Next Article in Journal
Learning Adversarially Robust Object Detector with Consistency Regularization in Remote Sensing Images
Previous Article in Journal
Post-Event Surface Deformation of the 2018 Baige Landslide Revealed by Ground-Based and Spaceborne Radar Observations
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Remote Sensing
Volume 15
Issue 16
10.3390/rs15163998
remotesensing-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

A Comparative Study of Ground-Gridded and Satellite-Derived Formaldehyde during Ozone Episodes in the Chinese Greater Bay Area

by
Yiming Zhao
1,
Xujun Mo
1,
Hao Wang
1,2,3,
Jiangyong Li
1,
Daocheng Gong
1,2,
Dakang Wang
4,5,
Qinqin Li
1,2,
Yunfeng Liu
1,
Xiaoting Liu
1,6,
Jinnian Wang
4,5 and
Boguang Wang
1,2,3,*
1
Institute for Environmental and Climate Research, Jinan University, Guangzhou 511443, China
2
Guangdong Provincial Observation and Research Station for Atmospheric Environment and Carbon Neutrality in Nanling Forests, Guangzhou 511443, China
3
Australia-China Centre for Air Quality Science and Management (Guangdong), Guangzhou 511443, China
4
Institute of Aerospace Remote Sensing Innovations (ARSI), Guangzhou University, Guangzhou 510006, China
5
School of Geography and Remote Sensing, Guangzhou University, Guangzhou 510006, China
6
Department of Ophthalmology, The First Affiliated Hospital of Jinan University, Guangzhou 510630, China
*
Author to whom correspondence should be addressed.
Remote Sens. 2023, 15(16), 3998; https://doi.org/10.3390/rs15163998
Submission received: 8 July 2023 / Revised: 5 August 2023 / Accepted: 7 August 2023 / Published: 11 August 2023
(This article belongs to the Section Atmospheric Remote Sensing)
Browse Figures
Versions Notes

Abstract

:
Formaldehyde (HCHO) plays an important role in atmospheric photochemical reactions. Comparative studies between ground-based and satellite observations are necessary to assess and promote the potential use of column HCHO as a proxy for surface HCHO and volatile organic compound (VOC) oxidation. Previous studies have only validated temporal and vertical profile variations at one point, with limited studies comparing horizontal spatial variations due to sparse monitoring sites. The photochemistry-active Chinese Greater Bay Area (GBA) is a typical megacity cluster as well as a large hotspot of HCHO globally, which recorded a high incidence of ozone (O3) pollution. Here, we conducted the first comparative study of ground-gridded (HCHOgg) and satellite-derived (HCHOsd) HCHO during typical O3 episodes in the GBA. Our results revealed a good correlation between HCHOgg and HCHOsd, with a correlation coefficient higher than 0.5. Cloud coverage and ground pixel sizes were found to be the dominant factors affecting the quality of HCHOsd and contributing to the varying satellite pixel density. Daily averages of HCHOsd effectively improved the HCHOsd accuracy, except in areas with low satellite pixel density. Furthermore, a new quality control procedure was established to improve HCHOsd from Level 2 to Level 3, which demonstrated good application performance in O3 sensitivity analysis. Our findings indicate that the correlation between satellite observations and surface air quality can be optimized by spatiotemporal averaging of hourly HCHOsd, given the advent of geostationary satellites. Considering the representative range of sampling sites in this comparative study, we recommend establishing VOC monitoring stations within a 50 km radius in the GBA to further analyze and control photochemical pollution.

