Impact Plan ML

We use machine learning techniques to detect methane plumes in satellite data. The first project focused on detection of methane plumes in TROPOMI data. We use machine learning techniques to scan through millions of methane observations to find methane plumes originating from super-emitters. When combined with targeted high-resolution observations, this allows us to pinpoint large leaks, enabling mitigation activities (see IMEO Mars). This work is described in Schuit et al, 2023, and forms the basis for our weekly plume detections.

In the next step of the project, we will apply machine learning to other (high-resolution) satellite datasets.