Climate AI Nordics

Climate AI Nordics

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

Master’s thesis projects in AI for Earth observation, ecology, and bioacoustics

The RIDR (RISE Deep Learning Research) group is offering several exciting Master’s thesis opportunities for students passionate about applying artificial intelligence to real-world ecological challenges. These projects offer a chance to work with unique datasets, collaborate with interdisciplinary teams, and contribute to impactful research with the potential for publication.

Successful completion of a project will be rewarded with SEK 30,000 per student. We encourage teams of two students to apply for each project.

AI-based species distribution modeling of predators in scandinavia using remote sensing

This project aims to develop robust species distribution models for large predators like bears, lynx, and wolves in Scandinavia. By combining predator observation data from ROVBASE with high-resolution remote sensing imagery, you will use AI-based methods to understand how environmental factors influence their distribution. This work, in collaboration with Norwegian and Swedish authorities, will provide valuable insights for wildlife management and conservation.

  • Location: Gothenburg
  • Supervision: Olof Mogren, Aleksis Pirinen

Multi-modal machine listening for bioacoustic analysis and modelling

Centered around the unique AukLab-Audio dataset from the Stora Karlsö Auk Lab, this project focuses on developing machine listening models that leverage multi-modal data. You will explore how visual cues, weather data, and other sensor information can be used to train and contextualize acoustic models of common guillemot vocalizations. The Stora Karlsö Auk Lab is a unique cliff-edge observatory for common guillemots (Uria aalge) in the Baltic Sea. This research contributes to a deeper understanding of this seabird’s vocal behavior.

  • Location: Gothenburg
  • Supervision: Olof Mogren, John Martinsson

Data-centric AI for acoustic discovery: anomaly detection and active learning in bioacoustics

Also utilizing the rich AukLab-Audio dataset, this project takes a data-centric approach to uncover new acoustic patterns in guillemot vocalizations. You will develop anomaly detection algorithms to identify rare sound events and implement active learning strategies to guide the annotation process, optimizing human labeling efforts and expanding our knowledge of this species’ acoustic repertoire.

  • Location: Gothenburg
  • Supervision: John Martinsson, Olof Mogren

Machine learning and Earth observation data for monitoring nature restoration

Based in Lund, this project focuses on using machine learning and satellite data to monitor the restoration of Swedish pastures and semi-natural grasslands, in line with the EU’s new nature restoration law. You will develop and train machine learning models to process multi-year time series of satellite imagery and generate reliable signals about on-the-ground restoration efforts. The project involves key stakeholders such as Arla, the Swedish Board of Agriculture, and the Swedish Environmental Protection Agency.

  • Location: Lund
  • Supervision: Aleksis Pirinen, Delia Fano Yela, Georg Andersson

For more details on these opportunities, including application instructions and deadlines, please visit the links below:

Supervision for all projects is provided by the RIDR group.

You May Also Enjoy

Self-supervised pre-training for glacier calving front extraction from synthetic aperture radar imagery

Event date:

Webinar with Nora Gourmelon, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU). The factors influencing melt at glacier fronts facing the ocean remain an active area of research. To better understand these factors, it is essential to monitor changes in glacier calving fronts. Due to the importance of weather-dependent and seasonal variations, Synthetic Aperture Radar (SAR) is the preferred imaging method for such monitoring. In recent years, deep learning models have been developed to automate the extraction of calving front positions, however, their performance on SAR data remains suboptimal. Limited labeled data and high variability in SAR images hinder traditional supervised learning. Foundation models pre-trained on large, diverse datasets could provide robust feature representations that require minimal labeled data for fine-tuning on specific SAR tasks. However, in preliminary experiments, we found that the domain gap is too large for the task of extracting calving fronts from SAR imagery, and foundation models fine-tuned on the "CAlving Fronts and where to Find thEm" (CaFFe) benchmark dataset performed subpar. Therefore, we compiled an unlabeled dataset of Sentinel-1 SAR image sequences of Arctic glaciers, each associated with a single Sentinel-2 optical reference image. Using this dataset, we developed a novel multi-modal self-supervised pre-training strategy and applied it to pre-train a hybrid CNN-transformer model. Fine-tuning on the CaFFe benchmark showed that the pre-trained model outperforms its non-pre-trained counterpart, enabling more robust calving front segmentation and demonstrating the potential of data-efficient learning for SAR imagery.

Climate AI Nordics Newsletter, September 2025

Published:

Welcome to the September edition of the Climate AI Nordics Newsletter! This month’s issue highlights upcoming events on Earth observation, AI, and climate research, along with our featured member, Mohammad Kakooei, whose work advances scalable AI–Earth Observation for global sustainability. You’ll also find recent community highlights and new job opportunities across the Nordic region.

Earth observation and deep learning for urban applications

Event date:

Webinar with Sebastian Hafner, RISE. Urbanization is progressing at an unprecedented rate in many places around the world, making Earth observation (EO) a vital tool for monitoring its dynamics on a global scale. Modern satellite missions provide new opportunities for urban mapping and change detection (CD) through high-resolution imagery. At the same time, EO data analysis has advanced from traditional machine learning approaches to deep learning (DL), particularly Convolutional Neural Networks (ConvNets). Yet, current DL methods for urban mapping and CD face key challenges, including effectively integrating multi-modal EO data, reducing reliance on large labeled datasets, and improving transferability across geographic regions. This talk will present methods to address these challenges in urban mapping and CD, with applications such as multi-hazard building damage detection. It will also highlight the IDEAMAPS Data Ecosystem project, showing how its community-centric approach advances our understanding of urban deprivation and enables DL-based deprivation mapping.