Graphical Abstract

1. Introduction

Formaldehyde (HCHO), one of the 187 Hazardous Air Pollutants (HAPs) categorized by the U.S. Environmental Protection Agency (EPA), is not only a known carcinogen and mutagen but also plays an important role in atmospheric photochemical reactions. HCHO photolysis has a great impact on the atmospheric oxidizing capacity (AOC), which is a fundamental driving factor for the formation of ozone (O3) and secondary organic aerosols (SOA) in the troposphere [1,2,3]. The oxidation of methane and methanol largely defines the global background HCHO [4,5], whereas regional enhancements of HCHO concentration are mainly produced by the oxidation of non-methane volatile organic compounds (VOCs) emitted from biogenic, anthropogenic, or wildfire sources [6,7,8]. Although primary emissions of HCHO in some areas are considered to be the major source (e.g., wildfires’ direct emissions in Alaska [9]), secondary formations of HCHO overall contributed the most in ambient air and can be regarded as an indicator of VOC oxidation [10,11].
Ambient HCHO can be observed from the ground and space. Ground-based HCHO can be detected via offline or online methods. The most common offline method is to use silica cartridges coated with 2,4-dinitrophenylhydrazine (DNPH) to collect air samples for high-performance liquid chromatography (HPLC) analysis [12,13,14,15], which has been considered the international standard [16]. Online methods include differential optical absorption spectroscopy (DOAS), Hantzsch method-based analyzers, and proton transfer reaction–mass spectrometry (PTR-MS) [14]. HCHO from space is detected by 325 to 360 nm solar ultraviolet backscattered radiation [17], and the vertical column densities (VCDs) in units of molecules cm−2 are acquired after the retrieval process. Presently, several polar-orbiting satellite instruments including OMI (Ozone Monitoring Instrument) aboard Aura satellite, OMPS (Ozone Mapping and Profiler Suite) aboard Suomi NPP, GOME-2 (Global Ozone Monitoring Experiment-2) aboard Metop, and TROPOMI (Tropospheric Monitoring Instrument) aboard Sentinel-5 Precursor are used to obtain HCHO VCDs. Among these satellite instruments, TROPOMI boasts the highest resolution [18]. Some geostationary satellites, such as GEMS (Geostationary Environment Monitoring Spectrometer), TEMPO (Tropospheric Emissions: Monitoring of Pollution), and Sentinel-4, will be used to detect hourly HCHO VCDs over eastern Asia, North America, and Europe, respectively, in the future [19,20,21].
Surface HCHO data were used to analyze spatiotemporal variations, determine sources [22,23,24], discuss regional transport characteristics [25,26], and investigate the effects of meteorological factors on HCHO concentrations [27,28]. Meanwhile, satellite retrievals of column HCHO have been extensively researched over the past few decades. Previous studies have used column HCHO data to top-down constraint on VOC emissions [6,7,8,29,30,31,32,33,34,35,36,37,38], quantify organic aerosol abundance [39], map hydroxyl variability [40,41], improve ground-level O3 spatial distribution [42,43], diagnose O3 sensitivity based on the HCHO to NO2 ratio (FNR) [44,45,46], and estimate cancer risks of outdoor HCHO exposure [47,48]. The use of column HCHO for air quality supervision is more prevalent than surface HCHO on a regional scale, as ground-based observations have limited regional coverage.
Comparative studies between ground-based and satellite observations are necessary to better characterize surface HCHO concentrations. HCHO and NO2 are regarded as proxies for O3 photochemistry [49,50]. Similar comparative studies of spatiotemporal correlation for surface and column NO2 have been carried out in many previous studies. Consistent NO2 trends and variations could be obtained between satellite observations from OMI or TROPOMI and ground-based observations from in situ monitoring [51,52,53,54]. However, most previous comparative studies have used multi-axis DOAS (MAX-DOAS) to validate temporal and vertical profile variations at one point [55,56], with fewer studies comparing horizontal spatial variations due to sparse monitoring sites [44]. Unlike the relatively complete surface NO2 monitoring networks, VOCs, such as HCHO, are only measured at a limited number of stations, and the sparse coverage of HCHO monitoring sites makes it difficult to capture their spatial variability. Therefore, satellite observations are becoming increasingly important for accurately characterizing surface HCHO concentrations. Recently, Wang et al. [57] evaluated HCHO in America between 2006 and 2015 by comparing surface measurements from the EPA and satellite observations from OMI. They found that the correlation between satellite observations and surface measurements had seasonal and regional variations. Explanations of these variations are insufficient for other hotspot areas of the world.
Photochemical pollution in China has become a significant concern. Ground-based HCHO measurements have mainly been concentrated in fast-developing areas, such as the North China Plain (NCP) [58,59], the Yangtze River Delta (YRD) [60,61], and the Greater Bay Area (GBA) [15,22,24]. Studies on column HCHO have also increasingly focused on China in recent years [11,62,63,64,65,66,67,68]. As a typical megacity cluster, GBA is a hotspot with active photochemical processes on a global scale, which are caused by the wide distribution of biogenic and anthropogenic sources, warm climate, and strong solar radiation. Autumn is the peak season for photochemical pollution in the GBA, with the highest O3 concentration of the year [69,70,71], while in recent years, serious O3 pollution events have also been found in the spring [72,73]. Compared with most areas in China, the level of HCHO concentration in GBA is relatively high [14,74], which contributes significantly to photochemical pollution. Therefore, research on the concentration characteristics and distribution of HCHO in the GBA has certain scientific significance for understanding photochemical pollution.
Herein, grid field measurements during O3 episodes in spring and autumn were carried out to evaluate the spatial agreement between ground-gridded (HCHOgg) and satellite-derived (HCHOsd) HCHO in the GBA. To improve the relevance of HCHOsd with the spatial distribution of HCHOgg, we improved HCHOsd to a new dataset. This study also assessed the potential of column HCHO as a proxy for surface HCHO. The different methods of surface O3 sensitivity analysis using satellite data and ground-based data were compared to evaluate the roles of the new satellite dataset.

