The AETHER project (Advancing Explainable AI for Earth Observation) has officially launched, marking the start of a one-year European research and innovation initiative funded by the European Space Agency (ESA) under the EXPRO+ programme.

The project kickoff took place in Q4 2025, bringing together representatives from all consortium partners: Wageningen University & Research (WUR), Linköping University (LiU), the Agricultural University in Krakow, Poland (URK), and the Norwegian Institute for Air Research (NILU), alongside colleagues from ESA’s Φ-lab. The meeting provided an important opportunity to align on the project’s scope, objectives, and implementation strategy.

During the kickoff meeting, partners discussed the overall project methodology, key deliverables, expected scientific and societal impacts, and the roles and responsibilities of each organisation. This collaborative start laid the foundation for effective coordination throughout the project’s ambitious timeline and ensured alignment with ESA’s strategic priorities for Artificial Intelligence in Earth Observation.

Project Objectives and Scope

Over the next 12 months, AETHER aims to advance the state of the art in explainable artificial intelligence (xAI) for Earth Observation by developing methods that improve transparency, interpretability, and trust in AI-driven EO applications. The project will demonstrate these innovations through three high-impact, real-world use cases:

• Urban Heat Islands – led by the Agricultural University in Krakow (URK) and supported by Wageningen Environmental Research (WENR), focusing on detecting and analysing urban temperature patterns to support climate-resilient cities.
• Crop Yield and Rapid Damage Assessment – led by Linköping University (LiU) and supported by WENR, addressing food security and disaster impacts through interpretable EO-based yield forecasting.
• Biodiversity Mapping – co-led by Wageningen University (WU) and WENR, aiming to map species richness loss due to climate change and identify key environmental drivers.

In addition, AETHER will deliver a Proof-of-Concept (TRL-5) software prototype featuring an intuitive user interface, enabling end users to interact with explainable AI models and EO data more effectively. The project also places strong emphasis on open science and community engagement, with plans to share tools, methods, and results through open-source releases, workshops, and scientific publications.

Looking Ahead

AETHER reflects a shared vision among the consortium and ESA: to make AI for Earth Observation transparent, explainable, and trustworthy, and to bridge advanced machine learning research with practical applications in climate action, agriculture, urban resilience, and biodiversity conservation.

The consortium looks forward to sharing regular updates on project progress, key milestones, publications, and events through this website and associated communication channels.