Skip to main content
NCBI home page
NASA Author Manuscripts logoLink to NASA Author Manuscripts
. Author manuscript; available in PMC: 2020 Dec 16.
Published in final edited form as: J Geophys Res Atmos. 2019 Nov 12;124(23):13560–13575. doi: 10.1029/2019jd030765

Comparison of Near-surface NO2 Pollution with Pandora Total Column NO2 during the Korea-United States Ocean Color (KORUS OC) Campaign

Anne M Thompson 1,2, Ryan M Stauffer 1,3, Tyler P Boyle 2, Debra E Kollonige 1,4, Kazuyuki Miyazaki 5, Maria Tzortziou 6, Jay R Herman 1,7, Nader Abuhassan 1,7, Carolyn E Jordan 8, Brian T Lamb 6
PMCID: PMC7477803  NIHMSID: NIHMS1558473  PMID: 32913698

Abstract

Near-surface air quality (AQ) observations over coastal waters are scarce, a situation that limits our capacity to monitor pollution events at land-water interfaces. Satellite measurements of total column (TC) nitrogen dioxide (NO2) observations are a useful proxy for combustion sources but the once daily snapshots available from most sensors are insufficient for tracking the diurnal evolution and transport of pollution. Ground-based remote sensors like the Pandora Spectrometer Instrument (PSI) that have been developed to verify space-based total column NO2 and other trace gases are being tested for routine use as certified AQ monitors. The KORUS-OC (Korea-United States Ocean Color) cruise aboard the R/V Onnuri in May-June 2016 represented an opportunity to study AQ near the South Korean coast, a region affected by both local/regional and long-distance pollution sources. Using PSI data in direct-sun mode and in situ sensors for shipboard ozone, CO and NO2, we explore, for the first time, relationships between TC NO2 and surface AQ in this coastal region. Three case studies illustrate the value of the PSI as well as complexities in the surface-column NO2 relationship caused by varying meteorological conditions. Case Study 1 (25-26 May 2016) exhibited a high correlation of surface NO2 to TC NO2 measured by both PSI and Aura’s Ozone Monitoring Instrument (OMI) but two other cases displayed poor relationships between in situ and TC NO2 due to decoupling of pollution layers from the surface. With suitable interpretation the PSI TC NO2 measurement demonstrates good potential for working with upcoming geostationary satellites to advance diurnal tracking of pollution.

Keywords: Pandora spectrometer, KORUS (2016), Korean air pollution, OMI NO2, 0345, 0365, 1610, 1640

1. Introduction

Total Column NO2 (TC NO2) measurements have been used extensively to characterize global pollution. With Aura’s OMI (Ozone Monitoring Instruments) satellite in operation for nearly fifteen years changes in TC NO2 have been tracked since late 2004 [Duncan et al., 2013; Krotkov et al., 2016; Levelt et al., 2018; Liu et al., 2017; Tan et al., 2018; Goldberg et al., 2019]. For example, Duncan et al. [2016] reported changes over 195 large cities worldwide. Every continent surveyed included cities with decreasing NO2, principally due to reduced NOx emissions from mobile sources. In many areas where economic growth rates have been high, roughly half to three-quarters of the cities recorded positive changes, typically from 10-30% from 2005-2014 for tropospheric column NO2. Total column NO2 over several middle Eastern, south and east Asian cities grew more than 40%. Even where megacities showed double-digit declines, large regions outside the “urban core” displayed positive changes (example for China in Figure 9, Duncan et al., 2016). Over the megacity of Seoul, South Korea there was a decrease (15% lower TC NO2 from 2005 to 2014) whereas the economically expanding Daesan-Pyeongtaek-Incheon area southwest of Seoul displayed an uptick of 20% or more (Figures 2 and 3 in Duncan et al., 2016). More rural areas in the Seoul region also showed increases in TC NO2.

Figure 9.

Figure 9.

Open in a new tab

Surface (red) and PSI TC (blue) NO2 amounts on 30 May from the Onnuri. PSI TC NO2 values close to the cruise average (dotted line) were measured during the DC-8 overpass. Dashed black line is the cruise mean in situ NO2. The time of the DC-8 overpass of the Onnuri is noted on the figure.

Figure 2.

Figure 2.

Open in a new tab

Onnuri surface NO2 (a), O3 (b), CO (c), and CL51 BLView aerosol mixed layer heights (d) for each day during the KORUS-OC cruise (5-minute averaged data). The three case study days are highlighted in red (CS1), black (CS2), and blue (CS3) dots, and non-case study days are shown in gray.

Figure 3.

Figure 3.

Open in a new tab

(a) In situ and TC NO2 readings from the RV Onnuri during KORUS-OC. Case study periods indicated. (b) Coincident surface and PSI TC NO2 measurements and linear least squares best fit lines for Case Study1 (red), 2 (black), and 3 (blue). The remainder of the coincident NO2 measurements are shown with open circles.

Since the Duncan et al. (2016) study with OMI, the Sentinel 5 Precursor (S5P) TROPOMI has become operational. With increased resolution over OMI (~5–6 times greater at nadir) S5P makes it easier to follow local and wide-area changes in TC NO2 as well as other pollutants, e.g. HCHO and SO2 [Veefkind et al., 2012; van Geffen et al., 2018]. The east Asian Geostationary Environment Monitoring Spectrometer (GEMS) instrument as part of KOMPSAT will be launched by a South Korean consortium (KARI (Korea Aerospace Research Institute), KMA (Korea Meteorological Administration) and NMSC (National Meteorological Satellite Center)) within a year or so, providing TC O3, NO2, SO2 and HCHO approximately hourly during daylight hours.

Accurate attribution of TC NO2 (or O3, NO2 or HCHO) to the column in the mixed or boundary layer (BL) depends on separation of stratospheric and tropospheric column amounts. Even when the stratospheric column has been removed, the relationship of satellite tropospheric column NO2 to surface NO2, which is an irritating pollutant and often the dominant source of surface ozone, can be complex [Knepp et al., 2015; Kollonige et al. 2018; Reed et al., 2015]. Thus, there is a need for independent verification of the satellite column amount and parallel measurement of surface NO2. Among a number of ground-based spectrometric approaches to TC NO2 [Piters et al., 2012; Kreher et al., 2017] is the Pandora Spectrometer Instrument (PSI). Recently Herman et al. [2019] surveyed OMI TC NO2 patterns over six multi-year PSI stations in Korea and the US. The typical mid-year elevated TC NO2 seen by OMI was confirmed by the ground-based instrument. However, over the sites that are often polluted, e.g., Seoul, Busan, Washington, DC, mean and daily PSI measurements of TC NO2 exceeded the satellite readings by 50% or more suggesting that satellite measurements of TC NO2 are not sufficient for estimating air quality.

