Plume Segmentation from MethaneSAT with Cross-Sensor Transfer Learning and Physics-Informed Postprocessing
Abstract
Automated detection and masking of individual methane plumes from satellite imagery is important for operational emission attribution and quantification. We present a machine learning framework for plume detection from MethaneSAT retrieved column-averaged dry-air mole fractions of methane. We address two core challenges: the scarcity of labeled MethaneSAT data and the need for inference reliability across diverse atmospheric and surface conditions. We first demonstrate that Mask R-CNN with a ResNet-50 backbone outperforms a U-Net semantic segmentation baseline on both MethaneAIR (an airborne version of MethaneSAT) and MethaneSAT data, with pixel-level F1 score gains of 10.49 and 5.48 percentage points respectively. To address MethaneSAT data scarcity, we evaluate three cross-sensor transfer strategies leveraging MethaneAIR flights and synthetic plumes. Mask R-CNN with ResNet-50 fine-tuned from MethaneAIR pre-trained weights is the most effective strategy, achieving instance-level precision of 0.60 and recall of 0.98 at the baseline operating point. A physics-informed post-processing pipeline converts detections into two operationally distinct modes. The first is a high-sensitivity mode that applies morphological filtering and proximity-based merging for comprehensive emission screening, achieving precision of 0.71 and recall of 0.94. The second is a high-precision mode that additionally applies a distribution-based classifier for confident source attribution, achieving precision of 0.92 and recall of 0.70. Manual review of detections classified as false positives against our wavelet-based reference plume mask labels shows that some correspond to real methane enhancements... Our data and code used in this work are publicly available at: https://doi.org/10.7910/DVN/FR959H
Related
- 3D seismic salt dome segmentation using a hybrid UNet with global attention, denoising pretraining, and discriminative transfer learning
- Tail Skill and Ensemble-Configuration Sensitivity in Six-Hour Precipitation Postprocessing
- Physics-informed deep learning links geodetic data and fault friction
- A Sentinel-2 benchmark dataset for deep-learning active-fire segmentation across 25 California wildfires, July 2020-August 2026
- Assessing Deep Learning Segmentation Models for Mapping Landslides From LiDAR‐Based Elevation Data
BibTeX
@misc{perezcarrasco2026plume,
title = {{Plume Segmentation from MethaneSAT with Cross-Sensor Transfer Learning and Physics-Informed Postprocessing}},
author = {Manuel Pérez-Carrasco and Maya Nasr and Zhan Zhang and Apisada Chulakadabba and Javier Roger and Raia Ottenheimer and Sébastien Roche and Maryann Sargent and Chris Chan Miller and Daniel Varon and Jack Warren and Luis Guanter and Kang Sun and Jonathan Franklin and Jia Chen and Cecilia Garraffo and Xiong Liu and Ritesh Gautam and Steven Wofsy},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2605.24273},
eprint = {2605.24273},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2605.24273},
}
RIS
TY - UNPB
TI - Plume Segmentation from MethaneSAT with Cross-Sensor Transfer Learning and Physics-Informed Postprocessing
AU - Manuel Pérez-Carrasco
AU - Maya Nasr
AU - Zhan Zhang
AU - Apisada Chulakadabba
AU - Javier Roger
AU - Raia Ottenheimer
AU - Sébastien Roche
AU - Maryann Sargent
AU - Chris Chan Miller
AU - Daniel Varon
AU - Jack Warren
AU - Luis Guanter
AU - Kang Sun
AU - Jonathan Franklin
AU - Jia Chen
AU - Cecilia Garraffo
AU - Xiong Liu
AU - Ritesh Gautam
AU - Steven Wofsy
PY - 2026
DA - 2026/10/10
PB - arXiv
DO - 10.48550/arxiv.2605.24273
UR - https://arxiv.org/abs/2605.24273
AB - Automated detection and masking of individual methane plumes from satellite imagery is important for operational emission attribution and quantification. We present a machine learning framework for plume detection from MethaneSAT retrieved column-averaged dry-air mole fractions of methane. We address two core challenges: the scarcity of labeled MethaneSAT data and the need for inference reliability across diverse atmospheric and surface conditions. We first demonstrate that Mask R-CNN with a ResNet-50 backbone outperforms a U-Net semantic segmentation baseline on both MethaneAIR (an airborne version of MethaneSAT) and MethaneSAT data, with pixel-level F1 score gains of 10.49 and 5.48 percentage points respectively. To address MethaneSAT data scarcity, we evaluate three cross-sensor transfer strategies leveraging MethaneAIR flights and synthetic plumes. Mask R-CNN with ResNet-50 fine-tuned from MethaneAIR pre-trained weights is the most effective strategy, achieving instance-level precision of 0.60 and recall of 0.98 at the baseline operating point. A physics-informed post-processing pipeline converts detections into two operationally distinct modes. The first is a high-sensitivity mode that applies morphological filtering and proximity-based merging for comprehensive emission screening, achieving precision of 0.71 and recall of 0.94. The second is a high-precision mode that additionally applies a distribution-based classifier for confident source attribution, achieving precision of 0.92 and recall of 0.70. Manual review of detections classified as false positives against our wavelet-based reference plume mask labels shows that some correspond to real methane enhancements... Our data and code used in this work are publicly available at: https://doi.org/10.7910/DVN/FR959H
ER -