Fei Liu's ROSES-24 Proposal Accepted for Work on Improved NOx Modeling

6.12.2025

slide graphic from MPS
Dr. Fei Liu

A proposal from a GMAO member was recently accepted as part of the ROSES-24 TEMPO/ACX Science and Applications Team. This research, led by Fei Liu (pictured above) is titled: "Spatiotemporal Fusion of TEMPO-Inferred and Bottom-Up Estimates for High-resolution Nitrogen Oxide Emissions”, Fei Liu (PI, 610.1/MSU), Can Li (614), Daniel Tong (George Mason University), and Joanna Joiner (614). The GEOS Composition Forecasting (GEOS-CF) system and other air quality models rely on bottom-up inventories for data on nitrogen oxides (NOx = NO2 + NO) emissions, which are key precursors to ozone (O3) and particulate matter. Bottom-up inventories typically estimate annual emissions at coarser resolution (e.g., at state/county level for NEI) and allocate them to finer resolution required by models, using temporal profiles and spatial-distribution proxies that are often highly uncertain. Additionally, updates to these inventories are often delayed because of the challenges of data collection and reporting requirements. While nitrogen dioxide (NO2) observations from Low Earth Orbit (LEO) satellites have been widely used to infer NOx emissions and supplement bottom-up inventories, they do not capture the diurnal variations crucial for understanding dynamic tropospheric O3 formation. The geostationary TEMPO instrument provides hourly NO2 data, yet the nonlinear relationship between satellite retrieved NO2 and NOx emissions necessitates a holistic approach. Integrating satellite data with bottom-up information is crucial to gain new insights into the diurnal variation of NOx emissions.

slide graphic from MPS
Figure 1. The proposed methodology detects the diurnal variation in NOx emissions, which is different from that in tropospheric NO2 vertical column densities (VCDs), by considering NOx transport and chemistry without additional inputs from Chemical Transport Models (CTMs). Data are averaged from May to September 2019. The synthetic NO2 VCDs around Washington D.C., (a, c) are simulated by a CTM using given emissions, “true emissions”. The proposed approach will enhance the ability of TEMPO to capture the spatial distribution of emissions (*remis > rVCD).

The team proposes a fusion NOx emission inventory that reconciles TEMPO-derived hourly daytime emissions with a state-of-the-science bottom-up inventory, delivering daily emissions with TEMPO-constrained spatiotemporal patterns at a spatial resolution of 4 km to satisfy the requirements for high-resolution atmospheric composition simulations and forecasts. A preliminary study using their proposed approach [Figure 1] shows that hourly emissions can be resolved from NO2 columns without running CTMs. This significantly boosts computational efficiency and enables emission estimates at a 4 km resolution, thereby maximizing the use of TEMPO’s high-resolution data. The TEMPO-derived emissions will reduce the bias and lag in bottom-up estimates, reducing uncertainties in simulated tropospheric composition. As the errors in bottom-up emissions tend to magnify with finer resolutions, enhancing the accuracy of emissions data becomes crucial, especially as air quality models continue to increase in complexity and resolution.

The GMAO congratulates Dr. Liu on her accepted ROSES proposal, and looks forward to the accomplishments from her team's work.