PhD in 3D Radiation Modeling in Clouds and Wildfire Plumes Using GPUs and Machine Learning

This PhD position at Wageningen University & Research focuses on revolutionizing atmospheric radiation modeling by creating a high-performance Monte Carlo ray tracer optimized for GPUs. The project, funded by the NWO Open Technology Programme, aims to replace traditional 1D radiation models with a 3D approach to better understand cloud dynamics, solar energy availability, climate sensitivity, and wildfire plume behavior. The candidate will collaborate with a PhD colleague at Leiden University on GPU optimization and machine learning denoising, ensuring the ray tracer is fast enough for operational use. The research integrates real-world applications through partnerships with KNMI, Whiffle, TNO, VORtech, and Columbia University, ensuring societal impact from the outset.

• Develop and implement a 3D thermal radiation solver within the RTE-RRTMGP-CPP ray tracer, incorporating efficient sampling strategies for radiation sources.
• Set up three benchmark simulations in the GPU-accelerated MicroHH model, focusing on solar energy forecasting, tropical climate, and real wildfire plumes.
• Quantify the impact of 3D radiation on cloud formation, surface irradiance, and plume dynamics compared to conventional 1D radiation models.
• Collaborate with a PhD colleague at Leiden University to co-develop a physics-informed machine-learning denoiser for operational use.
• Evaluate the ray tracer and denoiser with a user committee comprising KNMI, Whiffle, TNO, VORtech, and Columbia University.
• Publish findings in leading journals and present results at international conferences and to industry stakeholders.
• Contribute to the development of open-source simulation codes used by the research community.