Toward Explainable and Transferable Deep Downscaling of Atmospheric Pollutants

Understanding the intricate relationship between air quality and human health necessitates a comprehensive assessment of the spatial distribution of pollutants at high resolutions. Here, we use deep learning (DL) methodologies for downscaling atmospheric concentrations of pollutants demonstrated to adversely affect human health (NO2 and PM2.5). We train a model using data fusing multisource information from the Copernicus Atmosphere Monitoring Service (CAMS) and in situ observations from ground-based stations in North Italy. Our model demonstrates robust generalization capabilities by effectively improving bias compared with ground-truth station observations when applied to the state of California, at an order of magnitude different downscaling resolution, despite not previously being trained on or exposed to this region. In addition to the demonstrated transferability, we developed and applied an occlusion-based method over the fused data sources for interpretability. Our results suggest that our model adeptly leverages auxiliary informed data encompassing past, present, and future insights for inference. Thus, we are able to quantify the influence of the input variables on both the predictability and the uncertainty associated with the physical deterministic model.