Accelerating Top-Down Emission Constraints via a Hybrid Deep Learning and Data Assimilation Framework

This dataset supports the study entitled “Accelerating Top-Down Emission Constraints via a Hybrid Deep Learning and Data Assimilation Framework.” The archive includes: (1) final posterior emission files inferred using the conventional RAPAS system and the proposed RAPAS-DL system; and (2) ground-based CO and NO2 observations used in the data-assimilation procedure. The posterior emission files are provided on the model grid and support the national, regional, provincial, and temporal analyses presented in the study. The observational dataset contains the surface monitoring data used to constrain the emission estimates. Collectively, these files enable reproduction and independent examination of the emission adjustments inferred by RAPAS and RAPAS-DL and support further comparisons between conventional chemical transport model and DL-DA inversion approaches.