Theoretical study and algorithmic developments aimed at detecting methane emissions from multispectral data

Details

Supervisors

Dehant, Véronique

Faculty

Faculté des sciences

Degree label

Master [120] en sciences physiques, à finalité approfondie

Abstract

For the first time in history, the Earth is experiencing global warming as a result of human activity. Specifically, anthropogenic emissions of greenhouse gases are the chief cause of this threat to biodiversity. Among these gases, methane is of particular concern due to its significant presence in the atmosphere and its high global warming potential over a short period of time. Furthermore, 35% of methane emissions can be attributed to the fossil fuel industry, and numerous leaks occur at its facilities. This implies that monitoring and cutting those “Super-emitters” is one of the fastest opportunities we have to immediately slow the rate of global warming. This master’s thesis demonstrates the detection capabilities of methane emissions using public multispectral data from the Sentinel-2 space mission. It turns out that multispectral satellites can be valuable by compensating for their lack of spectral resolution with much better spatial resolution and revisit rate. Thanks to a theoretical study of radiative transfer in the atmosphere, the use of satellite data and algorithmic developments, we have successfully built our own methane retrieval model. Although imperfect, it constitutes a proof of concept that could help mitigate climate change.