Probabilistic Machine Learning for Spatio-Temporal Forecasting Under Climate and Environmental Uncertainty

I offer expertise in probabilistic machine learning, uncertainty quantification, Bayesian deep learning, and spatio-temporal modelling, with a strong focus on applying these methods to real-world scientific and societal challenges. My research develops ML models that provide not only accurate predictions but also reliable uncertainty estimates—crucial for decision-making in complex domains such as climate, land-use forecasting, environmental risk, and geospatial information systems.

A recent example is the DK-Future concept proposal , which integrates probabilistic modelling with geospatial data to predict future land-use scenarios in Denmark under uncertain climate pathways. The work combines Earth-observation datasets, climate projections, socio-economic variables, and regulatory frameworks with Bayesian deep learning, ensemble methods, and variational inference. These approaches produce probabilistic forecasts of long-term spatio-temporal processes, accounting for compound climate impacts and low-probability high-impact events.

I am interested in collaborating with MSCA Postdoctoral Fellows or Doctoral Networks to develop new probabilistic models for climate-related processes, environmental AI, uncertainty-aware forecasting, geospatial ML, or any application domain where robust, interpretable, and uncertainty-aware ML is needed.
Collaboration can include:
• development of new probabilistic ML methods;
• modelling spatio-temporal processes (e.g., climate, mobility, land use, environmental risks);
• Bayesian deep learning, ensembles, diffusion-based uncertainty modelling;
• integration of diverse data sources (Earth observation, climate scenarios, policy data);
• decision-support tools for science, sustainability, or public policy.

I welcome collaborations with researchers from computer science, geography, climate sciences, sustainability, urban planning, environmental modelling, or related fields requiring uncertainty-aware ML.