A PDE-Informed Latent Diffusion Model for 2-m Temperature Downscaling
This work addresses the challenge of dynamical downscaling for atmospheric science, representing an incremental improvement by fine-tuning an existing diffusion model with physics-based regularization.
The paper tackles the problem of reconstructing high-resolution 2-m temperature fields from atmospheric data by developing a physics-conditioned latent diffusion model that integrates a PDE loss term for physical consistency, resulting in enhanced physical plausibility of the generated fields.
This work presents a physics-conditioned latent diffusion model tailored for dynamical downscaling of atmospheric data, with a focus on reconstructing high-resolution 2-m temperature fields. Building upon a pre-existing diffusion architecture and employing a residual formulation against a reference UNet, we integrate a partial differential equation (PDE) loss term into the model's training objective. The PDE loss is computed in the full resolution (pixel) space by decoding the latent representation and is designed to enforce physical consistency through a finite-difference approximation of an effective advection-diffusion balance. Empirical observations indicate that conventional diffusion training already yields low PDE residuals, and we investigate how fine-tuning with this additional loss further regularizes the model and enhances the physical plausibility of the generated fields. The entirety of our codebase is available on Github, for future reference and development.