Leveraging JEDI for Atmospheric Composition: A unified framework for evaluating observations and model forecasts

Abstract. Accurate evaluation of both observations and forecasts is essential for advancing atmospheric composition research and improving operational prediction. Traditionally, this has relied on separate workflows with product-specific preprocessing, often limiting reproducibility and creating inconsistencies between models and observational datasets or across different products. Modern data assimilation systems provide precise observation operators for mapping model variables into observation space, yet these capabilities remain underutilized outside assimilation. Here, we demonstrate how the Joint Effort for Data assimilation Integration (JEDI) framework addresses this gap by offering a unified, modular, and model-agnostic system that integrates data assimilation with systematic evaluation. JEDI enables consistent intercomparisons of observations, forecasts, and reanalyses by interfacing with diverse forecast models and gridded datasets, while leveraging carefully designed observation operators to compute equivalent quantities for a wide range of observation products. These include satellite instruments such as Tropospheric Emissions: Monitoring of Pollution (TEMPO), The TROPOspheric Monitoring Instrument (TROPOMI), Moderate Resolution Imaging Spectroradiometer (MODIS), Visible Infrared Imaging Radiometer Suite (VIIRS), and Plankton, Aerosol, Cloud, and ocean Ecosystem (PACE), as well as ground-based networks like Aerosol Robotic Network (AERONET), Pandora, and U.S. Environmental Protection Agency (EPA) AirNow. Case studies illustrate the flexibility of this workflow: (1) NO₂ forecasts from the Weather Research and Forecasting model coupled with Chemistry (WRF-Chem) evaluated against TEMPO, TROPOMI, and Pandora retrievals; (2) surface fine particulate matter and ozone forecasts from WRF-Chem assessed against AirNow measurements using EPA regulatory thresholds; and (3) aerosol optical depth (AOD) retrievals from multiple satellites compared with Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2) and validated against AERONET. These examples highlight JEDI’s ability to detect systematic regional biases, reconcile complementary sampling characteristics across platforms, and assess the added value of unified observation operators for cross-comparison. Overall, JEDI provides a consistent and extensible framework for model validation and observation assessment, reducing redundant preprocessing and aligning evaluation with operational data assimilation practices, and ultimately advancing both research and operational applications in atmospheric composition.