A suite of PSIs were deployed throughout South Korea in May-June as part of a comprehensive ground-air-satellite joint Korea-US Air Quality (KORUS-AQ) field campaign [Al-Saadi et al., 2015]. A major goal of KORUS-AQ was to compare measurements of column NO2 among various instruments in anticipation of GEMS. In addition to OMI, total column NO2 was measured from nine PSI and partial column NO2 was provided by the airborne GEO-TASO spectrometer and from spirals of two other aircraft carrying NO2 instruments, the NASA-DC-8 and Korea’s NIER (National Institute of Environmental Research) Hanseo King Air. The aircraft carried out maneuvers over a number of the PSI sites (see list of PSI locations in Herman et al., 2018). Depending on the site, background TC NO2 averaged 0.1-0.3 DU (Dobson Units; 1 DU = 2.69×1016/cm2). and pollution is defined as a minimum of 1.5 times the background (Figure 3 in Herman et al., 2018). There were NO2 pollution episodes throughout South Korea during KORUS-AQ, although the frequency and magnitude varied with the station [Herman et al., 2018; 2019]. Remote sites averaged 3-4 times lower TC NO2 than the two most polluted stations in Seoul, Olympic Park and Yonsei University [Tzortziou et al., 2018]. At Olympic Park mean daytime TC NO2 from the PSI was ~1.5 DU

At the same time as KORUS-AQ a coastal cruise sponsored by the Korean Institute of Ocean Science and Technology (KIOST) and NASA deployed a smaller set of in situ instruments and a shipboard direct-sun PSI [Tzortziou et al., 2018] on the Research Vessel (R/V) Onnuri. Designated as part of KORUS-OC [Park et al., 2015], the cruise originated at the KIOST South Sea Research Institute dock on the island of Geoje, near the large city of Busan, on 20 May 2016 on the southern coast of the Korean Peninsula and it first sailed to the northeast in the East Sea. The Onnuri returned south after 27 May, then sailed west and to the north in the Yellow Sea where it was west of Pyeongtaek (30 km southwest of Seoul) before returning south and back to Geoje on 6 June 2016. Due to clouds there were no PSI data from 3-6 June. The cruise track appears in Figure 1; gaps denote data within 12 nm of shore that KIOST did not deposit in the KORUS-OC archive. Tzortziou et al. [2018] carried out a detailed comparison of PSI and OMI total column ozone and NO2 measurements over the KORUS-AQ period and the Onnuri cruise, although a temporary malfunction of OMI operation precluded comparisons after 29 May 2016. One of the main objectives of that study was to assess the impact of spatiotemporal variability in atmospheric trace gases on satellite retrievals of ocean biogeochemical processes. Column O3 was found to average 320 DU. Agreement between land-based PSI and OMI column ozone varied from site to site, ranging from 0.4-2% [Tzortziou et al., 2018], with OMI usually higher than the Pandora. For TC NO2, agreement varied considerably with the land-based PSI instruments recording more TC NO2 by 10-50% than OMI. Comparisons of OMI and a suite of PSIs during a prior northern mid-latitude campaign (DISCOVER-AQ in the Baltimore-Washington region, July 2011; Flynn et al., 2015; Knepp et al., 2015; Reed et al., 2015; Tzortziou et al., 2015) showed the ground-based instrument usually registering more TC NO2 than OMI because of averaging over the larger field of view of OMI.

Figure 1.

Figure 1.

Open in a new tab

R/V Onnuri ship track during KORUS-OC excluding locations within territorial waters. Individual case study locations are shown with arrows. Case Study 1 took place as the ship sailed toward the city of Pohang on the East Sea. Case Study 2 is based on data taken as the ship sailed on the Yellow Sea. Case Study 3 occurred as the Onnuri sailed toward Seoul. Data from missing segments, corresponding to South Korean coastal waters, were not available for analysis.

Here, for the first time, we compare total column NO2 observations from OMI and PSI with detailed Onnuri in-situ and ancillary observations to give a more comprehensive view of coastal air quality during the KORUS-OC cruise. That is the goal of this investigation. Namely, we focus on a survey of PSI TC NO2 comparisons with surface NO2, ozone and CO as measured on the Onnuri. After giving a description of the measurements and analytical tools employed (Section 2), we give an overview, then examine the relationship between the PSI TC NO2 and shipboard NO2 to see how well they track one another under various meteorological conditions. This approach will focus on case studies: (1) 25-26 May 2016, when the Onnuri sampled pollution near the surface; (2) 30 May where an encounter with the DC-8 provided profiles of trace gases for comparison with the ship; (3) 1 June where the Onnuri sampled large variations in TC NO2 not far from Seoul. Section 3 presents these case studies in detail to determine under which conditions remote sensors can be expected to provide meaningful information about poor AQ episodes. We also compare our findings to a prior coastal cruise in which a shipboard PSI was compared to surface NO2 and to OMI TC NO2. That expedition, the Deposition of Active Nitrogen to Coastal Ecosystem (DANCE) experiment was conducted in July-August 2014 in the western Atlantic off the Virginia (US) and North Carolina coast [Kollonige et al., 2018; Martins et al., 2016]. Section 4 is a summary.

2. Data and Methods

2.1. Total Column Measurements

2.1.1. Pandora Spectrometer Instrument (PSI)

Our research focuses on ground- and ship-based PSI [Herman et al., 2009] measurements during KORUS-OC and KORUS-AQ [Al-Saadi et al., 2015]. PSI measures direct solar irradiances to calculate TC NO2 and O3 (Table 1) [Herman et al., 2009; Tzortziou et al 2013; Reed et al., 2015]. There are two versions of this instrument; the Sun-only complementary metal-oxide-semiconductor (CMOS) detector PSI and the charge-coupled device (CCD) detector Sun-and-Sky PSI. Both models use similar wavelengths to retrieve TC measurements every 2 minutes [Tzortziou et al., 2013; Reed et al., 2015]. With clear skies, the spectrometer can return TC NO2 values with an absolute error of ± 0.05 Dobson Units (DU) and a precision of ± 0.01 DU [Herman et al., 2009]. PSI has a field of view (FOV) of 1.6° and operates at a spectral resolution of 0.6 nm. A more detailed technical description of the PSI spectrometer is provided in Herman et al. [2009]. Updates are described in Tzortziou et al. [2018] for the ship-board PSI sensor and Herman et al. [2018]. Herman et al. [2018] compared land-based NO2 and formaldehyde (HCHO) PSI observations with satellite TC NO2 observations from the Ozone Monitoring Instrument (OMI) and aircraft KORUS-AQ TC NO2 observations from the Spectrometer for Sun-tracking Sky-scanning Atmospheric Research (4STAR) on the DC-8. They found typical mean differences between PSI and OMI observations (PSI larger) at Busan were 0.35 DU, with 0.58 DU mean differences at Seoul. One significant cause of OMI underestimation at its overpass time compared to PSI is the large OMI satellite field-of-view (FOV) that includes regions containing low values of pollutants [Tzortziou et al. 2013; Herman et al. 2018]. Co-located observations from 4STAR had relatively good correlation with PSI observations.

Table 1.

Instrumentation used on the RV Onnuri during KORUS-OC.

Data Instrument Case Studies
*TC NO2 PSI Spectrometer 1, 2, and 3
*In situ NO2 Aerodyne CAPS 1, 2, and 3
*In situ O3 Thermo Model 49C 1, 2, and 3
*In situ CO Thermo 48C 1 and 3
*Aerosol backscatter Vaisala CL51 Ceilometer 1 and 3
TC NO2 OMI 1
Open in a new tab
*

Ship-based measurements

We use filtered TC NO2 measurements (those with Norm RMS < 0.01 and NO2 error < 0.05 DU) in each case study. As recommended by the data providers [Tzortziou et al., 2018], measurements flagged with a potential retrieval error (NO2 error > 0.05 DU) were also removed, whereas Herman et al. [2018] omitted observations with NO2 error > 0.1 DU. Measurements taken within 12 nm of the South Korean coastline, inside territorial seas, were also removed due to a data publication agreement with KIOST. We carefully examined the surface NO2, O3, and Pandora TC NO2 data for spikes that were obviously caused by sampling of the Onnuri exhaust. Pandora was located forward of the exhaust stack, so plumes rarely affected the Pandora TC NO2 measurements while the ship was moving. Ship traffic should not have affected the Pandora TC NO2 measurements analyzed here because encounters with ship traffic outside of the 12 nm excluded coastal zone were rare. After filtering, Onnuri PSI TC NO2 observations were resampled to 5-min averages for comparison to the in situ measurements.

