The concentrations of greenhouse gases in the atmosphere are constantly rising, and unexpected evolution of the carbon dioxide sinks and sources has drawn attention recently.
In my research, I am developing machine learning methods for assessing the evolution of carbon dioxide based on satellite–based measurements of the atmospheric composition. The aim is to create effective models to estimate carbon dioxide fluxes directly from satellite observations. Satellites cannot measure fluxes of the atmospheric gases; hence fast estimates could provide valuable information before computationally heavy atmospheric models have been run. The methods include versatile deep learning techniques to combine data from various earth observations missions such as OCO-2, GOSAT, TROPOMI etc. ESA’s upcoming CO2M mission will provide us even more precise measurements of the carbon dioxide. The developed machine learning model will highly benefit from this data further supporting the mission’s impact.
