E-CONTRAIL 2 presents project research at Contrails Analysis Workshop '26
Updated: 11 minutes ago
Researchers from the SESAR-funded E-CONTRAIL 2 project presented key advancements at the Contrails.org Contrails Analysis Workshop 2026 held at Imperial College London. The presentations highlighted progress across satellite AI detection, geostationary visibility modeling, and global ground-camera tracking networks.
1. Multi-Sensor Satellite AI and Project Roadmap
Dr. Hugues Brenot (BIRA-IASB) presented an overview of the E-CONTRAIL 2 project. Major highlights included:
Novel Satellite RGB Composites: Evaluating COBRA-derived RGB ice composites using Meteosat Third Generation (MTG-I1 / FCI) imagery to improve semantic segmentation and AI model training.
Physics-Informed Neural Networks (PINNs): Implementing PINN architectures for advection-diffusion tracking and performing cross-matching between geostationary MTG imagery and vertical profiles from ESA’s EarthCAREsatellite lidar.
Interactive Dashboard: Demonstrating real-time climate analytics, ADS-B flight impact assessments, and 24-hour AI contrail outbreak forecasts hosted at econtrail.aeronomie.be.
2. Physics-Based Satellite Visibility Metric
Addressing why coarse geostationary (GEO) satellite resolution (~2 km) leaves blind spots for narrow or young contrails, Vincent Meijer (TU Delft) presented a poster on new physics-based visibility metric:
Benchmark Dataset: Analyzed 109 contrails (64 visible in GEO infrared imagery) observed across 25 days at the ARM Southern Great Plains (SGP) site.
Integrated Inputs: Collocated ground all-sky camera polygons, vertical cross-sections from a High-Spectral Resolution LIDAR, GOES-East CONUS imagery, and ERA5-advected ADS-B trajectories.
Signal-to-Noise Formulation: Developed a brightness temperature contrast equation that demonstrates strong predictive skill in determining whether a contrail is visible in satellite imagery.
Download poster
3. Global Ground-Based Camera Networks
Emily Tracey (Western University) presented progress on adapting optical infrastructure from the Global Meteor Network (GMN)—comprising over 1,490 cameras in 45 countries—for automated contrail tracking:
Automated Data Pipeline: Deployed across 21 cameras to capture 5-second timelapses, integrate ERA5 wind fields, and perform automated flight attribution via ADS-B data.
Field Demonstration: Data from an August 2026 campaign in London, Ontario showed the pipeline monitoring 19,387 flights and automatically detecting 955 distinct contrails.
Comments