2. Methodology

2.1. Grid Sampling Campaigns

This study primarily focused on conducting a case analysis of two photochemical pollution events that occurred in the spring and autumn of the GBA, as indicated by the location shown in Figure S1. According to the air quality forecast results and monitoring network data analysis in Guangdong Province, weak winds and unfavorable atmospheric diffusion conditions caused O3 pollution risks in the GBA on 23 April (O3 episode in the spring, episode-S) and 17 September (O3 episode in the autumn, episode-A) of 2021. Therefore, 4-h grid sampling campaigns were conducted in the morning and early afternoon of these two days. These two specific time periods were chosen as suitable representations of HCHO characteristics. The morning sampling period represented the primary emissions resulting from human activities [75]. The early afternoon sampling period reflected the peak of photochemical pollution and secondary formation based on forecasts [24].
The spatial distribution of the sampling sites is shown in Figure 1. The 200 km × 200 km areas of the GBA region were divided into 100 grids (20 km × 20 km for each grid). Thirty-five sampling sites in 33 grids were chosen to carry out concurrent field measurements of HCHO. To ensure the representativeness of the measurements for regional HCHO levels, the 33 grids selected for HCHOgg covered a significant portion of the GBA, including urban, suburban, and rural areas. In order to capture the average concentration of each grid as accurately as possible, all sampling sites were carefully chosen to be located in open areas, avoiding direct influences from anthropogenic sources such as traffic and industrial emissions, as well as biological sources, such as plant emissions. The selection of sampling locations included official monitoring stations, parks, and building rooftops.
The meteorological parameters and trace gas data were obtained from the Guangdong Province Automatic Air Quality Monitoring Network (http://113.108.142.147:20061, accessed on 26 April 2022) and were jointly used for comparison with satellite observations. A total of 218 monitoring stations in the GBA were selected, and VOC data were only available from 9 of these sites during the sampling date. The VOC online analysis systems at each monitoring station were slightly different, but all devices are based on the principles of GC-FID or GC-MS/FID analysis. To obtain accurate and reliable VOC data and ensure comparability between different brands of instruments, each site performed a unified instrument quality control check every two months, including zero point noise (≤0.05 nmol/mol), accuracy (±10%), precision (≤10%), and instrument parallelism (≤20%). For detailed site information, please refer to Supplementary Tables S1 and S2. The hourly data of wind field direction used in Figure 2 and Figure S2 were obtained from the European Centre for Medium Range Weather Forecasts (ECMWF) ERA5 dataset.
According to the method proposed by Louie et al. [75], the GBA region was divided into three sub-regions, i.e., the Western GBA (WGBA), the Pearl River Estuary (PRE), and the Eastern GBA (EGBA). The PRE mainly includes southern Guangzhou, eastern Zhuhai, western Shenzhen, western Dongguan, Zhongshan, and Macao. This area has a high level of urbanization, high population density, a dense transportation network, and intensive industry [76]. In addition, this area gathers the main ports in the GBA with large ship emissions [77,78]. The PRE has been thought of as a key area for heavy O3 pollution in southern China [70], where O3 and its precursors can easily be transported and accumulated [79]. The EGBA mainly includes eastern Dongguan, eastern Shenzhen, Huizhou, and Hong Kong. There is a quantity of industry emissions in Dongguan, while Huizhou has a great proportion of emissions from biogenic sources [80]. The WGBA includes Zhaoqing, Foshan, Jiangmen, and western Zhuhai. Foshan has the largest number of motor vehicles and industrial resources in the WGBA [80], which causes the most serious air pollution.

