AI-based remote sensing of green-house gas emissions

Climate Change is driven by human-emitted green-house gases (GHGs), but pinpointing sources of emissions is not trivial without data of on-site activities. Having real-time maps of GHGs could help us quantify emissions from agriculture (livestock, fertilizers,...) and forestry practices, which remain relatively uncertain, as well as closely monitor emissions from industrial processes and operations ( e.g. methane (CH4) emissions from the oil & gas industry are major producers of methane [1]; Methane is 80 times more harmful than CO₂ for 20 years after it is released [2]).

The accurate monitoring of GHG emission concentrations, and pin-pointing the emission sources, would be valuable in guiding regulations or incentives that could lead to better land use and industrial practices. For example, data on emissions make it possible to set effective targets, and pinpointing the sources of emissions makes it possible to enforce regulations.

While GHGs are invisible to our eyes, we can observe these compounds with hyperspectral cameras, selecting the wavelengths of interest. This opens the path for remote sensing of emissions. Indeed, many satellites are equipped with hyperspectral cameras and can perform, to some extent, estimations of CO2, CH4 (methane), H2O, and N2O (nitrous oxide) emissions. While extremely useful for studying climate change, most of these satellites have very coarse spatial resolution and large temporal and spatial gaps, making them unsuitable for precise tracking of emissions. Standard satellite imagery provides RGB images with much higher resolution, which could be used in an ML algorithm to fill the gaps in hyperspectral data and obtain more precise information about emissions. Some preliminary work [3] has studied this possibility, but this remains largely an open problem with high potential impact.

Additional methods can be further explored, e.g. remote sensing data fusion can be a solution to obtain more accurate GHG emissions data. The fused data from sensors installed on satellites, UAV and ground stations, with different spatial and spectral resolutions, can provide even more detailed information to compensate for the limitations of each individual source [4]. Satellites can provide time-series data across broad areas, UAV-borne remote sensing can provide hyperspectral data with higher spatial and spectral resolution. More regional and precise measurements can be achieved by combining the satellite-based, UAV-based imaging data and (mobile) ground in-situ measurement data.

Satellites, UAVs,... are typically equipped with multiple sensors. Multi-sensor data fusion, combining hyperspectral and LiDAR sensor data can provide even more accurate information about emission sources.

In this project, you will build and investigate large, high-quality datasets and study and develop advanced algorithms that leverage remote sensing data. The aim is to precisely identify and monitor GHG emissions and their sources. These insights can then contribute to informed decision-making for targeted emission reductions and effective climate change mitigation strategies.

Our ideal candidate has the following skills and competences:

• Master in exact or applied sciences, engineering/bioengineering or environmental sciences
• Expertise in remote sensing and data science
• Experience working with remote sensing data & sensor data fusion (RGB, multi & hyperspectral imaging, LiDAR,...)
• Expertise in using machine learning techniques (on remote sensing data is an asset)
• Programming skills in Python, R, C++
• Fluent in English (incl. writing and presentation skills)
• Passionate about sustainability, strong interest in emissions monitoring,...
•