2.1.2. OMI

Total column NO2 observations are available once-daily from OMI on the NASA Aura satellite. Overpasses occur between 1300-1400 KST. OMI operates with a 13 km × 24 km nadir spatial resolution, observing direct and back scattered solar radiation between 264-504 nm to determine TC NO2 values (Levelt et al., 2006; Boersma et al., 2007; Duncan et al., 2013). Before entering temporary safe mode on 29 May, OMI provided both TC O3 and TC NO2 observations for the campaigns. Despite frequent cloud cover and the period of safe mode operations, OMI overpassed the Onnuri most days during the early portion of the cruise [Tzortziou et al., 2018]. We use one overpass during the operational period to determine the capability of the PSI to measure TC NO2-rich air masses offshore. OMI TC NO2 data are from the NASA L2 standard product available at https://aura.gesdisc.eosdis.nasa.gov/data/Aura_OMI_Level2/OMNO2.003/.

We use TC NO2 data to track pollution. For this approach to be valid, the stratospheric NO2 must be small compared to the values and changes in the tropospheric column. For the period for which OMI data were available, using the Taehwa site overpass data as a reference, the stratospheric column was ~4×1015 cm−2, ~0.14 DU. During Case Study 1 (Section 3.2.1 below) the PSI TC NO2 was well above this value and increased nearly 50% by the time that surface NO2 changed. In Case Study 3 the tropospheric NO2 increases by more than a factor of 5. Lamsal et al. [2014] describe several sources of uncertainty in the tropospheric column; these would be 20-25% for the data used here.

2.2. Surface-based Observations

2.2.1. In Situ Analyzers

Surface trace gas measurements on the Onnuri included NO2 (Aerodyne CAPS; Table 1), O3 (Thermo 49C), and CO (Thermo 48C). Further details on instrumentation are found in Martins et al. [2012; 2016]. We averaged all observations to 5-minutes to match the 5-minute PSI averages. Uncertainties for these instruments specified in Martins et al. [2012] and Kollonige et al. [2018] are 5% (NO2), 1.3% (O3) and 10 ppbv (CO) respectively. Observations influenced by ship exhaust (identified from coincident upward spikes in surface NO2 and downward spikes in surface O3) were omitted from the 5-minute averages and comparisons.

2.2.2. CL51 Ceilometer

Ancillary observations from a Vaisala CL51 ceilometer on the Onnuri (Table 1) were used to determine atmospheric stability during KORUS-OC. The ceilometer uses diode-laser technology to detect 910 nm backscatter between 0.05-15 km above ground level. In this work, we utilize aerosol-layer (or “mixed layer”) heights determined through post-processing software (BL-VIEW; Emeis et al., 2007; Lotteraner and Piringer, 2016; Knepp et al., 2017) to supplement analysis of three case studies.

2.3. Additional Analyses

2.3.1. Chemical Reanalysis

Chemical reanalyses provide another perspective of atmospheric composition over South Korea during KORUS-OC. We use model output from Miyazaki et al. [2019], who used a chemistry transport model combined with assimilated satellite and meteorological data to construct a reanalysis of global chemical conditions during the KORUS-AQ/KORUS-OC time periods. A strength of this reanalysis is its assimilation of multiple chemical constituents from a number of satellites (O3: MLS, OMI, and AIRS; NO2: OMI and GOME-2; CO: MOPITT; HNO3: MLS; SO2: OMI) and simultaneous optimizations of ozone precursor emissions and concentrations. For example, the chemical reanalysis showed reduced biases in ozone relative to models without data assimilation when comparisons were made to aircraft measurements and ozonesondes from KORUS-AQ [Miyazaki et al., 2019]. The spatial and vertical variations of O3 and its precursors obtained from the DC-8 measurements were mostly reproduced by the reanalysis. The chemical reanalysis has also shown reasonable agreements with ship measurements of O3 and CO over open oceans from 67°S to 75°N [Kanaya et al., 2019]. The two-hourly output is available on a 1.1 × 1.1° horizontal resolution with 32 levels from the surface to 4.4 hPa. Although the spatial resolution of the reanalysis is not quite sufficient to resolve detailed structures in the observations, it is useful when examining near-surface (i.e. the lowest model level) NO2, O3, and CO on 26 May, 30 May, and 1 June (the three case studies) to understand potential emissions sources and atmospheric transport of aged air.

2.3.2. HYSPLIT

Backward and forward trajectories from the Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT; Draxler and Hess, 1997) model were initialized from KORUS-OC and KORUS-AQ sites to visualize and understand air mass transport during each case study. Trajectories were run at https://www.ready.noaa.gov/HYSPLIT_traj.php. For 25–26 May we analyze backward trajectories initialized at 500 m, 1000 m, and 1500 m. We use trajectories initialized at the altitude of temperature inversions identified at 400 m and 1200 m for 30 May. Forward trajectories were run for 1 June. Further information on HYSPLIT can be found in Stein et al. [2015].

2.3.3. DC-8

We use 10-second averages of NO2, O3, CO, and temperature from the NASA Airborne Science Laboratory (DC-8) to supplement 30 May Onnuri observations. We use NO2 and O3 from the 4-channel chemiluminescence instrument (https://www2.acom.ucar.edu/engineering/4-channel-nono2noyozone-chemiluminescence-instrument). Carbon monoxide observations are from the Differential Absorption Carbon Monoxide Measurement (DACOM; Sachse et al., 1987) instrument (DC-8 data available at: https://www-air.larc.nasa.gov/cgi-bin/ArcView/ korusaq). Uncertainties for NO2, O3, and CO are 50 pptv ± 30%, 5 ppbv ± 10%, and 2 ppbv, respectively.

3. Results

3.1. KORUS-OC Campaign Overview

Figure 2 summarizes the variability in surface NO2, O3 and CO (Figures 2a–c) as well as the MLH computed from the ceilometer BLView software. Case Study (CS) 1 (in red) displays the highest pollution which occurred on 26 May. The most important O3 impacts in all three cases occurred in mid-late afternoon when the MLH (Figure 2d) was most elevated.

Unlike the relatively clean air masses with low surface O3 and TC NO2 observed off the Eastern US during DANCE [Kollonige et al., 2018; Martins et al. 2016], the pollution-rich environments of the Yellow and East Seas provided several opportunities to investigate the sensitivity of a PSI to poor AQ events (Figure 3a). For example, the Korean air quality standard for maximum 8-hour average surface O3 is 60 ppbv (1-hour standard = 100 ppbv), which was exceeded on 15 consecutive days (19 May-2 June) on the Onnuri during KORUS-OC.

Generally, lower in situ and TC NO2 amounts were recorded on the Onnuri during KORUS-OC than land-based sites during KORUS-AQ [Herman et al., 2018; Tzortziou et al., 2018]. We observed average PSI TC NO2 values near 0.20 DU, with the highest values occurring as the ship sailed on the Yellow Sea near the coast off Seoul on 1 June [Tzortziou et al., 2018]. Herman et al. [2018] reported TC NO2 of 0.3–0.5 DU near smaller Korean cities and frequent observations above 2 DU near Seoul. Both land-based and Onnuri PSI TC NO2 measurements were larger than the 0.15 DU mean TC NO2 from the DANCE cruise [Martins et al., 2016]. KORUS-OC in situ NO2 measurements averaged ≈ 2 parts per billion by volume (ppbv), with the highest NO2 values on 24 and 26 May as the ship sailed on the East Sea (Figure 1). These concentrations were smaller than the 3.3 ppbv median mixing ratios observed during DANCE [Martins et al., 2016]. Given the larger TC NO2 observed in coastal Korean waters (KORUS-OC mean TC NO2 = 0.21 DU), this suggests that NO2 may be concentrated at a higher altitude over the South Korean coast relative to the mid-Atlantic. The lowest surface NO2 was measured during the later portion of the cruise when the Onnuri was farther offshore.