2.2. Sampling Method and Analytical Instrument

To obtain ground-based HCHO data, a DNPH/HPLC method similar to that by Wang et al. [57], was chosen. Following the standard procedure of EPA TO-11A, the air samples were collected into DNPH-coated silica cartridges (Waters Sep-Pak DNPH-silica, WAT037500). The DNPH tube was connected in series with the O3 removal tube and sampling pump to sample without interference from O3 in the air. According to the results of the breakthrough test and pre-test, the sampling flow rate was set to 1 L/min and the sampling time was 4 h. After the sampling campaigns, all samples were stored in a portable refrigerator and transported back to the laboratory on the same day. All samples were extracted with acetonitrile within one week, and the eluates were stored at a temperature below 4 °C. Sample analysis was completed within one month using high pressure liquid chromatograph mass spectrometry (HPLC-MS, Agilent G6400). More details about the analytical instrument used and the types of data can be found in our recent studies [81].
Quality assurance and quality control (QA/QC) systems were implemented to ensure data quality. Analysis of certification blanks and laboratory blanks was performed before and after sampling. During the two grid sampling campaigns, a total of 140 field samples, 54 field blank samples, and 22 parallel samples were collected and analyzed. All blank results conformed to TO-11A blank criteria. The linearity of the standard curves of all samples was greater than 0.995. The instrumental detection limit of HCHO was 0.0006 μg/m3. In addition, all sampling and analysis procedures were well documented to maintain traceability. More details about QA/QC information could be referred to in the experimental research of Chi et al. [13].

2.3. Satellite Data (TROPOMI HCHO and NO2 Observations)

TROPOMI Level 2 products of tropospheric HCHO and NO2 VCDs were obtained in this research. TROPOMI provides daily global observations, and the overpass time is 13:30 local solar time, with the strongest photochemical reactions during the day [24]. The spatial resolution at nadir has been increased from 3.5 km × 7 km to 3.5 km × 5.5 km since 6 August 2019. The TROPOMI HCHO and NO2 retrieval algorithms are very similar to the OMI dataset reprocessed by QA4ECV (Quality Assurance for Essential Climate Variables project), while TROPOMI significantly improves the precision of the observations at short temporal scales [82], which is helpful for our case study.
Offline satellite data with a quality assurance value (QA_value) above 0.5 (no error flag, cloud radiance fraction (CRF) at 340 nm < 0.5, solar zenith angles ≤ 70°, surface albedo ≤ 0.2, no snow/ice warning, air mass factor > 0.1) were selected in order to avoid misinterpretation of the data quality. According to past research, early afternoon overpass time corresponds to the daily peak of HCHO [55,83]. Hence, in our study, HCHOgg in the afternoon when a high AOC was forecasted was mainly selected for comparison with HCHOsd. The oversampled method, following Zhu et al. [34] and Sun et al. [84], was used in this paper to characterize the spatial distribution of the TROPOMI 2021 average column HCHO for GBA (Figure 1). On average, the HCHO column amounts are higher where anthropogenic emissions of VOCs are high, including Shenzhen, Dongguan, Southern Guangzhou, and Foshan [80].

