Climate AI Nordics

Climate AI Nordics

Nordic network for collaborations around tackling climate change using AI and machine learning.

PhD Fellow in Machine Learning for Generative Emulators of the Arctic Ocean

This 3-year PhD fellow position at the Nansen Environmental and Remote Sensing Center (NERSC) is connected to the Data Assimilation group and conducted in collaboration with the Sea Ice and Ocean Modelling groups. The project, titled “The reliability of generative emulators of the Arctic Ocean,” aims to investigate how effectively the Arctic Ocean observing system can support building data-driven machine learning forecasts.

The PhD fellow will be responsible for designing the architecture and training a generative emulator of the coupled ice-ocean-biogeochemical model of the Arctic Ocean. Key tasks include assessing the emulator’s reliability—specifically whether predicted error bars encompass actual errors—and integrating these models into ensemble data assimilation for operational forecasting. The researcher will utilize the TOPAZ system, the neXtSIM sea ice model, the Ensemble Kalman Filter (EnKF), and remote sensing data.

The position is based in Bergen, Norway, at the Nansen Center, a recognized environmental research institute at the forefront of climate and environmental research. The PhD study is linked to the University of Bergen (UiB), Faculty of Mathematics and Natural Sciences, Department of Geophysics. This international environment employs around 75 people from more than 20 countries and maintains a strong commitment to work-life balance and diversity.

Key Qualifications:

  • A Master’s degree or equivalent in data science, applied mathematics, machine learning, informatics, or a related field.
  • Eligibility for registration as a PhD candidate at the University of Bergen.
  • Fluency in both spoken and written English.
  • Knowledge of physical oceanography is considered a plus.
  • Strong personal skills, including being self-motivated, structured, and having the ability to work efficiently in an international team.

Deadline: April 30th, 2026

Apply through the official recruitment system

You May Also Enjoy

Climate AI Nordics Newsletter, April 2026

Published:

The April 2026 Climate AI Nordics newsletter highlights the SafeCityLearn benchmark and researcher Miguel Nobre da Costa, while sharing a wealth of upcoming events, webinars, and job opportunities at the intersection of AI and climate action across the Nordic region.

HydroImaging: Mining Imaging Data for Hydrological and Environmental Modelling

Event date:

Submit your research to HydroImaging, a half-day IEEE ICIP 2026 workshop in Tampere, Finland (September 13–17, 2026). This workshop bridges computer vision, remote sensing, and environmental science to address climate change and the water cycle. Contributions on data-centric ML, multi-modal fusion, and disaster mapping are welcome.

Enhanced Flood Detection through Innovative integration of PolSAR, metaheuristic optimization, and deep learning-based segmentation

Event date:

Webinar with Solmaz Khazaei, KTH Royal Institute of Technology. Flood is the most common natural disaster in the world, and can have catastrophic impacts on human society and the environment, including infrastructure damage, agricultural losses, and casualties, resulting in widespread economic and social disruptions. In early studies, water body detection relied on on-the-spot investigation, hydrological models and common remote sensing techniques that face issues like slow processing and real-time delays. By addressing this challenges we propose a novel hybrid PoLSAR-metaheuristic-DL models and high-resolution remote sensing data to generate accurate and rapid flood mapping for one of the huge recent flood in France. Compared with standard synthetic aperture radars (SAR), polarimetric synthetic aperture radar (PolSAR) is an advanced technique of SAR remote sensing. So, by using polarimetric decomposition methods, features were extracted and feature selection problem, one of the most challenging, was solved by using metaheuristic techniques. The selected features fed into three deep learning-based segmentation models- U_Net_V3, Nested_UNet and Efficient_UNet. The reliability of the generated flood maps was evaluated using Accuracy, precision and recall metrics. Our experimental results indicate that Nested_UNet integrate with optimized PolSAR data achieves the highest segmentation performance, with an accuracy of 0.910, precision of 0.914, and recall of 0.909. These findings underscore the capability of Nested_UNet, demonstrates superior feature extraction abilities, making it a promising choice for real-time flood segmentation applications. Moreover, detecting the knowledge of flooded areas, officials can actively adopt steps to reduce the potential impact of flood, ensure the sustainable management of natural resources and mitigate flood impacts.