Maxime Rischard: Project Eucalyptus -- An Open-Source Toolkit for Satellite-Based Methane Detection and Quantification

Abstract

Robust evaluation of satellite-based methane retrievals remains limited by the scarcity of transparent algorithms and reproducible workflows. Project Eucalyptus presents an open-source implementation of Orbio Earth’s end-to-end pipelines for Sentinel-2, Landsat 8/9, and EMIT imagery. The release includes: (i) trained models for plume segmentation, (ii) a synthetic-plume-data generator for training and validation purposes, and (iii) a framework for detection-threshold evaluation and model benchmarking. By making these resources freely available for research and non-commercial use, Project Eucalyptus aims to lower barriers for comparative studies and to accelerate methodological convergence and innovation across the methane remote-sensing community.

Bio

Maxime Rischard is Director of Science at Orbio Earth, where he co-leads the development of machine-learning methods that convert satellite pixels into quantitative methane estimates. Trained in statistics and physics at the University of Cambridge and Harvard, Maxime has worked across environmental data science, atmospheric physics and geospatial software development for more than a decade.