in different environments at regional scales using EO datasets from Aotearoa New Zealand, the Andes, and Norway. The framework can work towards looking at changes over time. The successful candidate will be encouraged to explore geospatial portals such as Google Earth Engine or the Microsoft Planetary Computer as well to develop open-source scripts and routines.
The ideal candidate will have the following skills: Experience with programming, in particular machine learning libraries such as Tensorflow or Pytorch; Geospatial analysis including GIS and remote sensing is essential; Familiarity with methods such as differential radar interferometry (DInSAR), time series analysis, and topographic analysis are an advantage.