Sentinel-5P (S5P) plays a central role in global atmospheric and environmental monitoring, yet its coarse spatial resolution limits the analysis of localized emission sources and sharp concentration gradients. Super-resolution (SR) methods have been proposed to address this limitation, but most existing approaches rely on paired low- and high-resolution data that are unavailable for S5P, restricting their applicability in real-world settings.

In this talk, I will present a self-supervised hyperspectral SR framework specifically designed for S5P that enables training without high-resolution ground truth. The proposed framework integrates the S5P degradation operator and band-dependent noise characteristics derived from sensor signal-to-noise ratio metadata within a self-supervised learning strategy. Convolutional neural network (CNN) architectures based on Depthwise Separable Convolutions (DSC), tailored to S5P's spectral characteristics, are introduced to efficiently enhance spatial detail while preserving spectral fidelity. 

The framework is evaluated across all S5P spectral bands under two settings: (i) reference experiments where supervised and self-supervised learning can be directly compared using synthetic ground truth, and (ii) fully self-supervised settings where high-resolution reference data are unavailable, and assessment relies on physics-based consistency metrics. The proposed self-supervised models achieve performance comparable to their supervised counterparts while producing sharper spatial structures than standard bicubic interpolation. Additional validation using coincident EMIT hyperspectral observations demonstrates physically meaningful spatial enhancement, particularly along coastline regions. These results highlight the potential of self-supervised learning to improve the effective spatial resolution of atmospheric satellite observations in scenarios where high-resolution reference data are inherently unavailable.