T3-PhD3 Machine Learning for Mapping and Parameterising Permafrost Terrain Types

Anticipated start: Fall 2019 (flexible)

Supervisory team: Dr. Bernhard Rabus, Trevor Lantz (Univeristy of Victoria), Duane Froese (University of Alberta), Joe Melton (University of Victoria), Dr. Stephan Gruber (Carleton University)

Research description

Climate change is causing permafrost (permanently frozen regions of rock or soil) to deteriorate around the world. This will have a substantial effect on global ecology and infrastructure. As one of the largest and most northern nations in the world, Canada is uniquely positioned to experience and study this important environmental event. 

Canada’s permafrost knowledge base is more than two decades old, and in need of revision and improvement in order to understand what is taking place and how any negative impact can be prevented. NSERC has initiated a 5 year program to meet this challenge. 

Using the latest advances in satellite remote sensing technology, deep learning and statistical modelling, you will be part of a nationwide team of scientists tasked with creating a leading-edge knowledge base of Canada’s permafrost system. This system will document the current state of permafrost, predict how it will change in the future, and provide foundational data for future climate change research.


  • Research and develop state-of-the-art deep learning methods to fuse multiple remote sensor types (SAR, optical, hyper-spectral) with ground based sensors, along with existing ground surface data
  • Establish which data sources are most suitable for ground substrate and surface mapping
  • Research and develop Bayesian and other statistical techniques to robustly quantify uncertainty of deep learning models
  • Establish the reliability and suitability of these methods for use in a national permafrost knowledge-base

Your Profile

  • Passionate research interest and a background in remote sensing
  • Solid programming skills (including C++ and Python)
  • Strong scientific writing skills and high motivation to produce publishable results
  • A very good command of English, both spoken and written. Internet-based TOEFL (93 with minimum of 20 in each category), Paper-based TOEFL (580 and TWE 5), IELTS academic (overall band score of 7.0 of the academic (NOT general) test with a minimum of 6.5 in each section).

What We Offer

  • Fully-funded research for 12 months per year, starting at CAD$ 24,000/yr and increasing up to CAD$ 30,000/yr, for full concentration on research and network activities. Through the School of Engineering Science's Graduate Fellowship (GF) and/or Teaching Assistantship Program (TA), applicants generally receive an additional CAD$13,000/yr or more (this additional amount alone is sufficient to cover tuition fees).
  • Internship opportunities especially with partners in northern locations
  • A dynamic and multidisciplinary learning environment
  • State-of-the-art equipment and facilities

To submit an application: Send Dr. Bernhard Rabus (btrabus@sfu.ca) your letter of interest, CV, copy of transcript(s), and contact information for 3 references. Applications will be reviewed as they are received. All applications will be reviewed by the selection committee with respect to academic qualifications and integration within the network through the PermafrostNet lens on equity, diversity and inclusion. The position will remain open until filled. We thank all applicants for their interest, however, only those individuals selected for an interview will be contacted.