Skip to main content
NCBI home page
EPA Author Manuscripts logoLink to EPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jun 1.
Published in final edited form as: Atmos Res. 2022 Jun 1;270:1–14. doi: 10.1016/j.atmosres.2022.106076

Changes in the ozone chemical regime over the contiguous United States inferred by the inversion of NOx and VOC emissions using satellite observation

Jia Jung a, Yunsoo Choi a,*, Seyedali Mousavinezhad a, Daiwen Kang b, Jincheol Park a, Arman Pouyaei a, Masoud Ghahremanloo a, Mahmoudreza Momeni a, Hyuncheol Kim c,d
PMCID: PMC8972085  NIHMSID: NIHMS1789194  PMID: 35370333

Abstract

To investigate changes in the ozone (O3) chemical production regime over the contiguous United States (CONUS) with accurate knowledge of concentrations of its precursors, we applied an inverse modeling technique with Ozone Monitoring Instrument (OMI) tropospheric nitrogen dioxide (NO2) and total formaldehyde (HCHO) retrieval products in the summers of 2011, 2014, and 2017, years in which United States National Emission Inventory were based. The inclusion of dynamic chemical lateral boundary conditions and lightning-induced nitric oxide emissions significantly account for the contribution of background sources in the free troposphere. Satellite-constrained nitrogen oxide (NOx) and non-methane volatile organic compounds (NMVOCs) emissions mitigate the discrepancy between satellite and modeled columns: the inversion suggested 2.33–2.84 (1.07–1.34) times higher NOx over the CONUS (over urban regions) and 0.28–0.81 times fewer NMVOCs emissions over the southeastern United States. The model-derived HCHO/NO2 column ratio shows gradual spatial changes in the O3 production regime near urban cores relative to previously defined threshold values representing NOx and VOC sensitive conditions. We also found apparent shifts from the NOx-saturated regime to the transition regime (or the transition regime to the NOx-limited regime) over the major cities in the western United States. In contrast, rural areas, especially in the east-southeastern United States, exhibit a decreased HCHO/NO2 column ratio by –1.30 ± 1.71 with a reduction in HCHO column primarily driven by meteorology, becoming sensitive to VOC emissions. Results show that incorporating satellite observations into numerical modeling could help policymakers implement appropriate emission control policies for O3 pollution.

Keywords: HCHO/NO2 factor, Inverse modeling, Ozone chemical regime, OMI, CMAQ-DDM

1. Introduction

Ground-level ozone (O3) is primarily formed through complex nonlinear photochemical reactions between nitrogen oxide (NOx) and volatile organic compounds (VOCs) in the presence of sunlight, while stratospheric intrusion contributes a substantial amount of O3 in the upper troposphere (Holton et al., 1995; Knowland et al., 2017). In the United States (U.S.), as long-term exposure to high-level O3 concentrations poses substantial threats to human health and ecosystems, a number of national regulations have been implemented during the last few decades to reduce O3 and its precursors, resulting in a substantial reduction in O3 concentrations. For instance, the Nitrogen Oxides States Implementation Plan Call (NOx SIP Call), enforced by the U.S. Environmental Protection Agency (EPA) from 2003 to 2004, has reduced the interstate transport of O3, resulting in a remarkable reduction in O3 concentrations by 20% to 30% (Foley et al., 2015). Meanwhile, the long-range transport of O3 and its precursors (e.g., NOx, VOCs, and carbon monoxide (CO)), especially from Asia and Europe, becomes important due to their increasing contribution to background O3 levels in the U.S. (Jacob et al., 1999; Lee et al., 2021).

Depending on the relative availabilities of NOx and VOCs concentrations, the O3 formation regime is conventionally classified into three categories—NOx-limited, transition, and NOx-saturated (VOC-limited)—all of which determine the rate of O3 production. The substantial reduction in O3 precursor emissions could alter VOC/NOx ratio, resulting in changes in the O3 chemical regime. Previous studies have reported that the O3 chemical regime has shifted toward the NOx-limited regime in the U.S. urban environments in response to pronounced NOx emission reductions in warm seasons (Jin et al., 2020, 2017). Moreover, variations in changes of different emission sectors also contribute to the spatial variations in the concentrations of O3 precursors and the O3 chemical regime (Jiang et al., 2018).

The identification of the O3 chemical regime is essential to determining the effectiveness of O3 control policies. Based on analysis of a simple trajectory model, Sillman (1995) compared reactive odd nitrogen (NOy), O3/(NOy – NOx), HCHO/NOy, and hydrogen peroxide (H2O2)/nitric acid (HNO3) as indicator species of O3-NOx-VOC sensitivity. The following study by Tonnesen and Dennis (2000) found that the HCHO-to-NO2 ratio (HCHO/NO2) is more valuable than HCHO/NOy, owing to the short lifetime of NO2 and HCHO that can serve as proxies of NOx and non-methane volatile organic compounds (NMVOCs) emissions. Satellite data have broader spatial and temporal coverage than in situ observations; hence, numerous studies have employed satellite-derived HCHO/NO2 ratio (Choi et al., 2012; Duncan et al., 2010; Martin et al., 2004; Schroeder et al., 2017). Recent studies, however, have reported that changes in satellite-observed NO2 and HCHO columns do not necessarily reflect the trends in the U.S. National Emission Inventory (NEI) and near-surface air quality. While the decrease of the satellite-observed tropospheric NO2 column has become slower since 2009, the trend of NEI indicates a monotonic decrease in NOx emissions (Jiang et al., 2018; Silvern et al., 2019). Moreover, because of the significant reduction in surface emissions, the relative contribution of the background sources of NOx in the troposphere (e.g., lightning and soils) has increased, as has its impact on atmospheric chemistry (Kang et al., 2019; Silvern et al., 2019). In the case of HCHO, its trend in urban cores is consistent with the trend of NEI emissions. However, biogenic and biomass burning (open fire) emissions, primarily driven by the temperature, are generally dominant over the southeastern and western U.S., respectively, in summer months (Curci et al., 2010; Palmer et al., 2003; Zhu et al., 2017).

Chemical transport models (CTMs) are useful for not only connecting the column and surface air quality but also identifying the response of ozone to changes in the emissions of ozone precursors. The uncertainty of emission inventories, however, has resulted from the poor classification of emission sources and the limited knowledge of NOx and NMVOCs emission factors (Tian et al., 2010), which often cause significant bias in the results of models when they are compared to observed values. Thus, in conjunction with the broad coverage of satellite measurements, top-down emission estimation is well established and applied in a number of studies optimizing NOx and NMVOCs emission inventories from various areas around the globe. In particular, a rich legacy of NO2 and HCHO column retrievals observed by the National Aeronautics and Space Administration (NASA) Ozone Monitoring Instrument (OMI) has provided nearly continual daily column retrievals at a relatively high resolution since July 2004, which is suitable for longterm analysis and application. Since the public release of the data, the OMI column retrievals have been well validated (Baek et al., 2014; Herman et al., 2009; Lamsal et al., 2014; Wang et al., 2017; Zhu et al., 2016) and applied in numerous studies conducting a top-down assessment of emission inventories (Bae et al., 2020; Goldberg et al., 2019; Kaiser et al., 2018; Millet et al., 2008; Souri et al., 2020, 2016; Tang et al., 2013; Zhu et al., 2014).

In this study, we applied an analytical inverse modeling technique with OMI NO2 and HCHO column retrievals and estimated top-down NOx and NMVOCs emissions over the contiguous U.S. (CONUS) in the summers of 2011, 2014, and 2017, which were the base years that the U.S. NEI was formulated. From adjusted emissions, we examined changes in the ozone chemical regime over the CONUS throughout the study period.

2. Overview of measurements, the modeling system, and the inverse modeling method

2.1. OMI NO2 and HCHO column retrievals

The Ozone Monitoring Instrument (OMI) was launched in 2004 on the NASA Earth Observing System (EOS) Aura satellite, which is in a sun-synchronous polar orbit with a local equator crossing time of 13:45 ± 0:15. OMI is a nadir-viewing near-UV/visible spectrometer that provides complete global coverage of the daily measurements of several air quality components (e.g., NO2, sulfur dioxide (SO2), bromine monoxide (BrO), HCHO, and aerosols) with spatial resolutions of up to 13 × 24 km2 at the nadir, spectral range of 264–504 nm, and spectral resolutions between 0.42 nm and 0.64 nm (Levelt et al., 2006).

In this study, we use the tropospheric NO2 and total HCHO columns from the OMI operational retrieval products (Level 2 and version 3) released by the NASA Goddard Earth Sciences Data and Information Service Center (GES DISC). The tropospheric NO2 column retrievals are calculated through multiple procedures: using a spectral fitting algorithm from the normalized visible wavelength (402–165 nm) to retrieve the slant column, calculating the air mass factor (AMF), and separating troposphere and stratosphere column densities. Details of the data quality assurance and retrieval process are available in previous studies (Bucsela et al., 2013; Choi et al., 2020; Krotkov et al., 2017; Lamsal et al., 2014). To filter out unqualified pixels, we use pixels with a cloud fraction of less than 30%, a terrain reflectivity of less than 30%, a solar zenith angle of less than 65°, a data quality flag as zero, and a cross-track quality flag as zero to screen rows affected by row anomalies (Torres et al., 2018).

This study employed the total HCHO column retrieved by the new Smithsonian Astrophysical Observatory (SAO) algorithm (González Abad et al., 2015). Because of increased background HCHO and noise from the previous product, the retrieval algorithm has been updated in the processing of data such as the fitting window (328.5–356.5 nm), the AMF calculation, and the reference sector re-normalization of vertical column densities. With greatly reduced noise and increased temporal stability, the quality of the data has significantly improved (González Abad et al., 2015). In addition, the estimated detection limit is 1 × 1016 molecules cm−2 (González Abad et al., 2015). Similar to NO2, we use pixels with a cloud fraction of less than 40%, a solar zenith angle of less than 65°, a data quality flag as zero, a cross-track quality flag as zero, and a vertical column density with a range of −2 to 2 × 1015 molecules cm−2. For both the OMI NO2 and HCHO columns, we recalculated the AMF and replaced the prior profile of the OMI data with the profile of the CMAQ simulation of this study.