3.2. Case Studies

3.2.1. Case Study 1 (25-26 May)

We examine a 2-day pollution episode that encompasses both offshore and coastal pollution as seen in the PSI TC and surface NO2 time-series (Figure 3a, with details displayed in Figure 4a). Comprehensive OMI TC NO2 observations over the Korean peninsula and a close overpass of the Onnuri in the East Sea are available for 25 May but not 26 May. As denoted by TC NO2 measured over the city of Busan where a PSI was deployed (Figure 4 in Herman et al., 2018) and surface ozone, NO2 and CO (by NIER, not shown) are routinely measured, pollution increases from 25 May to 26 May were observed over the east coast of South Korea as well as on the Onnuri.

Figure 4.

Figure 4.

Open in a new tab

(a) Surface and PSI TC NO2 time-series from 22-26 May. (b) OMI TC NO2 from the 25 May overpass. Locations of the Onnuri during OMI overpass times on 25 and 26 May are noted on (b). Note poor air quality regions are near Seoul, Busan and the region sampled by the Onnuri on 25 May.

3.2.1.1. Case 1, Part 1 (25 May)

The second largest daily-averaged PSI TC NO2 amounts during the KORUS-OC cruise were recorded on 25 May as the Onnuri sailed on the East Sea (Figure 1). Despite intermittent cloud cover, the PSI detected increasing TC NO2 throughout the afternoon (Figure 4a). During an OMI overpass of the Onnuri between 1330-1400 KST, OMI and PSI measured TC NO2 ~0.24 DU (5 × 1015 molecules/cm2 in Figure 4a), smaller than the peak of 0.35 DU TC NO2 observed by the PSI at 1800 KST (Figure 4a). In contrast to the PSI TC NO2 levels exceeding the cruise average throughout the afternoon of 25 May, in situ NO2 was elevated only in the morning (several readings greater than 20 ppbv NO2 in Figure 4a).

Data from the CL51 ceilometer indicated a growing mixed layer height on 25 May with an altitude of 0.5 km during the overpass (Figure 5). Backscattered aerosols detected by the ceilometer throughout the day suggested the presence of a deep, well-mixed layer near the surface with enhanced aerosol backscatter up to 2 km altitude. The divergence of PSI/OMI TC NO2 and of in situ NO2 levels during a period of multiple mixed layers suggests the presence of a complex NO2 profile during the 25 May OMI overpass.

Figure 5.

Figure 5.

Open in a new tab

CL51 ceilometer backscatter profiles from CS1 on 25 and 26 May. The black dots indicate the BLView-calculated aerosol mixed layer heights.

Regional OMI observations provide context for pollution seen offshore on the Onnuri. On 25 May, OMI recorded TC NO2 observations greater than 0.5 DU at Busan and Seoul (Figure 4b). HYSPLIT back trajectories initialized during the overpass as well as patterns of enhanced OMI TC NO2 offshore indicate west-to-east flow across South Korea, influencing offshore air quality over the East Sea. Thus, we conclude that regional cities, e.g., Seoul and Busan, contributed to elevated TC NO2 observed by both PSI and OMI on 25 May.

3.2.1.2. Case Study 1, Part 2 (26 May 2016)

The largest surface O3 concentrations on the KORUS-OC cruise were measured on 26 May as the Onnuri sailed on the East Sea toward the industrial city of Pohang (Figure 1). Ozone increased throughout the morning, leading to values above South Korean standards for 1-hour mean ozone (100 ppbv vs 151 ppbv measured) and 8-hour mean ozone (60 ppbv, 105 ppbv measured) in the afternoon (Figure 6). Carbon monoxide reached 1000 ppbv at 1300 KST on 26 May, nearly an order of magnitude greater than earlier in the day. These data, along with elevated OMI TC NO2 data from the previous day (Figure 4b; cf Figure 5 in Tzortziou et al. [2018]), suggest the pollution measured on the Onnuri was influenced by South Korean cities.

Figure 6.

Figure 6.

Open in a new tab

O3 (red) and CO (black) measurements on 26 May. Ozone values increased throughout the day, exceeding both 1-hour and 8-hour local standards. At ~1300 KST, large CO concentrations were also measured, suggesting an influx of local pollution.

Surface and TC NO2 increases coincided with the increases of CO around 1300 KST on 26 May (Figure 6). Morning and afternoon surface NO2 measurements remained elevated compared to cruise mean (Figure 7). CL51 ceilometer backscatter profiles (Figure 5) show relatively high aerosol loading in the first 1-2 km. The backscatter values increase from 25-26 May and are greatest near the surface as the Onnuri approaches the coast on the afternoon of the 26th. BLView-calculated mixed layer heights suggest a lack of vertical mixing with mixed-layer heights remaining below 1 km. Values of TC NO2 increase simultaneously with the in situ surface NO2 at 1300 KST, leading to a high correlation between surface and TC NO2 on 26 May (Figure 3b, red line). Ceilometer measurements show a layer of enhanced aerosol backscatter near the surface. We hypothesize that this thin, well-mixed layer means that most of the pollution was confined to low altitudes, causing surface NO2 and TC NO2 measurements to track each other on 26 May.

Figure 7.

Figure 7.

Open in a new tab

Surface (red) and PSI TC (blue) NO2 comparisons on 26 May from the Onnuri. In situ NO2 values fluctuated throughout the day with the largest concentrations around 16.0 and 18.0 ppbv at 1300 KST and 1530 KST. Low TC NO2 values were measured throughout the morning. A peak concentration of 0.25 DU was measured at 1300 KST by the PSI. The mean in situ and TC NO2 for the cruise are denoted with dashed and dotted black lines, respectively.

HYSPLIT back trajectories (not shown), chemical reanalysis, and regional surface pressure analysis indicated that weak surface low and high pressure systems (“L” and “H” on Figure 8) led to light, westerly flow across South Korea on 26 May (see 850 hPa wind vectors on Figure 8). Winds advected NO2 emissions from major urban areas to the east of the Onnuri (Figure 8c,d, see also Figure 4b). Total Column NO2 from the Busan PSI and NIER surface NO2 data from Busan indicate a region of high NO2 emissions, similar to the chemical reanalysis (Figures 8c,d). Total Column NO2 measurements from the Busan PSI [Herman et al., 2018] showed a peak of 3 DU at 1100 KST, whereas NIER surface measurements displayed early morning and late afternoon concentration NO2 peaks ~50 ppbv. The model indicates the buildup of surface O3 to well above 100 ppbv in the region from 12 to 18 KST (Figures 8a,b) during the daytime. We conclude that a buildup and transport of NO2 emissions from continental sources over Korea promoted photochemical production of O3, which was subsequently advected toward the Onnuri causing the extreme surface O3 pollution measured on 26 May.

Figure 8.

Figure 8.

Open in a new tab

Chemical reanalysis output of O3 (a;b) and NO2 (c;d) from Case Study 1 at 03 UTC (a;c) and 09 UTC (b;d) 26 May 2016. Contoured colors are near-surface values of O3 and NO2 in ppbv, with 850 hPa wind vectors shown as gray arrows. Values for O3 are contoured every 10 ppbv, and NO2 values are contoured every 2 ppbv. The “L” and “H” indicate surface low and high MSLP centers, and the R/V Onnuri location is shown as the open circle on each panel (circle colors change for visibility).

