ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume V-3-2022
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., V-3-2022, 147–154, 2022
https://doi.org/10.5194/isprs-annals-V-3-2022-147-2022
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., V-3-2022, 147–154, 2022
https://doi.org/10.5194/isprs-annals-V-3-2022-147-2022
 
17 May 2022
17 May 2022

AUTOMATIC METHANE PLUMES DETECTION IN TIME SERIES OF SENTINEL-5P L1B IMAGES

E. Ouerghi1, T. Ehret1, C. de Franchis1,2, G. Facciolo1, T. Lauvaux3, E. Meinhardt1, and J.-M. Morel1 E. Ouerghi et al.
  • 1Université Paris-Saclay, CNRS, ENS Paris-Saclay, Centre Borelli, 91190, Gif-sur-Yvette, France
  • 2Kayrros SAS
  • 3Laboratoire des Sciences du Climat et de l’Environnement, CEA, CNRS, UVSQ/IPSL, France

Keywords: Methane, Hyperspectral, Time series, Atmospheric modeling, Anomaly detection

Abstract. Reducing methane emissions is essential to tackle climate change. Here, we address the problem of detecting automatically large methane leaks using hyperspectral data from the Level 1B product of the Sentinel-5P satellite. To do this, two features of TROPOMI (TROPOspheric Monitoring Instrument), the Sentinel-5P satellite sensor, are exploited. The first one is the fine spectral sampling of the data which allows to isolate features of the methane absorption spectrum in the shortwave infrared wavelength range (SWIR). The second one is the daily coverage of the whole Earth which allows to perform time series analysis. Our method involves three main steps: i) a pixel reconstruction, ii) an angle correction and iii) a plume detection with a time series. In the first step, a simplified absorption model is inverted to recover, for each pixel, a coefficient representative of the presence of methane which we call the methane coefficient. In the second step, a correction is made to the methane coefficient to take into account the viewing angle of the satellite. In the third step, the obtained coefficient is normalized spatially and then the detection is carried out pixel by pixel, by looking for anomalies in a time series. We validate our method by comparing the detected plumes against a recently published dataset of plumes manually detected in the Sentinel-5P L2 methane product. We then show how our method can complement the Sentinel-5P L2 methane product for the detection of methane plumes.