2.2. Modeling setup

To compute both atmospheric concentrations of gaseous and particulate pollutants and first-order sensitivity coefficients, required for conducting inverse modeling, we use the Community Multiscale Air Quality model configured with the decoupled direct method in three dimensions (CMAQ DDM—3D) based on CMAQ model version 5.2 (Cohan et al., 2005; Napelenok et al., 2006), developed and released by the EPA. Our study domain covering the CONUS utilizes 12 km grid spacing with 459 × 299 grid points (shown in Fig. 1), while the vertical extent between the surface and 100 hPa is discretized with 27 layers of variable thickness. We use carbon bond (CB6r3) for the gas phase and AERO6 for the aerosol chemical mechanism. We obtained the chemical initial and boundary conditions from the CMAQ model with the in-line dust model covering the entire Northern Hemisphere (HCMAQ) (Mathur et al., 2017), providing dynamic chemical lateral boundary conditions in a consistent physical configuration. Details of the HCMAQ simulation and its emission data can be found in Jung et al. (2021) and Vukovich et al. (2018). Since the HCMAQ simulation was based on the year 2016, we applied seasonal scaling for gaseous species based on the comparison with multiple satellite datasets such as the NASA OMI/Aura vertical O3 profile 1-Orbit level 2 data, NASA OMI tropospheric NO2 and total HCHO level 2 version 3 data, and the Measurement of Pollution in the Troposphere (MOPITT) CO level 2 version 8 data (shown in Fig. S1).

Fig. 1.

Fig. 1.

Open in a new tab

Map of the study domain and definition of urban (red) and rural (gray) regions over the CONUS. The yellow triangle symbols represent the location of ozonesonde stations (BLD: Boulder, HVA: Huntsville). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Using the Weather Research and Forecasting (WRF) model version 4.0, we estimated meteorological fields over the U.S. 12 km modeling domain (shown in Fig. 1) with the initial and boundary conditions from the National Centers for Environmental Prediction (NCEP) North American Regional Reanalysis (NARR), which has a 32 km horizontal resolution and 3 h of temporal resolution. Detailed configurations of both the CMAQ and WRF models can be found in Table 1. In an effort to enhance model performance in simulating meteorological fields, we applied indirect soil moisture and temperature nudging technique (Pleim and Gilliam, 2009; Pleim and Xiu, 2003) together with a four-dimensional data assimilation option every 6 h above the planetary boundary layer (PBL) for the temperature, the water mixing ratio, and wind components with an order of magnitude of 10−5 (Gilliam et al., 2012; Hogrefe et al., 2015).

Table 1.

Model configurations.

WRF version 4.0
Microphysics Morrison double-moment scheme
Longwave and shortwave radiation RRTMG scheme
Land surface The Pleim-Xiu land surface model (Pleim and Xiu, 1995; Xiu and Pleim, 2001)
Surface layer Pleim-Xiu surface layer (Pleim, 2006)
Planetary boundary layer The ACM2 planetary boundary layer model (Pleim, 2007a, 2007b)
Cumulus parameterization Kain-Fritsch (KF2) scheme with sub-grid cloud fraction interaction with radiation (Alapaty et al., 2012; Herwehe et al., 2014)
Four-dimensional data assimilation (FDDA) • Indirect soil moisture and temperature nudging technique (Pleim and Gilliam, 2009; Pleim and Xiu, 2003)
• A FDDA option every 6 h above the PBL for the temperature, the water vapor mixing ratio, and wind components (magnitude of 10−5) (Hogrefe et al., 2015)
Initial and boundary conditions for meteorology National Centers for Environmental Prediction (NCEP) North American Reginal Reanalysis (NARR) data
CMAQ-DDM version 5.2
Chemical mechanism CB6 and AERO6
Horizontal advection YAMO
Vertical advection WRF omega formula
Horizontal diffusion Multiscale
Vertical diffusion ACM2
Initial and boundary conditions for chemistry The CMAQ model version 5.3 with the in-line dust module covering the entire northern hemisphere in conjunction with scaling for gaseous species based on the comparison with satellite measurements
Open in a new tab

In this study, we conducted model simulations for the summers (June, July, and August) of 2011, 2014, and 2017. Each simulation was preceded by 10 days of model spin-up, allowing for initializing the deep soil moisture with the PX LSM.

2.3. Emission data

The emission data for this study comprises anthropogenic, lightning-induced, biogenic, and biomass burning emissions. Previous studies have addressed the significant contribution of non-anthropogenic sources, mainly from soil and lightning, to NOx emissions (Qu et al., 2021; Silvern et al., 2019). The absence of background NOx emissions, however, results in the significant underestimation of NO2 mixing ratios in the free troposphere and the misalignment of emission adjustments from the inversion. Therefore, before proceeding with inverse modeling, we must include proper non-anthropogenic background sources (Silvern et al., 2019; Tang et al., 2013; Travis et al., 2016).

To collect the anthropogenic emissions, we use the U.S. EPA NEI 2011 v6.3, 2014 v7.1, and 2017 (Eyth et al., 2016; Eyth and Vukovich, 2016) emission inventories for each corresponding year. The Sparse Matrix Operator Kernel Emission (SMOKE) system and associated tools were used for processing NEI to prepare the CMAQ-ready emissions. Details regarding NOx and NMVOCs emission are available in the NEI technical support documents released by the EPA. Briefly, NOx and NMVOCs emissions from NEI point sources comprising the commercial marine vessels, non-EGU (Electric Generating Utility) emissions, and oil and gas sectors are prepared using the State/Local/Tribal (S/L/T)-submitted emissions. NOx emissions from the EGU sector are based on the hourly Continuous Emissions Monitoring (CEM) emissions reported to the EPA Clean Air Markets Division (CAMD) database. The EPA Motor Vehicle Emissions Simulator (MOVES) model estimates daily emissions for NOx and calculates fuel consumption for the on-road sector based on S/L/T-submitted activity data. The MOVES also uses ratios from some primary emissions to calculate emissions for NMVOCs. Except for rail, shipping lanes (for coastal counties), and ports, which are more finely resolved to the county level, the NOx and NMVOCs emissions for stationary sectors, including non-point oil and gas, residential wood combustion, non-road, and other non-point are generated at the county level. The shipping lane emissions outside of 200 nautical miles from the U.S. coastline were filed in from non-NEI sources.

To compute biogenic emissions, we use the Biogenic Emission Inventoiy System (BEIS) version 3.61, which is built within the SMOKE, estimating hourly emissions of the biological activity of vegetative species and microbial activity from soils such as nitric oxide (NO), CO, and VOCs (e.g., isoprene, terpene, and sesquiterpene) with several meteorological variables (e.g., leaf area index, surface pressure, 2-m mixing ratio, precipitation, solar radiation, and soil properties). The use of PX LSM with indirect soil moisture and temperature nudging reduces bias in the 2 m air temperature and the 2 m water vapor mixing ratio used to calculate soil NO emission (Pleim and Gilliam, 2009; Pleim and Xiu, 2003).

We calculate lightning-induced NO emissions by the in-line lightning NO (LNO) production module implemented in the CMAQ. Estimates of the LNO emissions are based on hourly flash rates from the National Lightning Detection Network (NLDN) (Orville, 2008). A detailed description of the lightning NO production scheme can be found in (Allen et al., 2012; Kang et al., 2019). The contribution of LNO to the total NOx emissions ranges 16% and 20% over the Rocky Mountains area and the southeastern region of the U.S. for the summer of 2011 (Kang and Pickering, 2018) and 30% over the mountainous western states in July 2011 (Kang et al., 2020), affecting O3 concentrations in the upper troposphere and the surface as well as the deposition of nitrogen (Allen et al., 2012; Kang et al., 2020, 2019; Kang and Pickering, 2018).

For biomass burning emissions, we refer to the Fire Inventory from National Center for Atmospheric Research (FINN) version 1.5, which provides daily global estimates of trace gas and particulate matter emissions based on observation by the Moderate Resolution Imaging Spectroradiometer (MODIS) (Wiedinmyer et al., 2014, 2011, 2006). In this study, the gridded open fire emissions from the FINN are vertically distributed in the PBL (65%) and upper atmosphere (35%).

2.4. Inverse modeling of NOx and NMVOCs emissions

Given the relatively short lifetime of NO2 and HCHO, their satellite columns often correlate with NOx and NMVOCs emissions and can be applied to estimations of top-down emission inventories. We proceed the inverse modeling for total NOx and NMVOCs emissions after combining anthropogenic, biogenic, and biomass burning emissions. To constrain bottom-up emissions of total NOx and NMVOCs, we apply the Gauss-Newton method, an analytical inversion approach to finding solutions to problems that are not rigorously nonlinear, described in Rodgers (2000). This approach assumes no correlation between observations and emission errors, and the error covariances of observations and emissions with zero bias Gaussian probability density functions. The inversions with the Newtonian iterations gradually approach convergence at the zero of the gradients of the cost function, shown in Eq. (1). Posteriori emissions can be estimated by Eq. (2).

J(x)=12(y−F(x))TSo−1(y−F(x))+12(x−xa)TSe−1(x−xa) (1)
xi+1=xa+SeKiT(KiSeKiT+So)−1[y−F(xi)+Ki(xi−xa)] (2)

where x and xa represent a posteriori and a priori emissions, respectively. The observational error covariances (So) for the OMI NO2 and HCHO are based on column uncertainties provided in the OMI dataset in conjunction with uncertainties validated by previous studies (González Abad et al., 2015; Lamsal et al., 2014). To find the error covariance of emission (Se), we calculate the total error covariance by combining all sector emissions (NOx/NMVOCs) from anthropogenic (50%/150%), biogenic (200%/200%), and biomass burning (100%/300%) sources. We note that we considered NMVOCs species having relatively shorter lifetimes, such as HCHO, ethene, acetaldehyde, isoprene, ethanol, and acetone. This study applies the CMAQ DDM-3D model to correlate the columns of NO2 and HCHO with their emissions (F) (first-order sensitivity) in order to calculate the Jacobian matrix (K).