2.4. Observation-Based Model

F0AM (https://sites.google.com/site/wolfegm/models, accessed on 2 April 2023) combined with the latest version of MCM (MCM v3.3.1; http://mcm.leeds.ac.uk/MCM/, accessed on 2 April 2023) was primarily used to simulate in situ atmospheric photochemical processes and quantify key metrics, including O3 production rate, AOC, OH reactivity, and ROX radical budgets. The F0AM model used in this study included 142 non-methane VOCs and over 17,000 elementary reactions of 5900 species [85]. It considered the physical process of deposition within the boundary layer height (BLH) and inputted observation parameters, including gaseous pollutants (i.e., O3, CO, NO, NO2, HONO, SO2, and VOCs), meteorological parameters (i.e., T, P, and RH), and photolysis rate constants (JO1D, JNO2, JH2O2, JHONO, JHCHO, and JNO3), as constraints [86]. The photolysis rates of additional molecules, including OVOCs, were parameterized based on the solar zenith angle and then adjusted according to JNO2 measurements [87]. The model was pre-run for 5 days to initialize unmeasured compounds and radicals [88].
In this study, relative incremental reactivity (RIR) was used as an index to diagnose the sensitivity of O3 formation to precursors. RIR is defined as the ratio of the differences in O3 production rate to the difference in precursor concentrations [89]. The trace gas data, including VOCs from 9 monitoring stations, were used for F0AM to calculate RIR values. VOC components were the 56 O3 precursor mixtures (PAMS) designated by the US EPA, including 29 alkanes, 10 alkenes, 1 alkyne, and 16 aromatic hydrocarbons. The production rate of O3 (P(O3)), the destruction of O3 (D(O3)) and the net O3 production rate (Pnet(O3)) were calculated using the same equations as Liu et al. [90]. ΔX/X represents the percentage reduction in the input concentrations of each targeted O3 precursor group, and the value adopted was 20% [70]. The formula for RIR is given by:
RIR = ( ∆ P net ( O 3 ) / P net ( O 3 ) ) / ( ∆ X / X ) ,

3. Results

3.1. Data Overview

The average concentrations of ground-based HCHO, O3, NO, NO2, CO, and meteorological parameters during the two sampling periods are shown in Tables S3 and S4. It can be seen that the HCHOgg (5.86 ± 1.68 μg/m3 and 6.86 ± 2.02 μg/m3 for episode-S and episode-A, respectively) were comparable to that observed in Guangzhou during the 2018 O3 episodes reported by Wang et al. [91]. Compared to the morning, HCHOgg in the afternoon increased dramatically with the increased temperature and decreased humidity. This indicates the potential role of the secondary photochemical generation of HCHO [92].
Most of the in situ trace gas concentrations did not exceed the national ambient air quality standards in GB 3095-2012 Ambient Air Quality Standards [93], except for O3. Severe O3 pollution occurred in the episode-S and episode-A afternoon, with average levels of 163.08 ± 41.69 μg/m3 and 162.09 ± 37.07 μg/m3, respectively, especially for the PRE and EGBA regions. The more significant influence of photochemical pollution on HCHO is due to the higher secondary production of HCHO in environments with high levels of O3 [94]. Overall, photochemical pollution episodes characterized by high concentrations of O3 and HCHO were mainly concentrated in the PRE and EGBA areas during the sampling periods.