3.2.2. Case Study 2 (30 May 2016)

We compare remotely-sensed and in situ observations during an overpass of the R/V Onnuri by the KORUS-AQ DC-8 Airborne Science Laboratory on 30 May. Sampling was being conducted in the Yellow Sea region ~125 km west of the southeast Korean peninsula (Figure 1). The aircraft made a descending spiral to 130 AMSL near the Onnuri at 1035 KST, recording the vertical distribution of pollutants. As with Case Study 1, relatively low TC NO2 was measured by the shipboard PSI on the Onnuri, mostly 0.18-0.22 DU (Figure 9; cf Figure 3a). Below-average TC NO2 measurements occurred during the overpass, whereas above-average values were measured later on 30 May. Figure 3b, the scatterplot of PSI TC NO2 and in situ NO2 measurements, including the data depicted in Figure 9, shows poor correlation, in contrast to Case 1 (Figure 7e).

Throughout much of the 30 May spiral, there were large offsets between the ship trace gas measurements (NO2, ozone, CO) and temperature compared to those on the DC-8 (Figure 10; ship data on the abscissa). In situ NO2 concentrations on the Onnuri were 10-times larger than DC-8 measurements (2.5 vs. 0.2 ppbv; Figure 10a). However, in situ O3 was 25 ppbv lower on the ship than on the DC-8 (60 vs. 85 ppbv; Figure 10b). Carbon monoxide concentrations measured by the Onnuri were half of those measured by the DC-8 (170 vs. >300 ppbv at 400 m altitude; Figure 10c). These differences imply that the DC-8 was mostly in aged pollution in which NOx had been converted to ozone and the longer-lived CO tracer was still present. However, the DC-8 temperature data showed the vertical structure to be more complex. Three inversions detected by the DC-8 corresponded to three pollution layers in the lower atmosphere, which caused most pollution to be trapped several hundred m above the surface (Figure 10d). A stable surface layer below the DC-8 is indicated by a 2.6 K temperature difference between the Onnuri and what the DC-8 measured at 135 m; it is also consistent with CL51 aerosol backscatter (Figure 11), which shows multiple layers of enhancement within the first 1 km. These divergent layers led to NO2, O3 and CO discrepancies between ship and aircraft (Figures 10a,b,c). A second temperature inversion appeared to confine the O3- and CO-rich air to below 400 m; this is referred to as a Stable Transition Layer in Figure 10d. A separate “mixing layer” where NO2 measurements increased with altitude developed from a ~1 °C temperature inversion at 1200 m AMSL.

Figure 10.

Figure 10.

Open in a new tab

NO2 (a), O3 (b), CO (c), and potential temperature (θ;d) measurements from the DC-8 (circles) and Onnuri (crosses at bottom of each panel) on 30 May. Note the differing scale for Onnuri surface NO2 on panel (a). Small temperature inversions created a stable atmosphere, resulting in three distinct layers above the surface.

Figure 11.

Figure 11.

Open in a new tab

CL51 ceilometer backscatter profiles from CS2 on 30 May. The black dots indicate the BLView-calculated aerosol mixed layer heights. The time of the DC-8 overpass of the Onnuri is indicated on the figure.

Chemical reanalysis results were consistent with HYSPLIT back trajectories (not shown here) and indicated Chinese influence on the air masses sampled by the Onnuri and DC-8 (Figure 12). Model output from 29 May indicated large CO (Figure 12a) and NO2 (Figure 12b) emissions in that region, which were advected east over the Yellow Sea near the Onnuri (see 850 hPa wind vectors on Figure 12). Both trajectories and reanalysis winds display a pattern of emissions circulating around a surface high pressure system (“H” in Figure 12). Photochemical production of O3 from Chinese NOx emissions occurred over open water, evidently contributing to the large O3 values seen in the boundary and mixing layers by the DC-8 in Figure 10. The enhanced model CO concentrations above the Onnuri are consistent with pollution encountered by the DC-8 showing that an aged, polluted air mass originated from regional transport from China.

Figure 12.

Figure 12.

Open in a new tab

Chemical reanalysis output of CO (a) and NO2 (b) from Case Study 2 at 01 UTC on 30 May, 2016. Contoured colors are near-surface values of CO and NO2 in ppbv, with 850 hPa wind vectors shown as gray arrows. CO values are contoured every 25 ppbv, and NO2 values are contoured every 2 ppbv. The “H” indicates the surface high MSLP center, and the R/V Onnuri location is shown as the open circle on each panel (circle colors change for visibility).

The DC-8 overpass allowed us to characterize potential causes for disagreement between TC NO2 and surface NO2 on 30 May. In contrast to the well-mixed, near-surface atmosphere of Case Study 1, complex meteorology with multiple stable layers on 30 May created distinct offsets between the surface and air aloft. Low near-surface NO2, and high CO values indicated by the chemical reanalysis (Figure 12a) match the measurements found just above the surface by the DC-8. It is likely that the surface high pressure (“H” on Figure 12) contributed to the complex, stable atmosphere encountered by the DC-8 over the Onnuri on 30 May. Under these types of conditions over open water we find that it is difficult to relate surface NO2 to total column measurements.

3.2.3. Case Study 3 (1 June 2016)

The highest levels of PSI TC NO2 during the KORUS-OC cruise were observed on 1 June as the R/V Onnuri sailed toward Seoul (Figure 1; see also Tzortziou et al. 2018). The PSI TC NO2 began increasing at 0900 KST, reaching a peak of ≥ 0.5 DU at 1200 KST, more than twice the cruise average (Figure 13). In situ NO2 concentrations began increasing coincidently with TC NO2 at 0900 KST, but reached peak values, 6 ppbv, an hour earlier than the TC NO2. The in situ NO2 concentration fell to less than 2 ppbv at 1200 KST with only a handful of intermittent NO2 readings (fewer than 10%) reaching that level in the next 8 hours (red circles up to 2000 KST in Figure 13). The lag between the TC NO2 and surface NO2 peaks, along with larger than average TC NO2 (dotted line in Figure 13) and lower-than-average in situ NO2, led to poor correlation between the two sets of observations (see Figure 3b). The post-noon divergence of the surface and PSI TC NO2 may indicate that the somewhat elevated TC NO2 results from a high-NO2 layer above the surface. Data from the ceilometer (Figure 14a) showed an influx of air with high aerosol concentrations at the same time as increases in surface and TC NO2, as well as surface CO at 0900 KST (Figure 14b). The aerosol mixed layer continued to grow through the early afternoon over the shallow waters near Seoul (Figure 14a).

Figure 13.

Figure 13.

Open in a new tab

Surface (red) and PSI TC NO2 (blue) comparisons on 1 June as measured on the Onnuri. The largest column NO2 measurements KORUS-OC were observed at 1200 KST (outlined in black). Enhanced in situ NO2 values were observed at 1100 KST. The mean in situ and TC NO2 for the cruise are denoted with dashed and dotted black lines, respectively.

Figure 14.

Figure 14.

Open in a new tab

CL51 ceilometer backscatter profile (color bar; a) and surface CO time series (b) from 1 June on the Onnuri. The black line on (a) shows the BLView-calculated mixed layer height. The black dashed line denotes when the Onnuri encountered a sharp increase in near-surface pollution indicated by increasing CO and enhanced aerosol backscatter (yellow and orange colors on a).