3. Results

3.1. Changes in the NO2 and HCHO column retrievals and the contribution of background sources to the upper troposphere

Fig. 2 illustrates the spatial distribution of seasonally averaged tropospheric NO2 and total HCHO vertical column densities from the OMI for the summers (June to August) of 2011, 2014, and 2017 over the CONUS. The AMFs of both columns were adjusted based on the NO2 and HCHO columns simulated in this study. The OMI tropospheric NO2 columns showed large values in cities with high population density and industrial activity, such as Seattle, Los Angeles, Houston, Chicago, and New York. During the study period, the NO2 columns exhibited a similar spatial distribution, but with a domain-wide decrease by −6.10% and −7.95% in 2014 and 2017, respectively, with respect to 2011. With regard to HCHO columns, high values were predominant in the southeastern U.S., mainly resulting from oxidation of biogenically emitted isoprene (Palmer et al., 2003; Zhu et al., 2017). In this region, HCHO columns consistently declined by −26.16% in 2014 and – 33.68% in 2017 compared to those in 2011, while more widespread high values were found over the western area.

Fig. 2.

Fig. 2.

Open in a new tab

Spatial distribution of (a to c) the OMI tropospheric NO2 vertical column density and (d to f) the OMI HCHO vertical column density in the summers (June–August) of 2011, 2014, and 2017.

The variability of satellite-observed NO2 and HCHO columns reflects changes in bottom-up emissions and the contribution of background sources in the upper troposphere (Silvern et al., 2019). In Fig. 3, we plotted seasonal averaged changes in observed and simulated tropospheric NO2 and total HCHO column densities in four regions: domainwide (i.e., CONUS), urban and rural regions, and southeastern (SE) areas. We defined an urban area based on the U.S. census-based surrogate, where population density is at least 1000 people per square mile, denoted in red in Fig. 1; and we defined a rural region, designated by gray in Fig. 1, as more than 20 km away from both urban regions and power plants with a capacity of 1 MW or more, based on the U.S. Energy Information Administration. The southeastern (SE) area (a blue rectangle shown in Fig. 1) denotes an area with dominant lightning-induced (to the total NOx emissions) and biogenic emissions (to the total NMVOCs emissions) over the CONUS. In SE, NMVOCs emissions from biogenic sources are dominant, so we excluded the SE area when we analyzed modeled and observed HCHO columns over urban and rural regions in Fig. 3.

Fig. 3.

Fig. 3.

Open in a new tab

The long-term trend in summertime averaged tropospheric (a-1 to h-1) NO2 and (a-2 to h-2) HCHO columns observed by the OMI satellite (2011–2018) and simulated by the CMAQ model (2011, 2014, and 2017), and their relative changes in the CONUS and urban, rural, and southeastern U.S. regions (shown in Fig. 1). CMAQ simulations were conducted with the prior emission and three background setups (with dynamic lateral boundary conditions and lightning emissions (LNO), with dynamic lateral boundary conditions and without lightning emissions, and with profile lateral boundary conditions and without lightning emissions).

Also shown in Fig. 3 are modeled NO2 and HCHO columns with three different background setups that show the contribution of background sources to the upper troposphere: those with dynamic lateral boundary conditions and LNO emissions, those with dynamic lateral boundary conditions and no LNO emissions, and those with profile (static) lateral boundary conditions and no LNO emissions. We note that we weighted the contributions from lightning and the transport of air pollution for the background sources in the upper troposphere, while we included open fire and soil emissions in prior emissions, as addressed in Section 2.3. Also, surface meteorological variables (e.g., solar radiation, air temperature, and wind fields) vary concentrations of NO2 and HCHO by altering photolysis rates, lifetime, and dispersion (Goldberg et al., 2020), as well as rates of biogenic isoprene emissions (Guenther et al., 2012). Therefore, the model performance of the meteorological variable simulation is critically linked to the reliability of both concentrations of air pollutants and the estimates of top-down emissions. We evaluated modeled hourly air temperatures and wind fields against the U.S. EPA Air Quality System (AQS) measurements over the CONUS, listed in Table 2. We found that our model and the observations were comparable, showing a strong correspondence (air temperature: IOA > 0.89, wind fields: IOA > 0.61).

Table 2.

Model performance for the simulation of meteorological variables (i.e., air temperature, U and V wind components) compared to the AQS measurements.

Variables Number of stations Mean
 RMSE CORR IOA MB
OBS MOD
2011 JJA Air Temp ([°C]) 573 23.87 25.12 3.57 0.87 0.90 1.29
U wind ([m/s]) 552 0.21 0.59 2.12 0.50 0.63 0.38
V wind ([m/s]) 552 0.58 0.61 2.13 0.46 0.61 0.03
2014 JJA Air Temp ([°C]) 558 23.18 24.26 3.23 0.86 0.89 1.08
U wind ([m/s]) 620 0.21 0.56 2.04 0.50 0.63 0.34
V wind ([m/s]) 620 0.45 0.40 2.20 0.50 0.62 −0.05
2017 JJA Air Temp ([°C]) 600 24.10 24.95 3.15 0.87 0.90 0.84
U wind ([m/s]) 691 0.26 0.61 2.00 0.48 0.62 0.36
V wind ([m/s]) 691 0.31 0.32 2.16 0.49 0.62 0.01
Open in a new tab

We considered only stations with valid data for more than 50% of the study period. (RMSE: root-mean-square error, CORR: correlation, IOA: the index of agreement, MB: mean bias).

The OMI NO2 columns showed marginal changes over the CONUS during the study period; the change in urban regions, however, exhibited a decrease throughout the 2011–2017 period, which indicates the continuous efforts to reduce NOx emissions. The NO2 column in the urban and SE regions decreased by −20.83% and – 16.53%, respectively, while almost steady columns were found in the rural regions throughout the study period, possibly related to the increasing contribution of non-anthropogenic sources (e.g., soil and lightning) that compensated for the reduction in anthropogenic emissions (Li and Wang, 2019; Silvern et al., 2019). Although the model simulated a general decreasing tendency of the NO2 column, it showed a sustained underestimation compared to the OMI observations over the CONUS by 6.24 × 1014 – 1.37 × 1015 molecules cm−2 (11.58–62.78%) in the case of prior emissions and profile lateral boundary conditions. We found that the addition of dynamic lateral boundary conditions and the LNO emissions greatly reduced bias by 0.11–26.58%, resulting in stronger agreement with observations, dominantly in both rural regions (16.12–18.07%) and SE regions (4.65–10.54%), primarily benefited from LNO emissions.

The OMI HCHO columns showed a steady decrease, except for the rural regions (which decreased by 5.87% in 2014 and increased by 2.75% in 2017 with respect to 2011). Since meteorology, especially temperature, drives most of the variability of HCHO columns over the CONUS (Zhu et al., 2017), a significant decline in air temperature (shown in fig. S2), along with the reduction in NEI emissions, might cause a substantial decrease over the SE regions, where biogenic sources are dominant. Overall, the CMAQ model overestimated HCHO columns by up to 2.64 × 1015 molecules cm−2 (56.79%) compared to the OMI observations, and dynamic lateral boundary conditions amplified this overestimation (0.75 × 1015 – 3.58 × 1015 molecules cm−2 (5.30–67.60%)). In addition, the discrepancy between the OMI HCHO columns and the modeled columns were significantly large over the rural regions, which might be attributable to instrumental noise of the column retrieval, or rather the high uncertainty of the bottom-up emission inventory, particularly involving open fire emissions (e.g., injection height, the fraction of burned areas, classifications of land use and land cover, and unresolved small fires by satellites) (Hawbaker et al., 2008; Jeon et al., 2018; Mehmood et al., 2020; Shen et al., 2019; Souri et al., 2017; Wiedinmyer et al., 2011).

When compared with NO2 and O3 concentrations measured by P—3B aircraft during the NASA DISCOVER-AQ 2011 and 2014 campaigns, which took place over the Baltimore/Washington D.C. metropolitan area and northern Colorado, shown in Fig. 4 (flight tracks, flight time, and altitude were shown in Fig. S3), the base model tended to underpredict both NO2 and O3 concentrations in the upper troposphere, as reported in Napelenok et al. (2008), possibly owing to lack of emission sources and a chemical lifetime of NO2. Promisingly, the model performance greatly improved in the upper troposphere (above 2.5 km) with the dynamic lateral boundary conditions (as much as 23.24% for NO2 and 14.76% for O3) and LNO emissions (as much as 9.33% for NO2 and 26.34% for O3). An enhancement of model performance in the upper troposphere was also found by comparison with ozonesonde measurements, displayed in Fig. S4. The electrochemical concentration cell (ECC) attached to ozonesonde provides simultaneous measurements of ozone and meteorological quantities until the balloon bursts about at altitudes of 35 km (Tarasick et al., 2021). Improvements in the model results suggest the need to account for background sources in the free troposphere, which would prevent the misalignment of emission adjustments, discussed in the following section.

Fig. 4.

Fig. 4.

Open in a new tab

Comparisons of the vertical distribution of ozone and nitrogen dioxide (NO2) concentrations measured during the DISCOVER-AQ (a to d) 2011 (July 21 and 22) and (e to h) 2014 (July 29 and 31) campaigns and modeled concentrations with prior emissions and three background setups (i.e., with dynamic lateral boundary conditions and LNO emissions, with dynamic lateral boundary conditions and without LNO emissions, and with profile lateral boundary conditions and without LNO emissions). Error bars represent the standard deviation.

