- julien.vadnais@uib.no
- Visitor address
-
Allégaten 41
5007 Bergen - Room: 2G12c - 2148
- Postal address
-
P.O. Box 7803NO-5020 BergenNorway
Research groups
Research
I am a PhD candidate with a background in Geomatics and Remote Sensing.
In my current research, I develop geospatial deep learning algorithms to map marine oil slicks from satellite remote sensing, primarily SAR. Progress on this notoriously challenging task will enable new analyses of natural oil seepage systems which have important environmental implications at the basin scale (e.g. hydrocarbon migration, geological storage).
Outreach
I gave an interview at Les Années Lumière, the science radio programme of the Canadian public broadcaster, to speak about our recent discovery of an episodic seep.
I regularly take part in scientific conferences. In 2026, I will present new findings at IGARSS and the ISPRS congress.
Teaching
Universitetet i Bergen
Teaching Assistant
- GEOV205: Geographic Information Systems - Theory and Practice
- GEOV316: Practical Skills in Remote Sensing and Spatial analysis
Université de Montréal
Teaching Assistant
- GEO1542: Introduction to remote sensing
- GEO3015: Geographic information systems III
- GEO3532: Spatial modeling
- GEO6341: Complex socio-ecological systems modeling
- GEO6343: Advanced remote sensing
Publications
Academic article
- Julien Vadnais; Christian Haug Eide; Nicolas Pinet et al. (2026). Origin and fate of an episodic oil seep, Gulf of St. Lawrence (Canada). (external link)
- Julien Vadnais; Benjamin Aubrey Robson; Christian Haug Eide et al. (2025). Transfer learning between Sentinel-1 acquisition modes enhances the few-shot segmentation of natural oil slicks in the Arctic. (external link)
Conference lecture
- Julien Vadnais; Christian Haug Eide; Benjamin Aubrey Robson (2025). Advanced geomatics and machine learning to map and study spatio-temporal patterns of natural hydrocarbon seepage in the North Sea. (external link)
- Julien Vadnais; Benjamin Aubrey Robson; Christian Haug Eide (2025). On the Challenge of Oil Slick Lookalikes in SAR mapping: Transfer Learning With Limited Labels. (external link)