HYSPLIT forward trajectories and chemical reanalysis output show evidence of NO2 emissions transported from the Seoul region, contributing to the high surface and TC NO2 at the Onnuri. We ran forward trajectories from Seoul-Yonsei University and Olympic Park (two highly-urbanized KORUS-AQ sites; not shown) that show air from Seoul reaching the Onnuri at around 1200 KST at an altitude of 1500-1600 m AMSL (Figure S1). Reanalysis winds and CO (Figure 15a,b) and NO2 (Figure 15,c,d) indicate transport of CO and NO2-rich air from Seoul affecting the Onnuri.

Figure 15.

Figure 15.

Open in a new tab

Chemical reanalysis output of CO (a;b) and NO2 (c;d) from Case Study 3 at 01 UTC (a;c) and 03 UTC (b;d) 01 June, 2016. Contoured colors are near-surface values of CO and NO2 in ppbv, with 850 hPa wind vectors shown as gray arrows. CO values are contoured every 25 ppbv, and NO2 values are contoured every 2 ppbv. The “L” and “H” indicate surface low and high MSLP centers, and the R/V Onnuri location is shown as the open circle on each panel (circle colors change for visibility).

This case shows both surface and PSI TC NO2 responding simultaneously to a polluted air mass from Seoul, but TC NO2 measurements are offset in time from surface NO2 measurements. The initial (0900 KST) jump in surface NO2 might have been associated with a pulse of air enriched in CO and NO2 that mixed down over the ship (O3 dropped from ~40 to 35 ppbv coincident with the NO2 and CO increases). Tzortziou et al. [2018] examined the 1 June PSI TC NO2 on the Onnuri in detail, comparing the shipboard readings to the 1 June time-series from the PSI No. 40 in Seoul at Yonsei University. The Seoul PSI 40 TC NO2 levels began to increase from an overnight reading of ~0.4 DU at 0800. Like the Onnuri instrument, PSI 40 TC NO2 continued to increase steadily after 0900, reaching 0.8 DU at 1200 KST (Figure 8d in Tzortziou et al., 2018). There were two additional peaks in the PSI 40 TC NO2 time-series on 1 June, both reaching 1.2 DU over Seoul between 1400 and 1800 KST.

Tzortziou et al. [2018] also ran HYSPLIT back-trajectories from Onnuri locations, starting at four times on 1 June between 1000 and 1600 KST. Trajectories initiated over the Onnuri location before 1000 local time, when TC NO2 over the ocean was <0.25 DU, indicated transport of air masses from a region north of the city of Seoul where TC NO2 is typically relatively low (Figure 10 in Tzortziou et al., 2018). At 1200 local time, however, the air parcels reaching Onnuri were coming mostly from the city of Seoul, coinciding with the peak in TC NO2 observed over the ocean and similar to our HYPSLIT forward trajectories. Note that the Seoul PSI 40 TC NO2 averaged 1.0 (± 0.64) DU over the period 16 May to 15 June 2016 [Tzortziou et al., 2018]. OMI TC NO2 averaged 0.74 DU over the same period but there was a gap in OMI data from 29 May to 11 June 2016. The air parcel origins from the Onnuri the afternoon of 1 June (1300 and 1600 KST) had shifted southeast of Seoul where typical TC NO2, as recorded by two KORUS-AQ PSI during (No. 20 at Taehwa and No. 35 at Yeoju; see Figure 3 in Herman et al. 2018) ranged from 0.4-0.6 DU in May-mid-June, slightly more than half the levels measured by PSI 40. In summary, it appears that although Onnuri in situ NO2 and PSI TC NO2 both increased in response to pollution transport from Seoul on the morning of 1 June, the timing was different. After the midday peaks in both, the divergence of the two signals suggests that a somewhat elevated NO2 layer remained aloft as the wind shift brought less polluted air from south of Seoul toward the ship. This decoupling led to an apparent lack of correlation between surface and TC NO2 as illustrated in Figure 3b.

4. Summary and Conclusion

On the May-June 2016 KORUS-OC cruise trace gas measurements on the R/V Onnuri detected pollution over coastal South Korea and open waters due to transport from both local and regional sources. Each case study we examined represented differing near-surface meteorological conditions and transport patterns that were important factors in determining how closely the shipboard PSI TC NO2 and in situ NO2 analyzer tracked one another. Local sources contributed to elevated NO2 on both the 25-26 May and 1 June pollution events whereas a combination of pollution from upwind sources dominated during an episode on 30 May. On 26 May (Case Study 1), sources near Pohang confined to a shallow surface layer led to a strong correlation between TC NO2 and in situ NO2 measurements (r = 0.73). Long-range transport from China on 30 May (Case Study 2) led to high O3 and CO values measured by the DC-8 just above the surface, but several stable layers in the first 1.5 km AMSL complicated the relationship between surface and column NO2 on the Onnuri (r = −0.11). On 1 June (Case Study 3), NOx emissions from Seoul were transported, leading to the highest TC NO2 measurements during KORUS-OC, with similar, but slightly time-lagged increases in surface NO2. TC NO2 continued to be elevated later in the day, presumably due to transport from a less polluted region south of Seoul, but in situ NO2 was largely unaffected. Thus, the correlation between the two was insignificant (r = 0.04).

Assessing air quality in offshore environments requires a combination of in situ and remote sensing measurements along with detailed meteorological information to provide a complete picture of pollutant transformation and dynamics. These combinations also provide important information for evaluating models and data assimilation as illustrated by the KORUS-OC case studies we have analyzed. Relating surface trace gas observations to the PSI TC NO2 observations proved difficult. In only one of three situations was the PSI sensitive to the surface conditions for Case Study 1 (26 May).

Comparisons in total column NO2 measured by OMI and the PSI were limited due to a 2-week OMI data loss. However, Tzortziou et al. [2018] presented a summary of 7 days when it was possible to compare OMI and the Onnuri PSI TC NO2. Agreement was within ± 0.04 DU on five days; on the other two days PSI NO2 was 3 times higher than OMI (18 May) and 3 times lower than OMI (23 May). This disagreement between the PSI and OMI observations is largely due to the difference in the viewing geometry and spatial resolution of the two remote sensing approaches, and the strong spatial heterogeneity often observed in TC NO2 over the ocean. The OMI nadir view is ~13×24 km and many PSI comparisons were made with coarser footprints and compromised alignment of the two instruments. Preliminary comparisons of a given set of PSI readings and TROPOMI TC NO2 show better agreement and a stronger correlation than with OMI [Kaltenbaugh et al., 2018] due to the higher TROPOMI spatial resolution. Excellent PSI-satellite agreement is also expected for the upcoming GEMS (trace gas resolution 6×8 km) instrument that will launch within a year or so. Nonetheless, our analyses illustrate the challenge of interpreting poor air quality episodes over coastal East Asia, a region of rapidly changing NOx emissions, from column measurements alone.

Three points-

  • Pandora NO2 total column and coincident surface O3, NO2 and CO were measured off coastal Korea in May-June 2016

  • Relationship between Pandora column and surface NO2 depends on meteorology with only 1 of 3 cases well-correlated

  • Air pollution along Korean coast originates from both South Korean and regional sources

Acknowledgments

This work is based on the Senior Thesis of TPB (BS, Univ. Maryland-College Park [UMCP], May 2018). Comments and encouragement throughout the thesis were provided by Dr. Tim Canty (UMCP, Department of Atmospheric and Oceanic Sciences). Supplementary surface observations for Busan were made by the National Institute of Environmental Research (NIER). DC-8 profiles are courtesy of KORUS-AQ participants G. Diskin (NASA/LaRC) and A. Weinheimer (NCAR). We acknowledge the free use of OMI tropospheric NO2 column data from the NASA GES DISC (doi:10.5067/Aura/OMI/DATA2017). Support for the cruise came from the National Aeronautics and Space Administration (NASA) KORUS-OC/AQ field campaign (Grant: NASA NNX16AD60G). This study was partially funded by the U.S. Department of the Interior, Bureau of Ocean Energy Management through Interagency Agreement M17PG00026 with NASA (B. Duncan, PI). Part of this work was performed at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. All the KORUS-OC data used here are available at https://www-air.larc.nasa.gov/cgi-bin/ArcView/korusaq.

