Research Scientist at the Finnish Meteorological Institute (FMI)
Helsinki, Finland
I work at the intersection of physics-based numerical weather prediction, satellite remote sensing, and data-driven weather forecasting.
- Data-driven weather forecasting: Evaluating machine learning approaches, including graph neural networks, Transformers, and multivariate statistical postprocessing.
- Satellite data assimilation: Assimilating infrared and microwave radiances from various earth-orbiting satellites in the HARMONIE-AROME regional NWP model.
- Operational NWP and verification: System diagnostics, observation monitoring, and model verification with MetCoOp, ACCORD, HIRLAM, and Destination Earth On-Demand Extremes.
- Tools: Python, Fortran, R, Bash, Git, Linux-based HPC environments and Agentic coding.
- PhD thesis (2026): Extending Satellite Radiance Assimilation in High-Latitude Regional Numerical Weather Prediction, University of Helsinki.
- Schönach et al. (2025): “Assimilation of SEVIRI Water Vapour Radiances in HARMONIE-AROME at Large Satellite Zenith Angles.” Tellus A. doi:10.16993/tellusa.4106
- Schönach et al. (2020): “Postprocessing Ensemble Forecasts of Vertical Temperature Profiles.” ASCMO. doi:10.5194/ascmo-6-45-2020
Most of my operational work—including Fortran, Python, and HPC workflows for HARMONIE-AROME and AI forecasting—is developed in internal FMI and ACCORD/MetCoOp repositories. My public contributions include contributions in the Hirlam organisation.