3.2. Top-down estimation of NOx and NMVOCs emissions

Before investigating changes in the ozone chemical production regime, we adjusted total NOx and NMVOCs emissions with the OMI NO2 and HCHO column retrievals found by the inverse modeling approaches that we addressed in Section 2.4 in order to acquire accurate knowledge of the concentrations of ozone precursors. It is worth noting that we assumed the uncertainty of bottom-up emissions as the main contributor to the discrepancy between modeled (with prior emissions, dynamic boundary conditions, and LNO emissions) and satellite-observed columns; the discrepancy, however, could also be explained by several other deficiencies in chemical processes (e.g., a lifetime of alkyl nitrate) and meteorology (e.g., wind fields).

Figs. 5 and 6 display the spatial distribution of averaged tropospheric NO2 and total HCHO column densities simulated with prior and posterior emissions corresponding to the OMI observations (shown in Fig. 2) and their differences with respect to the column with prior emissions. For both the NO2 and HCHO columns, the modeled columns with posterior emissions more closely agreed with those from the OMI observations. In most areas, the modeled tropospheric NO2 column, compared to that of the OMI observations, exhibited an underestimation throughout the study period (Figs. 2 and 5), but using posterior emission (shown in Fig. S5), we observed a large reduction in bias, increasing the NO2 column by as much as 21.55 ± 28.76% nationwide. The adjustment ratio between the prior and posterior NOx emissions was slightly high in 2017 (2.84 ± 1.99) than in 2011 (2.33 ± 1.83) and 2014 (2.62 ± 2.19) over the CONUS (Fig. S8), predominantly in the central and western regions, shown in Fig. S5. The absolute amount of NOx emissions is much smaller over rural regions compared to urban regions, possibly resulting in an excessive emission adjustment ratio over the CONUS. As shown in Fig. S8, the posterior NOx emissions are 7–34% higher than that of prior emissions over urban regions. We also found increased NOx emissions and NO2 columns in Mexico and Canada, which in turn potentially enhanced the influence of O3 precursors transported from outside of the CONUS. With regard to HCHO, the model with prior emissions overestimated the total HCHO column over the southeastern U.S. by 6.77–14.85%. The inversion suggested 0.81 ± 0.30 times fewer NMVOCs emissions over the region in 2011, and the adjustment ratio decreased as 0.63 ± 0.28 in 2014 and 0.59 ± 0.24 in 2017, respectively (Fig. S5). In addition, we observed enhancement of the HCHO column over the northwestern U.S. with posterior emissions, possibly related to open fire events; the magnitude, however, was relatively small (Fig. S6).

Fig. 5.

Fig. 5.

Open in a new tab

Spatial distribution of the modeled NO2 vertical column densities with prior and posterior emissions and their relative differences with respect to the modeled column density with prior emissions in the summers (June–August) of 2011, 2014, and 2017.

Fig. 6.

Fig. 6.

Open in a new tab

Spatial distribution of the modeled HCHO vertical column densities with prior and posterior emissions, and their relative differences with respect to the modeled column density with prior emissions in the summers (June–August) 2011, 2014, and 2017.

To validate the model output with prior and posterior emissions, we compared the AQS measurements to modeled surface daytime (12–17 LST) averaged NO2 and maximum daily 8-h average (MDA8) O3 concentrations over the CONUS, shown in Fig. S9. As the number of AQS sites for each year varied throughout the study period, we used measurement data from monitoring sites (230 sites for NO2 and 1023 sites for O3) with more than 50% of temporal coverage to ensure both data quality and consistency. Following Lamsal et al. (2008), we also corrected NO2 measurements for the AQS sites to remove any interferences from several reactive oxidized nitrogen-containing species (e.g., alkyl nitrate, peroxyacetyl nitrate (PAN), and nitric acid (HNO3)) based on the modeled output. The performance of the model with the posterior emissions was superior, with a significantly reduced bias of 38–96% for NO2. However, posterior emissions, unfortunately, enhanced the tendency of the general model to overestimate O3 concentrations, frequently reported in previous studies (Hogrefe et al., 2018; Kitayama et al., 2019; Travis et al., 2016). More precisely, we compared model performance at the percentile of O3 distribution in Fig. S10. The model with the posterior emissions amplified the overestimation of ozone at the lower percentile (< 60%), where the majority of ozone data was presented. On the contrary, underestimation of ozone at a higher percentile (> 60%) slightly reduced with the posterior emission in 2011; however, model bias increased in 2014 and 2017. These results indicate that adjustments to the emissions of O3 precursors cannot fully address the total uncertainty of the model, and the bias is associated more with systematic problems in the model that modulate surface O3 such as vertical mixing highly related to meteorological factors (e.g., heat flux), as well as deficient ozone deposition in the boundary layer (Hogrefe et al., 2018; Travis et al., 2016).

3.3. Changes in the ozone chemical regime over the CONUS

The HCHO/NO2 ratio as an indicator of the ozone chemical regime is highly dependent on meteorological and photochemical conditions, as well as characteristics of satellite retrievals and CTMs, resulting in a wide range of values. For instance, Martin et al. (2004) proposed the HCHO/NO2 ratio of 1 for defining NOx-saturated (< 1) and NOx-limited (> 1) regimes. Choi et al. (2012) and Duncan et al. (2010) used the ratios of 1 and 2 to categorize the ozone chemical regime into NOx-saturated (< 1), NOx-limited (> 2), and transition (1 < HCHO/NO2 < 2) regimes. Chang et al. (2016) found different threshold values of ozone chemical regime with OMI (1.5 and 2.3) and Global Ozone Monitoring Experiment-2 (GOME-2) (1 and 1.85). Ozone production sensitivity to the HCHO/NO2 ratio derived by CTMs could vary depending upon the study domain, chemical mechanism, and meteorological conditions (Chang et al., 2016; Jin et al., 2017). To avoid the misclassification of the ozone chemical regime, we conducted an experiment to define threshold values by using the CMAQ model with posterior emission inventory based on the year 2011. Fig. 7 (a) represents the difference of averaged surface O3 concentrations between the baseline CMAQ model and the CMAQ with perturbed emission (20% reductions in NOx and NMVOCs emissions) along with the corresponding modeled HCHO/NO2 column ratio for the revisit time of the OMI satellite (13:00–14:00 LST) since we adjusted emissions of ozone precursors based on OMI column retrievals. In addition, the revisit time of the OMI satellite is beneficial to analyze the ozone chemical regime when the ozone production is maximum in a day in conjunction with the smallest solar zenith angle and most profound boundary layer height (Jin et al., 2017). To clearly examine the difference of O3 with perturbed emission, we considered model grids over polluted regions where the modeled tropospheric NO2 column is greater than 1 × 1015 molecules cm−2 (Boersma et al., 2016; Choi et al., 2012; Schaub et al., 2006). As a response to NOx emission reduction, surface O3 concentrations generally increased proportionally to the HCHO/NO2 ratio. The increase was dominant for a high HCHO/NO2 ratio. The change in surface ozone concentrations became moderate for the case of NMVOCs emission reduction (red color); still, noticeable changes in ozone concentration were found for a low HCHO/NO2 ratio.

Fig. 7.

Fig. 7.

Open in a new tab

(a) Difference between the surface ozone concentrations of the baseline model output (with posterior emissions) and emission reduction cases (20% reduction in NOx and NMVOCs emissions) with the ratio of the modeled HCHO column to the NO2 column in the summer of 2011; (b) the cumulative probability of NOx-limited and NOx-saturated conditions with respect to the ratio of the modeled HCHO column to the NO2 column following Jin et al. (2017). Both are estimated from daily averaged HCHO/NO2 values and surface O3 concentrations for the revisit time of the OMI satellite (13:00–14:00 LST) over polluted regions (modeled NO2 column >1 × 1015 molecules cm−2).

In order to ascertain threshold values of the chemical regime, we plotted the cumulative probabilities of NOx-limited and NOx-saturated conditions with respect to HCHO/NO2 column ratio by following Jin et al. (2017) as shown in Fig. 7(b). A more detailed description of the cumulative probabilities and defining the cut-off values can be found in Jin et al. (2017). Briefly, the NOx-saturated condition represents the model grids where surface O3 concentrations decrease with a 20% reduction of NOx emissions. The NOx-limited condition is defined in the model grids where the increase of O3 concentrations with a 20% reduction of NOx emissions is greater than that with a 20% reduction of NMVOCs emissions. The cumulative probability for NOx-saturated (NOx-limited) condition indicates the proportion of model grids that belong to the condition and have the HCHO/NO2 ratio less (greater) than a given value. As with Fig. 7 (a), both of the cumulative probabilities were estimated only for the revisit time of the OMI satellite (13:00–14:00 LST), and we excluded model grids having the modeled tropospheric NO2 column less than 1 × 1015 molecules cm−2. The threshold values for NOx-limited and NOx-saturated regimes are estimated as those corresponding to the 95% of cumulative probability. Accordingly, in this study, HCHO/NO2 ratios of 1.6 and 2.6 are determined as the regime thresholds. Thus, HCHO/NO2 ratios lower than 1.6 represent the NOx-saturated regime, ratios greater than 2.6 represent NOx-limited regime, while those in-between (1.6 < HCHO/NO2 < 2.6) defined the transition regime.

Fig. 8 (a) shows the spatial distribution of the ozone chemical production regime (based on the averaged HCHO/NO2 ratio) over the CONUS in the summer of 2011, while the results for 2014 and 2017 are plotted in Fig. S11. The entire CONUS is divided into three regimes according to the threshold values we defined earlier; model grids with NO2 column density less than 1 × 1015 molecules cm−2 are not considered and are denoted by the gray color. Several metropolitan cities were categorized as NOx-saturated and transition regimes, whereas the remaining rural areas were classified as NOx-limited regimes due to the nature of urban conditions with highly concentrated traffic density and amount of vehicle emissions. According to Fig. 8(b), showing the difference of HCHO/NO2 ratio between 2011 and 2017, the column ratio in rural areas decreased by −1.30 ± 1.71, mostly over the eastern and southeastern U.S. This regime modification toward the NOx-saturated condition could be explained by the increased NOx emissions over the northeastern U.S. (Fig. S7), as well as the sustained decrease of the HCHO column (Fig. 3 (h-2)) as a response to meteorological change (Fig. S2).