References

  1. Al-Saadi J, Carmichael G, Crawford J, Emmons L, Kim S, Song C, Chang L, Lee G, Kim J, and Park R (2015). NASA contributions to KORUS-AQ: An international cooperative air quality field study in Korea (White Paper; ). Retrieved from https://espo.nasa.gov/home/sites/default/files/documents/White%20paper%20outlining%20NASA%E2%80%99s%20contribution%20to%20KORUS-AQ.pdf. [Google Scholar]
  2. Boersma KF, Eskes HJ, Veefkind JP, Brinksma EJ, van der A RJ, Sneep M, van den Oord G,H,J, Levelt PF, Stammes P, Gleason JF, and Bucsela EJ (2007). Near-real time retrieval of tropospheric NO2 from OMI. Atmos. Chem. Phys, 7, 2103–2118, doi: 10.5194/acp$$-$$7$$-$$2103$$-$$2007. [DOI] [Google Scholar]
  3. Draxler RR, and Hess GD (1997), Description of the HYSPLIT_4 modeling system, NOAA Tech. Memo, ERL ARL-224, NOAA Air Resour Lab., 24 pp, Silver Spring, MD. [Google Scholar]
  4. Duncan B, Yoshida Y, De Foy B, Lamsal L, Streets D, Lu Z, Pickering K, and Krotkov N (2013). The observed response of ozone monitoring instrument (OMI) NO2 columns to NOx emission controls on power plants in the united states: 2005-2011. Atmos. Environ, 81, 102–111, doi: 10.1016/j.atmosenv.2013.08.068. [DOI] [Google Scholar]
  5. Duncan B, Lamsal L, Thompson A, Yoshida M, Hurwitz Y, Pickering M,M, E. K, Lu Z, and Streets DG (2016). A space-based, high-resolution view of notable changes in urban NOx pollution around the world (2005-2014). J. Geophys. Res, 121(2), 976–996, doi: 10.1002/2015JD024121. [DOI] [Google Scholar]
  6. Emeis S, Jahn C, nkel C, Münsterer C, and Schäfer K (2007). Multiple atmospheric layering and mixing-layer height in the inn valley observed by remote sensing, Meteorologische Zeitschrift, 16(4), 415–424, doi: 10.1127/0941-2948/2007/0203. [DOI] [Google Scholar]
  7. Flynn CM, Pickering KE, Crawford JH, Lamsal L, Krotkov N, Herman J, A. Weinheimer, Chen G, Liu X Szykman J, Tsay S-C, Loughner C, Hains J, Lee P, Dickerson RR, Stehr JW, and Brent L (2014), Relationship between column-density and surface mixing ratio: Statistical analysis of O3 and NO2 data from the July 2011 Maryland DISCOVER-AQ mission, Atmos. Environ, 92, 429–441, doi: 10.1016/j.atmosenv.2014.04.041. [DOI] [Google Scholar]
  8. Goldberg DL, Saide PE, Lamsal LN, de Foy B, Lu Z, Woo J-H, Kim Y, Kim J, Gao M, Carmichael G, and 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 (3), 1801–1818, 10.5194/acp-19-1801-2019. [DOI] [Google Scholar]
  9. Herman J, Abuhassan N, Kim J, Kim J-H, Dubey M, Raponi M, Tzortziou M (2019) Underestimation of column NO2 amounts from the OMI satellite compared to diurnally varying ground-based retrievals from multiple Pandora spectrometer instruments, Atmos. Chem. Phys, acpd-2019-123, accepted. [Google Scholar]
  10. Herman J, Cede A, Spinei E, Mount G, Tzortziou M, and Abuhassan N (2009). NO2 column amounts from ground-based Pandora and MFDOAS spectrometers using the direct-sun DOAS technique: Intercomparisons and application to OMI validation, J. Geophys. Res.-Atmos., 114(D13), doi: 10.1029/2009JD011848. [DOI] [Google Scholar]
  11. Herman J, Spinei E, Fried A, Kim J, Kim J, Kim W, Cede A, Abuhassan N, and Segal-Rozenhaimer M (2018). NO2 and HCHO measurements in Korea from 2012 to 2016 from PSI spectrometer instruments compared with OMI retrievals and with aircraft measurements during the KORUS-AQ campaign. Atmos. Meas. Tech, 1–60, 1–60. doi: 10.5194/amt-2018-56. [DOI] [Google Scholar]
  12. Jiang Z, McDonald BC, Worden H, Worden JR, Miyazaki K, Qu Z, Henze DK, Jones DBA, Arellano AF, Fischer EV, Zhu L, and Boersma KF (2018). Unexpected slowdown of us pollutant emission reduction in the past decade, Proceedings of the National Academy of Sciences, 115(20):5099–5104, doi: 10.1073/pnas.1801191115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kaltenbaugh AK, Thompson AM, Kollonige DE, Stauffer RM, Sullivan JT, McGee TJ, Swap R, Robinson J, Berkoff T, Knepp TN, Northam TE, and Brothers GB (2018). Comparison of OMI and Pandora Ozone and NO2 Measurements in the Mid-Atlantic Region during the OWLETS 2017 Campaign, AGU Fall Meeting, Washington D.C., Dec. 2018, 2018AGUFM.A51N2389K. [Google Scholar]
  14. Kanaya Y, Miyazaki K, Taketani F, Miyakawa T, Takashima H, Komazaki Y, Pan X, Kato S, Sudo K, Inoue J, Sato K, and Oshima K: Ozone and carbon monoxide observations over open oceans on R/V Mirai from 67° S to 75° N during 2012 to 2017: Testing global chemical reanalysis in terms of Arctic processes, low ozone levels at low latitudes, and pollution transport, Atmos. Chem. Phys. Discuss, 10.5194/acp-2018-1354, in review, 2019. [DOI] [Google Scholar]
  15. Knepp T, et al. (2015). Estimating surface NO2 and SO2 mixing ratios from fast-response total column observations and potential application to geostationary missions, J. Atmos. Chem, 72(3–4), 261–286, doi: 10.1007/s10874-013-9257-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Knepp TN, Szykman JJ, Long R, Duvall RM, Krug J, Beaver M, Cavender K, Kronmiller K, Wheeler M, Delgado R, Hoff R, Berkoff T, Olson E, Clark R, Wolfe D, Van Gilst D, and Neil D (2017). Assessment of mixed-layer height estimation from single-wavelength ceilometer profiles, Atmos. Meas. Tech, 10, 3963–3983, 10.5194/amt-10-3963-2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kollonige DE, Thompson AM, Josipovic M, Tzortziou M, Beukes JP, Burger R, Martins DK, van Zyl PG, Vakkari V, and Laakso L (2018). OMI satellite and ground-based Pandora observations and their application to surface NO2 estimations at terrestrial and marine sites, J. of Geophys. Res. Atmos, 123(2), 1441–1459, doi: 10.1002/2017JD026518. [DOI] [Google Scholar]
  18. Kreher K, Van Roozendael M, Hendrick F, Apituley A, Friess U, Lampel J, Piters A, Richter A, Wagner T, and CINDI- 2 Team (2017). First results of the CINDI-2 semi-blind MAX-DOAS intercomparison, EGU General Assembly, Vienna, Austria, 23–28 April 2017, EGU2017-13927. [Google Scholar]
