Modelling green-house gas emission and sinks using Earth Observation data

Organization
KappaZeta Ltd
Abstract
Why is modelling green-house gas fluxes important?
Stopping anthropogenic climate change is arguably among the most important challenges of our time. But even more important is to understand the Earth as a system to know better what is going on with our one and only home planet. See the big picture, do decisions based on actual data and not on subjective emotions. The more climate change is studied the more evident becomes the role of anthropogenic green-house gases in causing it. Therefore, to estimate green-house gas fluxes and their geographical distribution is uttermost important.

It is infeasible to measure green-house gases directly globally with in situ sensors. Earth Observation provides a solution to this problem. Whereas the sensor technology is not so mature yet to directly measure green-house gas fluxes from the orbit with high spatial resolution, it is possible to model them. Copernicus programme and other satellite missions’ data provide a comprehensive set of input features to model greenhouse gas fluxes with fine spatial resolution. Besides the Earth-saving motivation it is important for carbon trading applications – to know how much green-house gases a certain agricultural or forestry land parcel emits or absorbs.

Data Science role for solving the problem
While there is a comprehensive global data set of satellite measurements, the ground-reference data about actual green-house gas fluxes in various landscapes and ecological conditions is extremely limited. In data science language – there is a rich and relatively well calibrated set of input features, but very small set of labels. The challenge is to fit as accurate model as possible while using minimum amount of labelled ground reference data about actual green-house gas fluxes.
It is an interdisciplinary topic to be solved in cooperation with Tartu Observatory and Institute of Ecology and Earth Sciences, University of Tartu.
Graduation Theses defence year
2021-2022
Supervisor
Kaupo Voormansik
Spoken language (s)
Estonian, English
Requirements for candidates
The candidates should have basic software development skills and know the fundamentals of machine learning and deep learning. Interest towards Earth Observation and satellite technology is required, but prior knowledge is not. KappaZeta can provide the basic training needed to get going.
Level
Masters
Keywords
#machine_learning #deep_learning #earth_observation #satellite_data #climate_change

Application of contact

 
Name
Kaupo Voormansik
Phone
E-mail
kaupo.voormansik@kappazeta.ee
See more
https://kappazeta.ee/