Dr. Marieke Wesselkamp (she)
Forschungsgruppe Model-Daten-Integration
Projektgruppe EarthNet
Aufgabengebiet
I am a postdoctoral researcher in the EarthNet team, working at the intersection of machine learning and Earth system sciences. My research focuses on forecasting processes on the land surface, with a particular emphasis on short-term forecasting of land surface temperature using deep learning approaches.
Within the WeatherGenerator project, I develop prediction heads and benchmark models for land surface forecasting, contributing to the broader goal of building a coupled land-atmosphere foundation model for the Earth system. Central to my work is the integration of physical knowledge with data-driven methods, and asking how machine learning models can remain grounded in Earth system understanding. This includes questions of model uncertainty and verification, and the appropriate representation of land surface processes across spatial and temporal scales.
Vita
Postdoctoral Researcher -- Max Planck Institute for Biogeochemistry, Jena, Germany (2025–present)
Ph.D. Environmental Sciences -- University of Freiburg, Germany (2021–2025)
Visiting Doctoral Researcher -- ECMWF, Reading, UK (2023 / 2024)
M.Sc. Environmental Sciences -- University of Freiburg, Germany (2018–2021)
B.Sc. Geography -- University of Freiburg, Germany (2013–2018)
Peer-Reviewed Publications
Wesselkamp, M., Albrecht, J., Pinnington, E., Castillo, W., Pappenberger, F. & Dormann, C.F. (2025). Revisiting the ecological forecast limit: Potential, actual and relative system predictability. Methods in Ecology and Evolution, 16(7). https://doi.org/10.1111/2041-210X.70049
Wesselkamp, M., Chantry, M., Pinnington, E., Choulga, M., Boussetta, S., Kalweit, M., Bödecker, J., Dormann, C.F., Pappenberger, F. & Balsamo, G. (2025). Advances in Land Surface Forecasting: A comparative study of LSTM, Gradient Boosting, and Feedforward Neural Network Models as prognostic state emulators in a case study with ecLand. Geoscientific Model Development, 18, 921–937. https://doi.org/10.5194/gmd-18-921-2025
Wesselkamp, M., Moser, N., Kalweit, M., Boedecker, J. & Dormann, C.F. (2024). Process-informed neural networks: a hybrid modelling approach to improve predictive performance and inference of neural networks in ecology and beyond. Ecology Letters, 27, e70012. https://doi.org/10.1111/ele.70012
Wesselkamp, M., Roberts, D.R. & Dormann, C.F. (2024). Identifying potential provenances for climate-change adaptation using spatially variable coefficient models. BMC Ecology and Evolution, 24, 70. https://doi.org/10.1186/s12862-024-02260-z