Air Quality Downscaling with Station-Guided Pseudo-Supervision

2026-07-06

AI Summary

This study addresses the spatial support mismatch between coarse-resolution PM₂.₅ fields from atmospheric models and sparse ground-based monitoring stations by proposing a high-resolution (approximately 1 km, 40× super-resolution) air quality downscaling method that does not rely on temporal modeling. The approach integrates CAMS atmospheric reanalysis data with multisource geospatial covariates—including anthropogenic activity, land cover, elevation, satellite-derived aerosol optical depth, and wind fields—within a multiscale Transformer network to jointly perform super-resolution and bias correction. A novel station-guided pseudo-supervision strategy is introduced, which generates dense pseudo-labels via spatial Gaussian mixture interpolation from OpenAQ observations, effectively bridging the gap between point measurements and pixel-level predictions. Experiments across Europe demonstrate that the model substantially mitigates CAMS’s local systematic biases while recovering fine-scale spatial structures.

Abstract

Super-resolving coarse atmospheric fields to local PM$_{2.5}$ variations is uniquely challenged by a mismatch in spatial support: while pixels represent regional averages, ground-truth observations are discrete, unaligned samples of a continuous spatial signal. To bridge this gap, we present a station-guided framework for high-resolution PM$_{2.5}$ downscaling over Europe. Taking coarse CAMS atmospheric composition fields alongside heterogeneous side information (i.e., human activity, land cover, elevation, satellite aerosol observations, and wind fields) our framework jointly super-resolves ($\times 40$, $\approx$ 1 km) and bias-corrects CAMS rasters, without relying on temporal sequence modelling. To address the challenge of densely supervising our multi-scale transformer network with sparse in-situ data, we introduce a time-agnostic propagation strategy that utilises spatial Gaussian blending of interpolated OpenAQ observations. Extensive qualitative and station-level evaluations across Europe demonstrate that our model recovers fine-grained spatial structures and effectively mitigates localised CAMS biases.

Problem

• Air Quality Downscaling
• PM2.5
• Spatial Support Mismatch
• Super-resolution
• In-situ Observations

Innovation

• station-guided downscaling
• pseudo-supervision
• multi-scale transformer
• spatial Gaussian blending
• PM2.5 super-resolution