Fig. 8.

Fig. 8.

Open in a new tab

(a) Spatial distribution of the ozone chemical production regime over the CONUS for the year 2011. (b) The difference between the HCHO/NO2 ratios of 2011 and 2017. Gray areas represent tropospheric NO2 column densities that are less than 1 × 1015 molecules cm−2.

To examine the changes in the ozone chemical regime over major metropolitan cities more closely, we plotted the spatial distribution of the ozone production regime zoomed in over 5 boxes (Box A to Box E, denoted in Fig. 8) in Fig. 9, showing details of the selected metropolitan areas surrounding Seattle/Portland, Los Angeles/San Francisco, Dallas/Houston, Chicago, and Detroit/Pittsburgh/New York for the summers of 2011, 2014, and 2017. In addition, Fig. 10 shows changes in HCHO/NO2 ratio at urban cores for the major 12 cities over the CONUS. The change of the ozone production regime varied during the study period. As a response to the reduction of NO2 columns in urban regions, which was greater than that of HCHO columns (Fig. S8), the clear shifts of the ozone chemical regime from NOx-saturated to transition, or transition to NOx-limited were found partially over metropolitan areas (e.g., Seattle, Houston, and Pittsburgh) or completely even at urban cores (e.g., Houston, San Francisco, and Salt Lake City), especially from 2014 to 2017. By contrast, Los Angeles, New York, and Portland were the cities steadily remaining in the NOx-saturated regime throughout the study period. Also, we observed the sudden expansions of NOx-saturated conditions at the urban cores of Houston (in 2014) and Detroit (in 2017). In general, most cities in the eastern U.S., except Pittsburgh, tended to stay in the same regime to which they were in 2011 or even came back to the NOx-saturated regime compared to the cities in the western U.S. Many recent studies have reported the changes in ozone chemical regime over the major U.S. metropolitan areas toward NOx-limited regime in general, but with a narrow range in the reported year when the regime change occurred, and cities still remained in NOx-limited condition, which is related to the dataset, study period, and methodology of each study; Chang et al. (2016) showed that ozone formation in Boston, Pittsburgh, Philadelphia, and Washington has changed to NOx-limited or transition regimes in 2007–2014, while New York City mostly stayed in the transition regime (2007–2009) and VOC-limited regime (2007–2009 and 2014). Jin et al. (2020) addressed that NOx-saturated condition is occurring only in Los Angeles, Chicago, and New York City by 2013–2016, and Koplitz et al. (2021) showed regime change toward NOx-limited condition over most of the U.S., and a few cities such as New York City, Chicago, Minneapolis, San Francisco, and Los Angeles are still remained in NOx-saturated regime as in 2016.

Fig. 9.

Fig. 9.

Open in a new tab

Changes in the ozone production regime over the 5 boxes denoted in Fig. 8.

Fig. 10.

Fig. 10.

Open in a new tab

Changes in the HCHO/NO2 ratio from urban cores for 12 main cities over (a) the western U.S. (Seattle, Portland, San Francisco, Los Angeles, Salt Lake City, and Houston), and (b) the eastern U.S. (Minneapolis, Chicago, Detroit, Pittsburgh, New York, and New Orleans).

4. Conclusion and discussion

In this study, we examined changes in the ozone chemical regime over the CONUS in the summers of 2011, 2014, and 2017, years in which U.S. NEI data were based. To combine the benefits of both satellite retrievals and CTMs, we employed an inverse modeling technique with OMI tropospheric NO2 and total HCHO column retrievals to estimate NOx and NMVOCs emissions. To avoid misalignment of the emission adjustment, we applied dynamic chemical lateral boundary conditions and lightning-induced nitric oxide emissions, resulting in better representation of NO2 and O3 concentration in the upper atmosphere, confirmed by comparisons with measurements from the DISCOVER-AQ campaigns and two ozonesonde sites.

Compared to OMI observations, the model with the top-down estimated NOx and NMVOCs emissions showed remarkably better performance for simulating tropospheric NO2 and total HCHO columns and alleviated the nationwide underestimation of the NO2 column and the overestimation of the HCHO column appearing over the southeastern U.S. We then used the CMAQ model to determine the threshold values of the ozone production regime, which closely represented the ozone chemical conditions of this study: NOx-saturated (HCHO/NO2 < 1.6), a transition (1.6 < HCHO/NO2 < 2.6), and NOx-limited (HCHO/NO2 > 2.6), which is slightly distinct from other studies by Chang et al. (2016), Choi et al. (2012), Duncan et al. (2010) and Jin et al. (2017). This is because the threshold value of the HCHO/NO2 ratio represents the unique ozone chemistry of the CTM depending upon several influencing factors such as study domain, chemical mechanism, and meteorological condition; therefore, the applicability of the threshold value in other studies under different conditions is a subject of further investigation. During the study period, many of the major metropolitan areas over the CONUS, classified as NOx-saturated conditions in the year 2011, showed a clear shift toward a NOx-limited condition followed by a greater reduction in the NO2 column than that in the HCHO column. A few of the cities, mostly located in the eastern U.S., did not show a significant change in their ozone chemical regimes. Interestingly, we found that despite the substantial reduction in the NO2 columns, changes in meteorological conditions such as air temperature and precipitation exhibited a decrease in the HCHO column over the southeastern U.S. As a result, the ozone regime over the region slightly moved toward a NOx-saturated condition.

In spite of our best efforts to gain accurate knowledge of the emissions of ozone precursors, we observed an amplified overestimation of modeled ozone concentrations at the surface. This finding indicates systematic model bias stemming from uncertainty in chemical and physical processes, which lies beyond the scope of this study. Also, the inverse modeling method we applied in this study does not fully address the effect of transport, which could be mitigated by applying more advanced approaches such as the four-dimensional variational (4D-Var) method.

This study showed remarkable spatial and temporal variation of the ozone chemical regime over the CONUS. As a result of continuous efforts devoted to decreasing anthropogenic emissions and rapid climate change, the ozone chemical regime could vary more frequently, which might require further adjustments in current ozone control policies. Thus, armed with a better knowledge of ozone precursors and ozone chemical conditions, those charged with implementing appropriate emission control policies in a timely matter could benefit from the approach used in this study.

Supplementary Material

Supplement1

Acknowledgments

We authors truly appreciate Dr. Kiran Alapaty, Dr. Rohit Mathur, Dr. Barron Henderson, and Dr. David C. Wong (U.S. EPA) for their invaluable comments and suggestions. This study was supported by NASA’s Aura Science Team Grant (NNH19ZDA001N-AURAST). We are grateful for the support of the Research Computing Data Core at the University of Houston for assistance with the calculations carried out in this work.

Footnotes

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.atmosres.2022.106076.

Publisher's Disclaimer: Disclaimer

This paper has been subjected to an EPA review and approved for publication. The views expressed here are those of the authors and do not necessarily reflect the views or policies of the U.S. Environmental Protection Agency (EPA).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability statement

The OMI/Aura Level 2 tropospheric NO2 and total HCHO column data products are available from the NASA Goddard Earth Science (GES) Data and Information Services Center (DISC) (https://disc.gsfc.nasa.gov/datasets/OMNO2_003/summary?keywords=no2, https://disc.gsfc.nasa.gov/datasets/OMHCHO_003/summary?keywords=hcho). The NCEP NARR data can be downloaded from https://rda.ucar.edU/datasets/ds608.0/. The FINN data can be acquired from https://www.acom.ucar.edu/Data/fire. The lightning flash data are available for purchase from Vaisala Inc. (https://www.vaisala.com/en/products/systems/lightning-detection). For this study, the native form of the lightning flash data was provided by Dr. Hyuncheol Kim at the NOAA, and processed into the required format of the CMAQ model with the considerable assistance of Dr. Daiwen Kang at the EPA. The EPA AQS data (pre-generated) for meteorology and air quality measurement over the U.S. is available from https://aqs.epa.gov/aqsweb/airdata/download_files.html. Aircraft measurement during the DISCOVER-AQ 2011 (Baltimore-Washington, D.C.) and 2014 (Colorado) campaigns were distributed by NASA (https://www-air.larc.nasa.gov/missions/discover-aq/discover-aq.html). Ozonesonde data is downloaded from the NOAA Global Monitoring Laboratory/Earth System Research Laboratories (https://gml.noaa.gov/dv/data/index.php?category=Ozone&type=Balloon). The model outputs from the primary simulations can be downloaded from doi: https://doi.org/10.5281/zenodo.5590569.

