• /dataset contient un échantillon de données OCO-2 et des données d'inventaire ;
• /notebooks centralise les notebooks réalisés par l'équipe ;
• /pipeline comprend des scripts pour générer les données nécessaires.

Our goal is to localize CO² emission on earth. We are working with :

• Matthieu Porte, from IGN who submit the projet
• Marie Heckmann, from the French Ministry of Ecology
• Frederic Chevalier, from IPSL (!!Link) ,one of the author of https://www.atmos-chem-phys-discuss.net/acp-2020-123/acp-2020-123.pdf

The OCO-2 Satellite (Orbiting Carbon Observatory) use spectrometers to detect CO² in atmosphere :

!! More info here : https://oco.jpl.nasa.gov/instrument/ So it can't see through clouds or fog. And don't work the same over ground or water.

The swath of the satellite is small : only 10km : !!

And the coverage is partial, no orbit are contiguous.

More info on the mission on https://earth.esa.int/web/eoportal/satellite-missions/o/oco-2.

So that's very limitative and frustrating. We don't have a high resolution images.

NASA made global CO² image :

But it is extrapolation of the data, it's not what the satellite really see.

• Detect peak in data by looking at CO² Plume
• Agregate known sources of CO²
• Compare peak to known sources to correct them
• Machine Learning to find new source
• Display the result on a comprehensive map

• XXX Picture of CO² Plume

• Find nearest inventory from peak position, using the wind vector.

• Better peak detection: So far, we are fitting gaussian curves to detect relevant peaks. 2 issues:
  • we use SciKit Learn curve_fit. Do you know a better algorithme or how to tune parameters of curve_fit ?
  • we are looking at other methodologies to detect anomalies (our 'peaks') in the concentrations - any idea?
• Wind modeling to estimate emission from detected concentration - any idea? (inverting the gaussian plume model)
• Interactive dasboard to share our work on the web (Streamlit ?)