Fully Automatic Trace Gas Plume Detection

2026-05-05

AI Summary

This study addresses the challenges posed by the rapidly growing volume of imaging spectrometer data by proposing a fully automated framework for trace gas plume detection. The approach integrates machine learning–driven morphological analysis with physically constrained spectral fitting, enabling both routine screening and retrospective analysis. It achieves the first end-to-end automatic detection in EMIT data and successfully extends to underexplored gases such as NH₃, NO₂, and CO, uncovering at least 25% more plumes than previously identified through manual inspection—including the first satellite-observed CO plume. In routine operation, the system efficiently detects large plumes with exceptionally low false-positive rates, substantially enhancing the coverage and reliability of trace gas emission monitoring.

Abstract

Future imaging spectrometers will increase data volumes by orders of magnitude, requiring automated detection of trace gas point sources. We present a fully automated framework that combines machine learning-based morphological analysis with physics-based spectroscopic fitting to detect plumes without human participation. Applied to EMIT imaging spectrometer data, the system operates in two modes: "daily digest" that runs automatically on all downlinked data, flagging the largest events for immediate response, and a retrospective analysis that identifies plumes missed by prior human review. The daily digest demonstrates that a significant fraction of the largest plumes can be detected automatically with negligible false positives, while retrospective analysis suggests at least 25% of plumes may have been overlooked. In addition to the previously observed methane point sources, we extend detection to three understudied trace gases: NH3, NO2 and the first observations of carbon monoxide (CO) plume in EMIT imagery.

Problem

• trace gas plume detection
• automated detection
• imaging spectrometers
• point source emissions
• data volume

Innovation

• automated plume detection
• machine learning
• spectroscopic fitting
• imaging spectrometer
• trace gas