References

  1. Alapaty K, Herwehe JA, Otte TL, Nolte CG, Bullock OR, Mallard MS, Kain JS, Dudhia J, 2012. Introducing subgrid-scale cloud feedbacks to radiation for regional meteorological and climate modeling. Geophys. Res. Lett 39 10.1029/2012GL054031. [DOI] [Google Scholar]
  2. Allen DJ, Pickering KE, Pinder RW, Henderson BH, Appel KW, Prados A, 2012. Impact of lightning-NO on eastern United States photochemistry during the summer of 2006 as determined using the CMAQ model. Atmos. Chem. Phys 12, 1737–1758. 10.5194/acp-12-1737-2012. [DOI] [Google Scholar]
  3. Bae C, Kim HC, Kim B-U, Kim S, 2020. Surface ozone response to satellite-constrained NOx emission adjustments and its implications. Environ. Pollut 258, 113469 10.1016/j.envpol.2019.113469. [DOI] [PubMed] [Google Scholar]
  4. Baek KH, Kim JH, Park RJ, Chance K, Kurosu TP, 2014. Validation of OMI HCHO data and its analysis over Asia. Sci. Total Environ 490, 93–105. 10.1016/j.scitotenv.2014.04.108. [DOI] [PubMed] [Google Scholar]
  5. Boersma KF, Vinken GCM, Eskes HJ, 2016. Representativeness errors in comparing chemistry transport and chemistry climate models with satellite UV–Vis tropospheric column retrievals. Geosci. Model Dev 9, 875–898. 10.5194/gmd-9-875-2016. [DOI] [Google Scholar]
  6. Bucsela EJ, Krotkov NA, Celarier EA, Lamsal LN, Swartz WH, Bhartia PK, Boersma KF, Veefkind JP, Gleason JF, Pickering KE, 2013. A new stratospheric and tropospheric NO2 retrieval algorithm for nadir-viewing satellite instruments: applications to OMI. Atmos. Meas. Tech 6, 2607–2626. 10.5194/amt-6-2607-2013. [DOI] [Google Scholar]
  7. Chang C-Y, Faust E, Hou X, Lee P, Kim HC, Hedquist BC, Liao K-J, 2016. Investigating ambient ozone formation regimes in neighboring cities of shale plays in the Northeast United States using photochemical modeling and satellite retrievals. Atmos. Environ 142, 152–170. 10.1016/j.atmosenv.2016.06.058. [DOI] [Google Scholar]
  8. Choi Y, Kim H, Tong D, Lee P, 2012. Summertime weekly cycles of observed and modeled NOx and O3 concentrations as a function of satellite-derived ozone production sensitivity and land use types over the Continental United States. Atmos. Chem. Phys 12, 6291–6307. 10.5194/acp-12-6291-2012. [DOI] [Google Scholar]
  9. Choi S, Lamsal LN, Follette-Cook M, Joiner J, Krotkov NA, Swartz WH, Pickering KE, Loughner CP, Appel W, Pfister G, Saide PE, Cohen RC, Weinheimer AJ, Herman JR, 2020. Assessment of NO2 observations during DISCOVER-AQ and KORUS-AQ field campaigns. Atmos. Meas. Tech 13, 2523–2546. 10.5194/amt-13-2523-2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cohan DS, Hakami A, Hu Y, Russell AG, 2005. Nonlinear response of Ozone to emissions: source apportionment and sensitivity analysis. Environ. Sci. Technol 39, 6739–6748. 10.1021/es048664m. [DOI] [PubMed] [Google Scholar]
  11. Curci G, Palmer PI, Kurosu TP, Chance K, Visconti G, 2010. Estimating European volatile organic compound emissions using satellite observations of formaldehyde from the Ozone Monitoring Instrument. Atmos. Chem. Phys 10, 11501–11517. 10.5194/acp-10-11501-2010. [DOI] [Google Scholar]
  12. Duncan BN, Yoshida Y, Olson JR, Sillman S, Martin RV, Lamsal L, Hu Y, Pickering KE, Retscher C, Allen DJ, Crawford JH, 2010. Application of OMI observations to a space-based indicator of NOx and VOC controls on surface ozone formation. Atmos. Environ 44, 2213–2223. 10.1016/j.atmosenv.2010.03.010. [DOI] [Google Scholar]
  13. Eyth A, Vukovich J, 2016. Technical Support Document (TSD) Preparation of Emissions Inventories for the Version 6.3, 2011 Emissions Modeling Platform. US Environmental Protection Agency, Office of Air Quality Planning and Standards. [Google Scholar]
  14. Eyth A, Vukovich J, Farkas C, Strum M, 2016. Technical Support Document (TSD): Preparation of Emissions Inventories for the Version 7.1–2016 North American Emissions Modeling Platform. [Google Scholar]
  15. Foley KM, Hogrefe C, Pouliot G, Possiel N, Roselle SJ, Simon H, Timin B, 2015. Dynamic evaluation of CMAQ part I: separating the effects of changing emissions and changing meteorology on ozone levels between 2002 and 2005 in the eastern US. Atmos. Environ 103, 247–255. 10.1016/j.atmosenv.2014.12.038. [DOI] [Google Scholar]
  16. Gilliam RC, Godowitch JM, Rao ST, 2012. Improving the horizontal transport in the lower troposphere with four dimensional data assimilation. Atmos. Environ 53, 186–201. 10.1016/j.atmosenv.2011.10.064. [DOI] [Google Scholar]
  17. Goldberg DL, Saide PE, Lamsal LN, de Foy B, Lu Z, Woo J-H, Kim Y, Kim J, Gao M, Carmichael G, Streets DG, 2019. A top-down assessment using OMI NO2 suggests an underestimate in the NOx emissions inventory in Seoul, South Korea, during KORUS-AQ. Atmos. Chem. Phys 19, 1801–1818. 10.5194/acp-19-1801-2019. [DOI] [Google Scholar]
  18. Goldberg DL, Anenberg SC, Griffin D, McLinden CA, Lu Z, Streets DG, 2020. Disentangling the Impact of the COVID-19 Lockdowns on Urban NO2 from Natural Variability. Geophys. Res. Lett 47 10.1029/2020GL089269. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. González Abad G, Liu X, Chance K, Wang H, Kurosu TP, Suleiman R, 2015. Updated Smithsonian Astrophysical Observatory Ozone Monitoring Instrument (SAO OMI) formaldehyde retrieval. Atmos. Meas. Tech 8, 19–32. 10.5194/amt-8-19-2015. [DOI] [Google Scholar]
  20. Guenther AB, Jiang X, Heald CL, Sakulyanontvittaya T, Duhl T, Emmons LK, Wang X, 2012. The Model of Emissions of Gases and Aerosols from Nature version 2.1 (MEGAN2.1): an extended and updated framework for modeling biogenic emissions. Geosci. Model Dev 5, 1471–1492. 10.5194/gmd-5-14712012. [DOI] [Google Scholar]
  21. Hawbaker TJ, Radeloff VC, Syphard AD, Zhu Z, Stewart SI, 2008. Detection rates of the MODIS active fire product in the United States. Remote Sens. Environ 112, 2656–2664. 10.1016/j.rse.2007.12.008. [DOI] [Google Scholar]
  22. Herman J, Cede A, Spinei E, Mount G, Tzortziou M, Abuhassan N, 2009. NO 2 column amounts from ground-based Pandora and MFDOAS spectrometers using the direct-sun DOAS technique: Intercomparisons and application to OMI validation. J. Geophys. Res 114 10.1029/2009JD011848. [DOI] [Google Scholar]
  23. Herwehe JA, Alapaty K, Spero TL, Nolte CG, 2014. Increasing the credibility of regional climate simulations by introducing subgrid-scale cloud-radiation interactions: RCM sims with Cu-radiation interactions. J. Geophys. Res.-Atmos 119, 5317–5330. 10.1002/2014JD021504. [DOI] [Google Scholar]
  24. Hogrefe C, Pouliot G, Wong D, Torian A, Roselle S, Pleim J, Mathur R, 2015. Annual application and evaluation of the online coupled WRF–CMAQ system over North America under AQMEII phase 2. Atmos. Environ 115, 683–694. 10.1016/j.atmosenv.2014.12.034. [DOI] [Google Scholar]
  25. Hogrefe C, Liu P, Pouliot G, Mathur R, Roselle S, Flemming J, Lin M, Park RJ, 2018. Impacts of different characterizations of large-scale background on simulated regional-scale ozone over the continental United States. Atmos. Chem. Phys 18, 3839–3864. 10.5194/acp-18-3839-2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Holton JR, Haynes PH, McIntyre ME, Douglass AR, Rood RB, Pfister L, 1995. Stratosphere-troposphere exchange. Rev. Geophys 33, 403. 10.1029/95RG02097. [DOI] [Google Scholar]
  27. Jacob DJ, Logan JA, Murti PP, 1999. Effect of rising Asian emissions on surface ozone in the United States. Geophys. Res. Lett 26, 2175–2178. 10.1029/1999GL900450. [DOI] [Google Scholar]
  28. Jeon W, Choi Y, Souri AH, Roy A, Diao L, Pan S, Lee HW, Lee S-H, 2018. Identification of chemical fingerprints in long-range transport of burning induced upper tropospheric ozone from Colorado to the North Atlantic Ocean. Sci. Total Environ 613–614, 820–828. 10.1016/j.scitotenv.2017.09.177. [DOI] [PubMed] [Google Scholar]
  29. Jiang Z, McDonald BC, Worden H, Worden JR, Miyazaki K, Qu Z, Henze DK, Jones DBA, Arellano AF, Fischer EV, Zhu L, Boersma KF, 2018. Unexpected slowdown of US pollutant emission reduction in the past decade. Proc. Natl. Acad. Sci 115, 5099–5104. 10.1073/pnas.1801191115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Jin X, Fiore AM, Murray LT, Valin LC, Lamsal LN, Duncan B, Folkert Boersma K, De Smedt I, Abad GG, Chance K, Tonnesen GS, 2017. Evaluating a space-based indicator of surface ozone-NO x -VOC sensitivity over midlatitude source regions and application to decadal trends: space-based indicator of O 3 sensitivity. J. Geophys. Res.-Atmos 122, 10,439–10,461. 10.1002/2017JD026720. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Jin X, Fiore A, Boersma KF, Smedt ID, Valin L, 2020. Inferring changes in summertime surface Ozone–NO x –VOC chemistry over U.S. urban areas from two decades of satellite and ground-based observations. Environ. Sci. Technol 54, 6518–6529. 10.1021/acs.est.9b07785. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Jung J, Choi Y, Wong DC, Nelson D, Lee S, 2021. Role of Sea fog over the yellow sea on air quality with the direct effect of aerosols. Geophys. Res. Atmos 126 10.1029/2020JD033498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kaiser J, Jacob DJ, Zhu L, Travis KR, Fisher JA, González Abad G, Zhang L, Zhang X, Fried A, Crounse JD, St. Clair JM, Wisthaler A, 2018. High-resolution inversion of OMI formaldehyde columns to quantify isoprene emission on ecosystem-relevant scales: application to the southeast US. Atmos. Chem. Phys 18, 5483–5497. 10.5194/acp-18-5483-2018. [DOI] [Google Scholar]
  34. Kang D, Pickering K, 2018. Lightning NOx emissions and the implications for surface air quality over the contiguous United States. EM (Pittsburgh Pa) 11, 1–6. [PMC free article] [PubMed] [Google Scholar]
  35. Kang D, Foley KM, Mathur R, Roselle SJ, Pickering KE, Allen DJ, 2019. Simulating lightning NO production in CMAQv5.2: performance evaluations. Geosci. Model Dev 12, 4409–4424. 10.5194/gmd-12-4409-2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Kang D, Mathur R, Pouliot GA, Gilliam RC, Wong DC, 2020. Significant ground-level ozone attributed to lightning-induced nitrogen oxides during summertime over the Mountain West States, npj Clim. Atmos. Sci 3, 6. 10.1038/s41612-020-0108-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Kitayama K, Morino Y, Yamaji K, Chatani S, 2019. Uncertainties in O3 concentrations simulated by CMAQ over Japan using four chemical mechanisms. Atmos. Environ 198, 448–462. 10.1016/j.atmosenv.2018.11.003. [DOI] [Google Scholar]
  38. Knowland KE, Ott LE, Duncan BN, Wargan K, 2017. Stratospheric Intrusion-Influenced ozone Air Quality Exceedances Investigated in the NASA MERRA-2 Reanalysis: SI-INFLUENCED O 3 EXCEEDANCES IN MERRA-2. Geophys. Res. Lett 44, 10,691–10,701. 10.1002/2017GL074532. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Koplitz S, Simon H, Henderson B, Liljegren J, Tonnesen G, Whitehill A, Wells B, 2021. Changes in ozone Chemical Sensitivity in the United States from 2007 to 2016. ACS Environ. Au 10.1021/acsenvironau.1c00029, 1c00029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Krotkov NA, Lamsal LN, Celarier EA, Swartz WH, Marchenko SV, Bucsela EJ, Chan KL, Wenig M, Zara M, 2017. The version 3 OMI NOs standard product. Atmos. Meas. Tech 10, 3133–3149. 10.5194/amt-10-3133-2017. [DOI] [Google Scholar]
  41. Lamsal LN, Martin RV, van Donkelaar A, Steinbacher M, Celarier EA, Bucsela E, Dunlea EJ, Pinto JP, 2008. Ground-level nitrogen dioxide concentrations inferred from the satellite-borne ozone monitoring Instrument. J. Geophys. Res 113, D16308. 10.1029/2007JD009235. [DOI] [Google Scholar]
  42. Lamsal LN, Krotkov NA, Celarier EA, Swartz WH, Pickering KE, Bucsela EJ, Gleason JF, Martin RV, Philip S, Irie H, Cede A, Herman J, Weinheimer A, Szykman JJ, Knepp TN, 2014. Evaluation of OMI operational standard NO2 column retrievals using in situ and surface-based NO2 observations. Atmos. Chem. Phys 14, 11587–11609. 10.5194/acp-14-11587-2014. [DOI] [Google Scholar]
  43. Lee H-J, Chang L-S, Jaffe DA, Bak J, Liu X, Abad GG, Jo H-Y, Jo Y-J, Lee J-B, Kim C-H, 2021. Ozone continues to increase in East Asia despite decreasing NO2: causes and abatements. Remote Sens. 13, 2177. 10.3390/rs13112177. [DOI] [Google Scholar]