  19. Lamsal LN, Krotkov NA, Celarier EA, Swartz WH, Pickering KE, Bucsela EJ, … Knepp TN (2014). Evaluation of OMI operational standard NO2 column retrievals using in situ and surface-based NO2 observations. Atmospheric Chemistry and Physics, 14(21), 11,587–11,609. 10.5194/acp-14-11587-2014. [DOI] [Google Scholar]
  20. Levelt P, Van den Oord G, Dobber M, Malkki A, Visser H, De Vries J, Stammes P, Lundell JOV, and Saari H (2006). The ozone monitoring instrument, IEEE Transactions on Geosci, and Remote Sensing, 44(5), 1093–1101. [Google Scholar]
  21. Levelt PF, Joiner J, Tamminen J, Veefkind JP, Bhartia PK, Zweers DCS, Duncan BN, Streets DG, Eskes H, van der A R, McLinden C, Fioletov V, Carn S, de Laat J, DeLand M, Marchenko S, McPeters R, Ziemke J, Fu D, Liu X, Pickering K,Apituley A, Abad GG, Arola A, Boersma F, Miller CC, Chance K, de Graaf M,Hakkarainen J, Hassinen S, Ialongo I, Kleipool Q, Krotkov N, Li C, Lamsal L, Newman P, Nowlan C, Suleiman R, Tilstra LG, Torres O, Wang H, and Wargan K (2018). The ozone monitoring instrument: overview of 14 years in space, Atmospheric Chemistry and Physics, 18(8):5699–5745, doi: 10.5194/acp-18-5699-2018. [DOI] [Google Scholar]
  22. Liu F, Beirle S, Zhang Q, van der A RJ, Zheng B, Tong D, and He K (2017), NOx emission trends over Chinese cities estimated from OMI observations during 2005 to 2015, Atmos. Chem. and Phys, 17 (15), 9261–9275, 10.5194/acp-17-9261-2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lotteraner C and Piringer M (2016). Mixing-Height time series from operational ceilometer aerosol-Layer heights. Boundary-Layer Meteorology: An International Journal of Physical, Chemical and Biological Processes in the Atmospheric Boundary Layer, 161(2), 265–287, doi: 10.1007/s10546-016-0169-2. [DOI] [Google Scholar]
  24. Martins DK, Najjar RG, Tzortziou M, Abuhassan N, Thompson AM, and Kollonige DE (2016), Spatial and temporal variability of ground and satellite column measurements of NO2 and O3 over the Atlantic ocean during the Deposition of Atmospheric Nitrogen to Coastal Ecosystems Experiment (DANCE), J. Geophys. Res, 121(23), 175–14. doi: 10.1002/2016JD024998. [DOI] [Google Scholar]
  25. Martins DK, Stauffer RM, Thompson AM, Knepp TN, and Pippin M (2012), Surface ozone at a coastal suburban site in 2009 and 2010: Relationships to chemical and meteorological processes, J. Geophys. Res, 117(D5). doi:10/1029/2011JD016828. [Google Scholar]
  26. Miyazaki K, Sekiya T, Fu D, Bowman KW, Kulawik SS, Sudo K, et al. (2019). Balance of emission and dynamical controls on ozone during the Korea-United States Air Quality campaign from multiconstituent satellite data assimilation. J. Geophys. Res. Atmos., 124, 387–413. 10.1029/2018JD028912 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Park Y, Yu-Hwan A, and GEO-CAPE Ocean Science Working Group (2015), Risk reduction measurement for GEO-CAPE: A US-Korea joint field campaign (US-Korea JFC) in the East Sea and Yellow Sea (White Paper; ). [Google Scholar]
  28. Reed AJ, Thompson AM, Kollonige DE, Martins DK, Tzortziou MA, Herman JR, Berkoff TA, Abuhassan NK, and Cede A (2015), Effects of local meteorology and aerosols on ozone and nitrogen dioxide retrievals from OMI and Pandora spectrometers in Maryland, USA during DISCOVER-AQ 2011, J Atmos. Chem, 72(3–4), 455–482. doi: 10.1007/s10874-013-9254-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Sachse GW, Hill GF, Wade LO, and Perry MG (1987), Fast-response, high-precision carbon monoxide sensor using a tunable diode laser absorption technique, J. Geophys. Res, 92(D2), 2071–2081, doi: 10.1029/JD092iD02p02071. [DOI] [Google Scholar]
  30. Stein AF, Draxler RR, Rolph GD, Stunder BJB, Cohen MD, and Ngan F (2015). Noaa’s hysplit atmospheric transport and dispersion modeling system. Bull. Am. Meteor. Soc, 96(12), 2059–2077. doi: 10.1175/BAMS-D-14-00110.1 [DOI] [Google Scholar]
  31. Steyn DG, Baldi M, and Hoff RM (1999), The detection of mixed layer depth and entrainment zone thickness from lidar backscatter profiles, J Atmos. And Oceanic Technology, 16(7), 953–953. doi: [DOI] [Google Scholar]
  32. Tan W, Liu C, Wang S, Xing C, Su W, Zhang C, Xia C, Liu H, Cai Z, and Liu J (2018), Tropospheric NO2, SO2, and HCHO over the East China Sea, using ship-based MAX-DOAS observations and comparison with OMI and OMPS satellite data, Atmos. Chem. and Phys, 18 (20), 15,387–15,402, 10.5194/acp-18-15387-2018. [DOI] [Google Scholar]
  33. Tzortziou M, Herman JR, Loughner CP, Cede A, Abuhassan N, and Naik S (2015). Spatial and temporal variability of ozone and nitrogen dioxide over a major urban estuarine ecosystem, J. Atmos. Chem., doi: 10.1007/s10874-013-9255-8. [DOI] [Google Scholar]
  34. Tzortziou M, Thompson AM, and Herman J (2015), Dynamics of atmospheric trace gases and aerosols in Korean coastal waters: Impacts on ocean color atmospheric correction and surface air pollution studies (NASA Project Description, Grant # NNX16AD60G, PI: Tzortziou). [Google Scholar]
  35. Tzortziou M, Parker O, Lamb B, Herman J, Lamsal L, Stauffer R, and Abuhassan N (2018). Atmospheric trace gas (NO2 and O3) variability in Korean coastal waters, implications for remote sensing of coastal ocean color dynamics, Remote Sens, 2018, 10, 1587; doi: 10.3390/rs10101587. [DOI] [Google Scholar]
  36. van Geffen JHGM, Eskes HJ, Boersma KF, Maasakkers JD, & Veefkind JP (2018). TROPOMI ATBD of the total and tropospheric NO2 data products (issue 1.2.0) Royal Netherlands Meteorological Institute (KNMI; ), De Bilt, the Netherlands, s5P-KNMI-L2-0005-RP. [Google Scholar]
  37. Veefkind J, Aben I, McMullan K, Forster H, de Vries J, Otter G, et al. (2012). TROPOMI on the ESA Sentinel-5 Precursor: A GMES mission for global observations of the atmospheric composition for climate, air quality and ozone layer applications. Remote Sens. of Environ, 120, 70–83, doi: 10.1016/j.rse.2011.09.027. [DOI] [Google Scholar]

RESOURCES