Improved western U.S. background ozone estimates via constraining nonlocal and local source contributions using Aura TES and OMI observations

Western U.S. near‐surface ozone (O 3 ) concentrations are sensitive to transported background O 3 from the eastern Pacific free troposphere, as well as U.S. anthropogenic and natural emissions. The current 75 ppbv U.S. O 3 primary standard may be lowered soon, hence accurately estimating O 3 source contributions, especially background O 3 in this region has growing policy‐relevant significance. In this study, we improve the modeled total and background O 3 , via repartitioning and redistributing the contributions from nonlocal and local anthropogenic/wildfires sources in a multi‐scale satellite data assimilation system containing global Goddard Earth Observing System–Chemistry model (GEOS‐Chem) and regional Sulfur Transport and dEposition Model (STEM). Focusing on NASA's ARCTAS (Arctic Research of the Composition of the Troposphere from Aircraft and Satellites) field campaign period in June–July 2008, we first demonstrate that the negative biases in GEOS‐Chem free simulation in the eastern Pacific at 400–900 hPa are reduced via assimilating Aura‐Tropospheric Emission Spectrometer (TES) O 3 profiles. Using the TES‐constrained boundary conditions, we then assimilated into STEM the tropospheric nitrogen dioxide (NO 2 ) columns from Aura‐Ozone Monitoring Instrument to indicate U.S. nitrogen oxides (NO x = NO 2 + NO) emissions at 12 × 12 km 2 grid scale. Improved model skills are indicated from cross validation against independent ARCTAS measurements. Leveraging Aura observations, we show anomalously high wildfire NO x emissions in this summer in Northern California and the Central Valley while lower anthropogenic emissions in multiple urban areas than those representing the year of 2005. We found strong spatial variability of the daily maximum 8 h average background O 3 and its contribution to the modeled total O 3 , with the mean value of ~48 ppbv (~77% of the total).