Foundational ecosystem models
24.10.2023While foundational models are well-known in NLP and beginning to be explored in weather forecasting, such model does not yet exist for ecosystem data. This is a gap, and one that we intend to fill!
The ERA5 dataset is a valuable source for globally gridded atmospheric (rain, air temperature, air moisture, etc. at certain heights above the ground) and ground (soil moisture, radiation, etc.) data. Most weather forecast models are at least partially trained on this dataset.
While foundational models are well-known in NLP and beginning to be explored in weather forecasting, such model does not yet exist for ecosystem data. This is a gap, and one that we intend to fill!
You will need advanced experience with Pytorch and neural networks in general. Previous experience with large datasets is a must. An independent and self-driven work ethic is highly beneficial.
Supervisor: Pascal Janetzky