3.2. Comparison of Surface and Column HCHO

3.2.1. Spatial Distribution Characteristics

Figure 2 presents the spatial distribution characteristics of ground-based HCHO, O3, CO, and NO2. During the morning sampling periods, primary emissions influenced the spatial distribution of HCHOgg, which was consistent with CO and NO2. However, in the afternoon, secondary generation became the main source of HCHO, and the spatial distribution of HCHOgg was more correlated with O3. HCHO pollution was mainly concentrated in the central urban agglomeration of the GBA with densely populated traffic and industrial sources, with concentrations decreasing radially outwards. The lightest pollution was found in Zhuhai and Jiangmen in the WGBA. The O3 pollution diffusion directions were consistent with the wind field directions (Figure 2d,l). This suggests that the regional transport of O3 contributed significantly to EGBA and PRE during the O3 episodes. In episode-S, high concentrations of HCHO and O3 in the afternoon were further north compared to episode-A, likely due to seasonal variations. As shown in Figure S3, high values of the TROPOMI average column HCHO in spring were further north than in autumn because of the monsoon [95].
The spatial distribution characteristics of HCHOsd are shown in Figure 3. It can be seen that HCHOsd was relatively high at the convergence of Guangzhou, Foshan, Dongguan, Huizhou, and Shenzhen, and the HCHO pollution in Dongguan was the most serious. During episode-S and episode-A, the pollution in northern and western Dongguan was significantly higher than in other areas, which was similar to the results from grid field measurements. The weak spatial consistency between HCHOgg and HCHOsd in some regions (e.g., the northern part of Guangzhou) was likely due to limited sampling sites.
In this study, grid sampling sites were considered to represent the average HCHO of the 20 km × 20 km grids they were located in. To verify the representativeness of these sites for the grids, the average values of the HCHOsd in the selected grids were used to make a spatial comparison with HCHOgg in the WGBA, PRE, and EGBA. As shown in Figure 4, during the O3 episodes, the agreement in spatial patterns between HCHOgg and HCHOsd was relatively high. The surface and column HCHO in the EGBA and the PRE were much higher than those in the WGBA, while the WGBA was the region with the highest average column HCHO from TROPOMI during the spring and autumn of 2021. Again, this points to the large regional transport contribution to photochemical pollution in the EGBA and PRE during episode-S and episode-A. Additionally, the HCHOsd during the O3 episodes were two or even four times higher than the seasonal average values, which was consistent with previous studies [94].

3.2.2. Correlation Comparison

HCHOgg were compared with overlapping pixels for HCHOsd at 35 sampling sites, revealing a good relationship between HCHOgg and HCHOsd during the spring (R = 0.77, p < 0.01) and autumn (R = 0.52, p < 0.01) sampling periods (Figure 5a,b). This indicated that column HCHO could represent surface HCHO environments, as revealed by previous studies [44,45,50]. However, as shown in Figure 5c,d, the spatial correlation between HCHOgg and average HCHOsd during episode-S (R = 0.74, p < 0.01) was much higher compared to episode-A (R = 0.22, p = 0.24). This can be attributed to the low density of satellite pixels in some grids, such as E2, F4, and J1, as highlighted in Figure S4b. The spatial correlation coefficient increased to 0.43 when these grids were excluded. The pixel density was important for the spatial correlation between HCHOgg and average HCHOsd. Overall, grid field sampling sites represented the average HCHO of 20 km × 20 km grids to a certain extent. The correlation coefficients in Figure 5c,d were lower than in Figure 5a,b, indicating that referencing the HCHOsd from other areas within the grids we had divided did not effectively improve the correlation between HCHOgg and HCHOsd. In contrast, including HCHOsd farther from the sampling sites had a negative impact on the correlation.

3.2.3. Dominant Factors

Greater solar zenith angles, frequent cloud coverage, and lower HCHO emissions were found to be the dominant factors causing the insignificant correlations between ground-based and satellite observations in the USA [57]. However, in other regions, the dominant factors were different due to different geographical locations and atmospheric environments. Further analysis of pixel density and meteorological conditions revealed that the weak correlation in the GBA during episode-A was attributed to several factors, including significant cloud coverage, larger ground pixel sizes, and different horizontal wind speeds between 1000 hPa and 850 hPa.
Cloud coverage was a well-known factor affecting the quality of satellite observations, leading to missing data and retrieval errors [54,57]. As shown in Figure S5, GBA was barely affected by cloud coverage at 13:00 LT during episode-S, while more cloud coverage was present during episode-A. Additionally, daily observation pixel sizes are affected by the distances between satellite and ground pixels [96]. The pixel size of OMI varies from 13 km × 24 km at nadir to 150 km × 40 km at the edges of the swath [97], and the largest pixels were excluded in some studies to reduce spatial smearing of satellite data [51]. For TROPOMI, the largest pixels are around 14 km wide [82]. Although the increased spatial resolution reduced the effect of pixel size, larger pixels reduced the amount of available HCHOsd in the GBA region during episode-A (Figure S4). The different cloud coverage and the ground pixel sizes affected the data quality and contributed to the varying densities of satellite pixels.
Wind fields between 1000 hPa and 850 hPa (Figure S2) revealed that wind directions were relatively consistent during episode-S, but chaotic and differing during episode-A. Although the chaotic wind directions might contribute to the weak correlation, their impact was relatively small because the wind directions below 950 hPa, where high HCHO concentrations are primarily distributed [94], were consistent during episode-A. Additionally, NO2 showed a consistent correlation between monitoring stations and TROPOMI observations during both episode-S and episode-A with r values of 0.72 and 0.77, respectively. Based on the technical documents (NO2 ATBD), the data quality of TROPOMI NO2 was less affected by cloud coverage than that of TROPOMI HCHO.