  44. Levelt PF, van den Oord GHJ, Dobber MR, Malkki A, Visser Huib, de Vries Johan, Stammes P, Lundell JOV, Saari H, 2006. The ozone monitoring instrument. IEEE Trans. Geosci. Remote Sens 44, 1093–1101. 10.1109/TGRS.2006.872333. [DOI] [Google Scholar]
  45. Li J, Wang Y, 2019. Inferring the anthropogenic NOx emission trend over the United States during 2003–2017 from satellite observations: was there a flattening of the emission trend after the Great recession? Atmos. Chem. Phys 19, 15339–15352. 10.5194/acp-19-15339-2019. [DOI] [Google Scholar]
  46. Martin RV, Fiore AM, Van Donkelaar A, 2004. Space-based diagnosis of surface ozone sensitivity to anthropogenic emissions: surface ozone sensitivity to emissions. Geophys. Res. Lett 31, n/a–n/a. 10.1029/2004GL019416. [DOI] [Google Scholar]
  47. Mathur R, Xing J, Gilliam R, Sarwar G, Hogrefe C, Pleim J, Pouliot G, Roselle S, Spero TL, Wong DC, Young J, 2017. Extending the Community Multiscale Air Quality (CMAQ) modeling system to hemispheric scales: overview of process considerations and initial applications. Atmos. Chem. Phys 17, 12449–12474. 10.5194/acp-17-12449-2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Mehmood K, Wu Y, Wang L, Yu S, Li P, Chen X, Li Z, Zhang Y, Li M, Liu W, Wang Y, Liu Z, Zhu Y, Rosenfeld D, Seinfeld JH, 2020. Relative effects of open biomass burning and open crop straw burning on haze formation over central and eastern China: modeling study driven by constrained emissions. Atmos. Chem. Phys 20, 2419–2443. 10.5194/acp-20-2419-2020. [DOI] [Google Scholar]
  49. Millet DB, Jacob DJ, Boersma KF, Fu T-M, Kurosu TP, Chance K, Heald CL, Guenther A, 2008. Spatial distribution of isoprene emissions from North America derived from formaldehyde column measurements by the OMI satellite sensor. J. Geophys. Res 113, D02307. 10.1029/2007JD008950. [DOI] [Google Scholar]
  50. Napelenok S, Cohan D, Hu Y, Russell A, 2006. Decoupled direct 3D sensitivity analysis for particulate matter (DDM-3D/PM). Atmos. Environ 40, 6112–6121. 10.1016/j.atmosenv.2006.05.039. [DOI] [Google Scholar]
  51. Napelenok SL, Pinder RW, Gilliland AB, Martin RV, 2008. A method for evaluating spatially-resolved NOx emissions using Kalman filter inversion, direct sensitivities, and space-based NO2 observations. Atmos. Chem. Phys 12. [Google Scholar]
  52. Orville RE, 2008. Development of the National Lightning Detection Network. Bull. Am. Meteorol. Soc 89, 180–190. 10.1175/BAMS-89-2-180. [DOI] [Google Scholar]
  53. Palmer PI, Jacob DJ, Fiore AM, Martin RV, Chance K, Kurosu TP, 2003. Mapping isoprene emissions over North America using formaldehyde column observations from space. J. Geophys. Res.-Atmos 108 10.1029/2002JD002153, 2002JD002153. [DOI] [Google Scholar]
  54. Pleim JE, 2006. A simple, efficient solution of flux–profile relationships in the atmospheric surface layer. J. Appl. Meteorol. Climatol 45, 341–347. 10.1175/JAM2339.1. [DOI] [Google Scholar]
  55. Pleim JE, 2007a. A combined local and nonlocal closure model for the atmospheric boundary layer. Part II: application and evaluation in a mesoscale meteorological model. J. Appl. Meteorol. Climatol 46, 1396–1409. 10.1175/JAM2534.1. [DOI] [Google Scholar]
  56. Pleim JE, 2007b. A combined local and nonlocal closure model for the atmospheric boundary layer. Part I: model description and testing. J. Appl. Meteorol. Climatol 46, 1383–1395. 10.1175/JAM2539.1. [DOI] [Google Scholar]
  57. Pleim JE, Gilliam R, 2009. An indirect data assimilation scheme for deep soil temperature in the Pleim–Xiu land surface model. J. Appl. Meteorol. Climatol 48, 1362–1376. 10.1175/2009JAMC2053.1. [DOI] [Google Scholar]
  58. Pleim JE, Xiu A, 1995. Development and testing of a surface flux and planetary boundary layer model for application in mesoscale models. J. Appl. Meteorol 34, 16–32. [Google Scholar]
  59. Pleim JE, Xiu A, 2003. Development of a land surface model. Part II: Data assimilation. J. Appl. Meteorol 42,1811–1822. . [DOI] [Google Scholar]
  60. Qu Z, Jacob DJ, Silvern RF, Shah V, Campbell PC, Valin LC, Murray LT, 2021. US COVID-19 shutdown demonstrates importance of background NO 2 in inferring NO x emissions from satellite NO 2 observations. Geophys. Res. Lett 10.1029/2021GL092783. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Rodgers CD, 2000. Inverse methods for atmospheric sounding: theory and practice, series on atmospheric, oceanic and planetary physics. World Scientific. 10.1142/3171. [DOI] [Google Scholar]
  62. Schaub D, Boersma KF, Kaiser JW, Weiss AK, Folini D, Eskes HJ, Buchmann B, 2006. Comparison of GOME tropospheric NO2 columns with NO2 profiles deduced from ground-based in situ measurements. Atmos. Chem. Phys 6, 3211–3229. 10.5194/acp-6-3211-2006. [DOI] [Google Scholar]
  63. Schroeder JR, Crawford JH, Fried A, Walega J, Weinheimer A, Wisthaler A, Müller M, Mikoviny T, Chen G, Shook M, Blake DR, Tonnesen GS, 2017. New insights into the column CH 2 O/NO 2 ratio as an indicator of near-surface ozone sensitivity: CH 2 O/NO 2 as Indicator of O 3 Sensitivity. J. Geophys. Res.-Atmos 122, 8885–8907. 10.1002/2017JD026781. [DOI] [Google Scholar]
  64. Shen L, Jacob DJ, Zhu L, Zhang Q, Zheng B, Sulprizio MP, Li K, De Smedt I, González Abad G, Cao H, Fu T, Liao H, 2019. The 2005–2016 trends of formaldehyde columns over China observed by satellites: increasing anthropogenic emissions of volatile organic compounds and decreasing agricultural fire emissions. Geophys. Res. Lett 46, 4468–4475. 10.1029/2019GL082172. [DOI] [Google Scholar]
  65. Sillman S, 1995. The use of NOy, H2O2, and HNO3 as indicators for ozone-NOx-hydrocarbon sensitivity in urban locations. J. Geophys. Res 100, 14175. 10.1029/94JD02953. [DOI] [Google Scholar]
  66. Silvern RF, Jacob DJ, Mickley LJ, Sulprizio MP, Travis KR, Marais EA, Cohen RC, Laughner JL, Choi S, Joiner J, Lamsal LN, 2019. Using satellite observations of tropospheric NO 2 columns to infer long-term trends in US NO x emissions: the importance of accounting for the free tropospheric NO 2 background. Atmos. Chem. Phys 19, 8863–8878. 10.5194/acp-19-8863-2019. [DOI] [Google Scholar]
  67. Souri AH, Choi Y, Jeon W, Li X, Pan S, Diao L, Westenbarger DA, 2016. Constraining NOx emissions using satellite NO2 measurements during 2013 DISCOVER-AQ Texas campaign. Atmos. Environ 131, 371–381. 10.1016/j.atmosenv.2016.02.020. [DOI] [Google Scholar]
  68. Souri AH, Choi Y, Jeon W, Kochanski AK, Diao L, Mandel J, Bhave PV, Pan S, 2017. Quantifying the impact of biomass burning emissions on major inorganic aerosols and their precursors in the U.S.: burning Impact on Inorganic Aerosols. J. Geophys. Res.-Atmos 122, 12,020–12,041. 10.1002/2017JD026788. [DOI] [Google Scholar]
  69. Souri AH, Nowlan CR, González Abad G, Zhu L, Blake DR, Fried A, Weinheimer AJ, Wisthaler A, Woo J-H, Zhang Q, Chan Miller CE, Liu X, Chance K, 2020. An inversion of NOx and non-methane volatile organic compound (NMVOC) emissions using satellite observations during the KORUS-AQ campaign and implications for surface ozone over East Asia. Atmos. Chem. Phys 20, 9837–9854. 10.5194/acp-20-9837-2020. [DOI] [Google Scholar]
  70. Tang W, Cohan DS, Lamsal LN, Xiao X, Zhou W, 2013. Inverse modeling of Texas NOx emissions using space-based and ground-based NO2 observations. Atmos. Chem. Phys 13, 11005–11018. 10.5194/acp-13-11005-2013. [DOI] [Google Scholar]
  71. Tarasick DW, Smit HGJ, Thompson AM, Morris GA, Witte JC, Davies J, Nakano T, Van Malderen R, Stauffer RM, Johnson BJ, Stübi R, Oltmans SJ, Vömel H, 2021. Improving ECC ozonesonde data quality: assessment of current methods and outstanding issues. Earth Space Sci 8. 10.1029/2019EA000914. [DOI] [Google Scholar]
  72. Tian D, Cohan DS, Napelenok S, Bergin M, Hu Y, Chang M, Russell AG, 2010. Uncertainty analysis of ozone formation and response to emission controls using higher-order sensitivities. J. Air Waste Manage. Assoc 60, 797–804. 10.3155/1047-3289.60.7.797. [DOI] [PubMed] [Google Scholar]
  73. Tonnesen GS, Dennis RL, 2000. Analysis of radical propagation efficiency to assess ozone sensitivity to hydrocarbons and NOx: 2. Long-lived species as indicators of ozone concentration sensitivity. J. Geophys. Res.-Atmos 105, 9227–9241. 10.1029/1999JD900372. [DOI] [Google Scholar]
  74. Torres O, Bhartia PK, Jethva H, Ahn C, 2018. Impact of the ozone monitoring instrument row anomaly on the long-term record of aerosol products. Atmos. Meas. Tech 11, 2701–2715. 10.5194/amt-11-2701-2018. [DOI] [Google Scholar]
  75. Travis KR, Jacob DJ, Fisher JA, Kim PS, Marais EA, Zhu L, Yu K, Miller CC, Yantosca RM, Sulprizio MP, Thompson AM, Wennberg PO, Crounse JD, St. Clair JM, Cohen RC, Laughner JL, Dibb JE, Hall SR, Ullmann K, Wolfe GM, Pollack IB, Peischl J, Neuman JA, Zhou, 2016. Why do models overestimate surface ozone in the Southeast United States? Atmos. Chem. Phys 16, 13561–13577. 10.5194/acp-16-13561-2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Vukovich JM, Eyth A, Henderson B, Allen C, Beidler J, 2018. Development of 2016 Hemispheric Emissions for CMAQ. [Google Scholar]
  77. Wang Y, Beirle S, Lampel J, Koukouli M, De Smedt I, Theys N, Li A, Wu D, Xie P, Liu C, Van Roozendael M, Stavrakou T, Móller J-F, Wagner T, 2017. Validation of OMI, GOME-2A and GOME-2B tropospheric NO2, SO2 and HCHO products using MAX-DOAS observations from 2011 to 2014 in Wuxi, China: investigation of the effects of priori profiles and aerosols on the satellite products. Atmos. Chem. Phys 17, 5007–5033. 10.5194/acp-17-5007-2017. [DOI] [Google Scholar]
  78. Wiedinmyer C, Quayle B, Geron C, Belote A, McKenzie D, Zhang X, O’Neill S, Wynne KK, 2006. Estimating emissions from fires in North America for air quality modeling. Atmos. Environ 40, 3419–3432. https://doi.org/10.1016/j.atmosenv.2006.02.010. [Google Scholar]
  79. Wiedinmyer C, Akagi SK, Yokelson RJ, Emmons LK, Al-Saadi JA, Orlando JJ, Soja AJ, 2011. The Fire INventory from NCAR (FINN): a high resolution global model to estimate the emissions from open burning. Geosci. Model Dev 4, 625–641. 10.5194/gmd-4-625-2011. [DOI] [Google Scholar]
  80. Wiedinmyer C, Yokelson RJ, Gullett BK, 2014. Global Emissions of trace gases, particulate matter, and hazardous air pollutants from open burning of domestic waste. Environ. Sci. Technol 48, 9523–9530. 10.1021/es502250z. [DOI] [PubMed] [Google Scholar]
  81. Xiu A, Pleim JE, 2001. Development of a land surface model. Part I: application in a mesoscale meteorological model. J. Appl. Meteorol 40, 192–209. . [DOI] [Google Scholar]
  82. Zhu L, Jacob DJ, Mickley LJ, Marais EA, Cohan DS, Yoshida Y, Duncan BN, González Abad G, Chance KV, 2014. Anthropogenic emissions of highly reactive volatile organic compounds in eastern Texas inferred from oversampling of satellite (OMI) measurements of HCHO columns. Environ. Res. Lett 9, 114004 10.1088/1748-9326/9/11/114004. [DOI] [Google Scholar]
  83. Zhu L, Jacob DJ, Kim PS, Fisher JA, Yu K, Travis KR, Mickley LJ, Yantosca RM, Sulprizio MP, De Smedt I, González Abad G, Chance K, Li C, Ferrare R, Fried A, Hair JW, Hanisco TF, Richter D, Jo Scarino A, Walega J, Weibring P, Wolfe GM, 2016. Observing atmospheric formaldehyde (HCHO) from space: validation and intercomparison of six retrievals from four satellites (OMI, GOME2A, GOME2B,OMPS) with SEAC4RS aircraft observations over the southeast US. Atmos. Chem. Phys 16, 13477–13490. 10.5194/acp-16-13477-2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  84. Zhu L, Mickley LJ, Jacob DJ, Marais EA, Sheng J, Hu L, Abad GG, Chance K, 2017. Long-term (2005-2014) trends in formaldehyde (HCHO) columns across North America as seen by the OMI satellite instrument: evidence of changing emissions of volatile organic compounds: HCHO Trend across North America. Geophys. Res. Lett 44, 7079–7086. 10.1002/2017GL073859. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement1
NIHMS1789194-supplement-Supplement1.docx (4.7MB, docx)

