Impacts of Horizontal Resolution on Global Data Assimilation of Satellite Measurements for Tropospheric Chemistry Analysis

Abstract

We present the results from a global 0.56°‐resolution chemical data assimilation that integrates satellite observations of ozone, NO2, CO, HNO3, and SO2 from OMI, GOME‐2, SCIAMACHY, TES, MOPITT, and MLS. The assimilation is based on an ensemble Kalman filter technique and simultaneously optimizes ozone precursor emissions and concentrations of various species. The data assimilation at 0.56° resolution reduced model errors against independent surface, aircraft, and ozonesonde observations, which was larger than at coarser resolutions for many cases. By the data assimilation, surface model errors over major polluted regions were reduced by 33%–75% for NO2 and by 15%–18% for ozone. Agreements against assimilated observations for NO2 were improved using the data assimilation at 0.56° resolution by a factor of 1.5–3 compared to 2.8° resolution over major polluted regions. The estimated global total NOx emission over medium and strong source areas were smaller by 15% at 0.56° resolution than at 2.8° resolution associated with resolving small‐scale transport and chemistry processes, while 2%–26% smaller emissions were found for regional total emissions over Europe, the United States, China, India, and South Africa, with larger differences over megacities such as Los Angeles (−41%). The estimated ship emissions were 5%–7% smaller at 0.56° resolution over the Pacific and Atlantic. The 0.56°‐resolution data assimilation provides globally consistent analyses of the emissions and concentrations on a megacity scale, which benefit studies on air quality and its impact on human health at various spatial scales over different regions of the world.

Inferring and evaluating satellite-based constraints on NOx emissions estimates in air quality simulations

“…In addition to smearing effects, coarse resolution models can artificially alter nonlinear NO2 chemistry, leading to biases in inferences of NOx emissions from satellite NO2 columns (Valin et al, 2011;Sekiya et al, 2021;Lamsal et al, 2011).…”

“…Although emissions smearing in the approach is mitigated by only analyzing the lower portion of the model column, our emissions changes may be less precise than targeted assimilation methods, such as 4DVAR adjoint-based methods. Further, coarse grid resolution exacerbates biases in modeled NO2 columns (Valin et al, 2011) and inferred NOx emissions (Sekiya et al, 2021). The air quality model used here does not include stratospheric chemistry, which could affect comparisons against NO2 retrievals.…”

“…Similarly, Kalman filtering and related approaches have been used but are computationally-intensive (e.g. Napelenok et al, 2008;Ding et al, 2020;Ding et al, 2015;Mijling and Van Der A, 2012;Miyazaki and Eskes, 2013;Miyazaki et al, 2017;Miyazaki et al, 2012a;Miyazaki et al, 2012b;Sekiya et al, 2021). Mass balance inversion approaches, which scale model emissions by directly comparing model estimates and satellite observations, were introduced by Martin et al (2003), updated by Lamsal et al (2011), and have been widely used in research and forecasting (e.g.…”

Advances in Simulating the Global Spatial Heterogeneity of Air Quality and Source Sector Contributions: Insights into the Global South

“…25 In general, these models are revealing that global HRMs largely boost the ability to capture fine-scale variation of air quality while sustaining the feedbacks of longrange transport at the global scale. 26,27 Despite the insights into northern midlatitudes from regional models and the pioneering research using global models managed by national laboratories, there is a dearth of understanding of the spatial distribution of air pollution in the Global South. High-resolution simulations are especially important to understand population exposure in the Global South with fast-growing cities and sporadic ground monitors.…”

The worldwide COVID-19 lockdown impacts on global secondary inorganic aerosols and radiative budget

“…To assess the climate impact during the COVID-19 lockdowns, bottom-up emission inventories based on mobility data ( 8 , 9 ) have been commonly used; however, they are highly uncertain due to deviation of mobility datasets from traffic counts and energy consumption statistics, difficulties to accessing datasets that accurately represent industry and residential sectors, as well as the impacts of different lifestyles on emission reductions for residential sector ( 9 , 24 , 25 ). Satellite observations can be used to derive time-varying emission maps of aerosol precursor gases and its impacts on aerosol formations through the state-of-the-art multiconstituent satellite data assimilation system ( 26 , 27 ).…”