3.3. Development of HCHOsd

3.3.1. Mathematical Techniques for Accuracy Improvement

In order to better characterize surface HCHO, we tested methods to improve the accuracy of HCHOsd. Mathematical techniques, including averaging and kriging methods, have been found appropriate to a certain extent [54,98,99]. The technical documents (HCHO ATBD) also recommended increasing precision by averaging the satellite observations over space or time. The TROPOMI vertical column product requirements stated that the single-measurement precision should be 12 × 1015 molec. cm−2 but 4 × 1015 molec. cm−2 at a 20 km spatial resolution [82]. De Smedt et al. [82] demonstrated a significant improvement in TROPOMI HCHO precision when daily averages within 20 km radius circles were considered instead of individual pixels. Our study also yielded similar results, with the spatial correlation between HCHOgg and HCHOsd improving to 0.82 and 0.59 during episode-S and episode-A, respectively (Figure S6). The ordinary kriging method is proposed to address the issue of sparse and uneven data caused by gaps due to cloud coverage and produces a final regular sampling with a resolution of 1 km × 1 km [54]. Figure S7 shows the spatial correlation between HCHOgg and HCHOsd after applying the kriging method with an improvement to 0.82 and 0.65 during episode-S and episode-A, respectively. Both the averaging and kriging methods improved the accuracy of HCHOsd by adding nearby pixels as a data reference, which increased the correlation between HCHOgg and HCHOsd.

3.3.2. Sensitivity Tests

To evaluate the robustness of the two methods and increase the correlation between HCHOgg and HCHOsd, sensitivity tests were performed. Different radii (5.5–90 km) were used around the sampling sites to detect possible spatial resolution effects. The same procedures were applied to the HCHOsd after kriging. As shown in Figure 6a, the correlation showed a marginally small dependency on the radius during episode-S, indicating a relatively large natural dispersion of the HCHO columns [82]. During episode-A, the correlation clearly increased with the radius (0.34 at 5.5 km, 0.68 at 50 km). This meant that further neighboring pixels were needed to improve HCHOsd accuracy under low pixel density. The kriging method tried to maximize the correlation between HCHOgg and HCHOsd; hence, lower correlation appeared with the increasing radius of kriging averages (Figure 6b). Spatial correlation coefficients obtained by the direct averaging method (Figure 6a) and the averaging method after kriging (Figure 6b) declined significantly when the radius was greater than 50 km, suggesting a representative range of the grid sampling sites within 50 km.
In addition, during episode-A, although relatively high r values appeared in the kriging method, the long-distance satellite data may overly intervene, leading to values of HCHOsd that are too similar for each point, and the slope between HCHOgg and HCHOsd become too small (slope < 1), as seen in Figure S7. Therefore, the kriging method may not be appropriate when the pixel density was low, although the spatial correlation between HCHOgg and HCHOsd was increased. Similar issues can arise when daily averages within large-range radius circles (>20 km) were used (Figure S8). To ensure the credibility of the data, it is recommended to use daily averages within 20 km radius circles.
Different CRFs were used to evaluate the impact of clouds on the spatial correlation between HCHOgg and average HCHOsd within 20 km radius circles. The results in Table 1 showed that applying a cloud correction did not significantly alter the spatial correlation coefficient of HCHO when the ranges of CRF were adjusted between 0.4 and 0.6, while the correlation coefficient dropped sharply beyond this range. The impact of clouds was more prominent during episode-A. A more stringent cloud filter reduced the amount of available satellite data, while the spatial correlation coefficient did not improve due to a decrease in pixel density. Conversely, a more lenient cloud filter could increase pixel density but could result in lower data quality. Therefore, it appeared limited to improve the correlation between HCHOgg and HCHOsd through cloud correction in the GBA.

