Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Black carbon plumes from gas flaring in North Africa identified from multi-spectral imagery with deep learning

Tuel Alexandre, Kerdreux Thomas, Thiry Louis

TL;DR

This study tackles the challenge of measuring black carbon (BC) emissions from gas flaring, a major and poorly quantified climate forcer when relying on emission factors. It introduces a ConvLSTM-based framework applied to Sentinel-2 RGB time-series data, augmented with synthetic plumes from a 2D large-eddy simulation model and post-processing using the remaining spectral bands. Applied to North Africa (Algeria, Libya, Egypt) in 2022, the method detects BC plumes at 372 flare clusters and estimates total BC emissions of 6.2 Gg, with Algeria contributing about 75% and the top-10 sites driving a substantial portion of the total. In CO2-equivalent terms, this corresponds to roughly 3.1–6.2 million tCO2, and the work demonstrates the potential for operational, large-scale monitoring to guide mitigation of flaring-related BC and co-emitted pollutants.

Abstract

Black carbon (BC) is an important pollutant aerosol emitted by numerous human activities, including gas flaring. Improper combustion in flaring activities can release large amounts of BC, which is harmful to human health and has a strong climate warming effect. To our knowledge, no study has ever directly monitored BC emissions from satellite imagery. Previous works quantified BC emissions indirectly, by applying emission coefficients to flaring volumes estimated from satellite imagery. Here, we develop a deep learning framework and apply it to Sentinel-2 imagery over North Africa during 2022 to detect and quantify BC emissions from gas flaring. We find that BC emissions in this region amount to about 1 million tCO$_{2,\mathrm{eq}}$, or 1 million passenger cars, more than a quarter of which are due to 10 sites alone. This work demonstrates the operational monitoring of BC emissions from flaring, a key step in implementing effective mitigation policies to reduce the climate impact of oil and gas operations.

Black carbon plumes from gas flaring in North Africa identified from multi-spectral imagery with deep learning

TL;DR

This study tackles the challenge of measuring black carbon (BC) emissions from gas flaring, a major and poorly quantified climate forcer when relying on emission factors. It introduces a ConvLSTM-based framework applied to Sentinel-2 RGB time-series data, augmented with synthetic plumes from a 2D large-eddy simulation model and post-processing using the remaining spectral bands. Applied to North Africa (Algeria, Libya, Egypt) in 2022, the method detects BC plumes at 372 flare clusters and estimates total BC emissions of 6.2 Gg, with Algeria contributing about 75% and the top-10 sites driving a substantial portion of the total. In CO2-equivalent terms, this corresponds to roughly 3.1–6.2 million tCO2, and the work demonstrates the potential for operational, large-scale monitoring to guide mitigation of flaring-related BC and co-emitted pollutants.

Abstract

Black carbon (BC) is an important pollutant aerosol emitted by numerous human activities, including gas flaring. Improper combustion in flaring activities can release large amounts of BC, which is harmful to human health and has a strong climate warming effect. To our knowledge, no study has ever directly monitored BC emissions from satellite imagery. Previous works quantified BC emissions indirectly, by applying emission coefficients to flaring volumes estimated from satellite imagery. Here, we develop a deep learning framework and apply it to Sentinel-2 imagery over North Africa during 2022 to detect and quantify BC emissions from gas flaring. We find that BC emissions in this region amount to about 1 million tCO, or 1 million passenger cars, more than a quarter of which are due to 10 sites alone. This work demonstrates the operational monitoring of BC emissions from flaring, a key step in implementing effective mitigation policies to reduce the climate impact of oil and gas operations.
Paper Structure (10 sections, 4 equations, 6 figures)

This paper contains 10 sections, 4 equations, 6 figures.

Table of Contents

  1. Introduction
  2. Proposed Methodology
  3. Model and training data
  4. Formatting of Sentinel-2 data
  5. Post-processing of model output
  6. Evaluation
  7. Conclusion
  8. Appendix
  9. Estimating BC emission rates
  10. Supplementary figures

Figures (6)

  • Figure 1: Map of our region of study (Algeria, Libya and Egypt) showing the GGFR flaring locations (green dots) and the set of 28 Sentinel-2 tiles we downloaded for this study (red squares, tile name shown in red).
  • Figure 2: Top panels: example of detected BC plume in Algeria; from left to right: RGB image, plume prediction mask, and BC concentration map (obtained with $\mu=8$ m$^2$/g, see Appendix \ref{['appendix']}). Bottom panel: Map of BC emission rates (in t/yr) across North Africa. The size and colour of the dots depends on the annual flaring frequency of each location.
  • Figure S1: Identifying flares across North Africa. (a) Spatial distribution of the clusters of flares across North Africa. The size and colour of the dots depends on the annual flaring frequency of each location. (b) Distribution of flaring frequencies.
  • Figure S2: Example of synthetic BC plume integration in Sentinel-2 RGB scenes (top panels). The corresponding plume simulations are shown on the bottom panels.
  • Figure S3: Map of BC emission rates across North Africa. The size and colour of the dots depends on the annual flaring frequency of each location.
  • ...and 1 more figures