Towards Operational Automated Greenhouse Gas Plume Detection and Delineation

TL;DR

This paper reviews and demonstrates how convolutional neural networks can be effectively used for operational greenhouse gas plume detection and delineation, addressing key challenges like data quality and bias.

Contribution

It introduces a multitask CNN model for simultaneous detection and segmentation, and establishes best practices for operational GHG plume monitoring.

Findings

CNNs achieve operational detection performance when obstacles are addressed.

Multitask models improve plume delineation accuracy.

Thresholds for reliable detection vary by source and region.

Abstract

Operational deployment of a fully automated facility-scale greenhouse gas (GHG) plume detection system remains challenging for fine spatial resolution imaging spectrometers, despite recent advances in deep learning approaches. With the dramatic increase in data availability, however, automation continues to increase in importance for emissions monitoring. This work reviews and addresses several key obstacles in the field: data and label quality control, prevention of spatiotemporal biases, and correctly aligned modeling objectives. We demonstrate through rigorous experiments using multicampaign data from airborne and spaceborne instruments that convolutional neural networks (CNNs) are able to achieve operational detection performance when these obstacles are alleviated. We demonstrate that a multitask model that learns both instance detection and pixelwise segmentation simultaneously can successfully lead towards an operational pathway. We evaluate the model's plume detectability across emission source types and regions, identifying thresholds for operational deployment. Finally, we provide analysis-ready data, models, and source code for reproducibility, and work to define a set of best practices and validation standards to facilitate future contributions to the field.