Traditional approaches for deriving NOxemissions from satellite observations assume observed NO2column loadings are certain, and apply either a fixed and smooth wind/NO2gradient, or complex modeling systems which are inflexible to changes in emissions of NO, NO2, and other co-emitted substances and emissions variability at high spatial/temporal resolution. This analysis leverages a mass-conserving and flexible system to compute daily and grid-based NOxemissions explicitly considering uncertainties in the observed TROPOMI NO2columns, wind divergence, and first-order chemical, physical, and thermodynamic processes. Considering a conservative 10% TROPOMI uncertainty, 84% of grids are revealed to have robust emissions and 48% of grids have 12-months of data per year, while community-standard approaches fail to retrieve more than 6-months of data per year on any grid in Northern China. The major factors behind these differences are the improved computation of the effects of natural wind variation due to geophysical factors with the dual-gradient term, handling the strong non-linearity of the spatial gradient operator, as well as buffering from first order chemical, physical, and thermodynamic terms. These findings aim to offer a more precise, accurate, and robust assessment of the observational uncertainties involved. In accordance with the principles of MVR (Monitor, Verify, and Report) mandated by the US and China EPAs, this study proposes a new approach to enhance both monitoring and verification techniques. By accurately identifying the temporal and spatial occurrence of NOxemissions, more effective and trustworthy mitigation strategies are supported.

Painterly image harmonization aims to insert photographic objects into paintings and obtain artistically coherent composite images. Previous methods for this task mainly rely on inference optimization or generative adversarial network, but they are either very time-consuming or struggling at fine control of the foreground objects (e.g., texture and content details). To address these issues, we propose a novel Painterly Harmonization stable Diffusion model (PHDiffusion), which includes a lightweight adaptive encoder and a Dual Encoder Fusion (DEF) module. Specifically, the adaptive encoder and the DEF module first stylize foreground features within each encoder. Then, the stylized foreground features from both encoders are combined to guide the harmonization process. During training, besides the noise loss in diffusion model, we additionally employ content loss and two style losses, i.e., AdaIN style loss and contrastive style loss, aiming to balance the trade-off between style migration and content preservation. Compared with the state-of-the-art models from related fields, our PHDiffusion can stylize the foreground more sufficiently and simultaneously retain finer content. Our code and model are available at https://github.com/bcmi/PHDiffusion-Painterly-Image-Harmonization