The AETHER project focusses on three use cases:
Urban heat islands
Urban areas often experience significantly higher temperatures than surrounding rural regions due to dense construction, limited vegetation, and human activities — a phenomenon known as Urban Heat Islands (UHIs).
This use case focuses on the detection and temporal evolution of urban temperature patterns in Dutch cities, Kraków (Poland), and Guatemala City, combining EO data, climatic variables, and AI models.
By applying explainable AI techniques, the project aims to:
- Identify the main drivers of urban heat dynamics (e.g., land use, surface materials, vegetation cover).
- Provide clear explanations for observed temperature anomalies over time.
- Support urban planners and local authorities in designing targeted mitigation strategies, such as green infrastructure or cooling interventions.
This use case illustrates how xAI can turn complex thermal and land surface data into actionable insights for climate-resilient urban planning.
Crop yield prediction
This use case addresses both yield prediction and post-disaster damage assessment for key crops such as wheat and maize, leveraging the CY-Bench dataset across 29+ countries as well as agroforestry systems in Africa (both smallholder and plantation contexts).
By integrating EO imagery, agro-climatic indicators, and machine learning, AETHER aims to:
- Improve the accuracy and interpretability of crop yield forecasting models, supporting food security monitoring at regional and national scales.
- Provide rapid, explainable assessments of crop damage following extreme events (e.g., droughts, floods, storms), enabling faster response and recovery.
- Demonstrate how xAI can highlight the key environmental and agronomic factors influencing yield variability and damage, empowering local agencies and farmers to make informed decisions.
This use case showcases the potential of explainable AI to strengthen climate resilience and disaster preparedness in agriculture.
Biodiversity
Climate change is accelerating biodiversity loss worldwide, affecting both ecosystems and human livelihoods. This use case focuses on mapping species richness loss for birds (United States & Kenya) and butterflies (United Kingdom) by combining EO data, climatic variables, and species distribution models.
Through explainable AI approaches, AETHER will:
- Quantify and visualize changes in species richness over time under different climate scenarios.
- Identify which environmental factors (e.g., temperature shifts, land cover changes) are most responsible for biodiversity decline.
- Provide interpretable, spatially explicit outputs that can inform conservation strategies, land-use planning, and biodiversity policy at local and international levels.
This use case demonstrates the value of xAI in tackling one of the most pressing environmental challenges — understanding and mitigating biodiversity loss in a changing climate.