Data Availability Statement

The OMI/Aura Level 2 tropospheric NO2 and total HCHO column data products are available from the NASA Goddard Earth Science (GES) Data and Information Services Center (DISC) (https://disc.gsfc.nasa.gov/datasets/OMNO2_003/summary?keywords=no2, https://disc.gsfc.nasa.gov/datasets/OMHCHO_003/summary?keywords=hcho). The NCEP NARR data can be downloaded from https://rda.ucar.edU/datasets/ds608.0/. The FINN data can be acquired from https://www.acom.ucar.edu/Data/fire. The lightning flash data are available for purchase from Vaisala Inc. (https://www.vaisala.com/en/products/systems/lightning-detection). For this study, the native form of the lightning flash data was provided by Dr. Hyuncheol Kim at the NOAA, and processed into the required format of the CMAQ model with the considerable assistance of Dr. Daiwen Kang at the EPA. The EPA AQS data (pre-generated) for meteorology and air quality measurement over the U.S. is available from https://aqs.epa.gov/aqsweb/airdata/download_files.html. Aircraft measurement during the DISCOVER-AQ 2011 (Baltimore-Washington, D.C.) and 2014 (Colorado) campaigns were distributed by NASA (https://www-air.larc.nasa.gov/missions/discover-aq/discover-aq.html). Ozonesonde data is downloaded from the NOAA Global Monitoring Laboratory/Earth System Research Laboratories (https://gml.noaa.gov/dv/data/index.php?category=Ozone&type=Balloon). The model outputs from the primary simulations can be downloaded from doi: https://doi.org/10.5281/zenodo.5590569.

RESOURCES