TEMPO’s hourly, high-resolution observations of nitrogen dioxide (NO2) over North America offer new opportunities to characterize diurnal variability of reactive trace gases, evaluate the performance of atmospheric composition models, and improve air-quality analyses and forecasts. We present progress toward end-to-end use of TEMPO NO2 in NASA’s GEOS Composition Forecasting (GEOS-CF) system using the Joint Effort for Data assimilation Integration (JEDI).
We jointly assimilate TEMPO and TROPOMI tropospheric NO2 column retrievals in a four-dimensional ensemble-variational (4DEnVar) system, combining TEMPO’s hourly measurements with TROPOMI’s global polar-orbiting coverage. We assess how TEMPO constrains the spatial distribution and diurnal evolution of NO2 in different air quality regimes. We characterize TEMPO observation errors and evaluate both the retrievals and the model analyses against Pandonia Global Network measurements. Using the same ensemble framework, we also present initial results on using TEMPO to constrain surface NOx emissions.
We also introduce LOLA-JEDI, a lightweight longitude-latitude model interface for JEDI. LOLA-JEDI allows output from models on regular longitude-latitude grids to be used directly with JEDI, without requiring a full model-specific interface. We demonstrate its use for observation-space evaluation of atmospheric composition models against satellite retrievals, including comparisons across operational forecasting systems.
Finally, we describe progress in developing the SWELL workflow, including a near-real-time monitoring system that ingests various atmospheric composition observations and evaluates them against GEOS-CF forecasts. Together, these developments establish a reproducible pathway from geostationary retrievals to quantitative analyses, emission estimates, and forecast diagnostics, and lay the groundwork for operational atmospheric composition prediction with JEDI.