3.3.3. Improvement from TROPOMI Level 2 to Level 3

Based on the above analysis, we found that the average HCHOsd within a 20 km radius range could effectively improve the spatial correlation with HCHOgg, especially under high pixel density, while the spatial correlation was poor in some areas with low pixel density. To improve the relevance of TROPOMI Level 2 HCHO and better characterize surface HCHO concentrations, a new HCHO dataset was created as Level 3. This dataset uses satellite data within a 20 km radius circle from a ground point only if there are more than 10 pixels; otherwise, the data at this point will be considered poor quality and screened out. Considering that the averaging method did not improve the spatial resolution of satellite data, Level 3 HCHOsd were mapped onto the grids of 5.5 km × 5.5 km, which was similar to the TROPOMI spatial resolution at nadir.
To assess the quality and capability of Level 3 HCHOsd, we re-gridded Level 2 HCHOsd to the same spatial resolution. When we passed from Level 2 to Level 3 (Figure 7), the spatial distribution of the HCHOsd became more uniform, and some individual points missed due to the poor data quality of single pixels were repaired by the averaging method. However, it could be observed that the impact of cloud coverage on Level 3 was more obvious. The persistent challenge of cloud coverage remains a significant obstacle to satellite HCHO detection. To compare with the kriging results (Figure 3), Level 3 HCHOsd displayed similar spatial distribution during episode-A, while high values of Level 3 HCHOsd in western Dongguan were found more obvious during episode-S, which was consistent with the results of grid field measurements. In terms of spatial patterns (Figure 8), Level 3 HCHOsd showed significantly better agreement with HCHOgg compared to Level 2 HCHOsd. This improvement was particularly evident during episode-A from 0.52 to 0.7, despite a slight decrease in the number of available data from 27 to 24.

3.4. Applications of TROPOMI Level 3

Since this study primarily evaluated the ability of column HCHO to characterize surface HCHO and assessed the potential of column HCHO as a proxy for surface HCHO during photochemical pollution episodes in the GBA, the application of Level 3 HCHOsd should also involve the analysis of surface O3 pollution issues. FNR has been used in previous studies to diagnose O3 sensitivity [44,45,46]. Wang et al. [67] reported the FNR thresholds for major Chinese cities (including the GBA region), which defined three regimes: a VOC-limited regime occurs when FNR < 2.3, a NOx-limited regime occurs when FNR > 4.2, and a transition regime occurs when 2.3 < FNR < 4.2. The spatial distribution characteristics of Level 3 FNR and the increase in O3 (ΔO3) were shown in Figure S9. ΔO3 was an indicator of O3 production and could be calculated as the difference between daytime O3 (12:00–15:00 LT) and nighttime O3 (00:00–04:00 LT). The results showed that the high values of the ΔO3 always appeared in the VOC-limited regime and the transition regime. This finding was consistent with previous assumptions that the peak values of ΔO3 occurred during the transition from a VOC-limited to NOx-limited regime [46,94].
This study calculated the FNR during two O3 episodes using Level 2 and Level 3 datasets and compared them with model-calculated RIR to assess the application of satellite data. Few studies compared RIR and FNR to diagnose the sensitivity of O3 formation, while we found similar diagnostic results between them by analyzing the data including VOCs from 9 monitoring stations during the sampling period. Figure 9 presents the RIR values for major groups of O3 precursors, including anthropogenic VOCs and biogenic VOCs, during episode-A. The in situ O3 production was highly sensitive to VOC, especially for anthropogenic VOCs, with RIR values ranging from 0.28 to 3.46. This was followed by CO (RIR: 0–0.85) and biogenic VOCs (RIR: 0.12–0.49). These results suggest that anthropogenic activities and flourishing vegetation emissions have significant impacts on O3 production [100,101]. The RIR values for NOx (RIR-NOx) were mostly negative, ranging from −4.71 to −0.62, with the exception of CHTH (RIR: 0.01), HGS (RIR: 0.03), and ZML (RIR: 0.48). These small and even negative RIR-NOx indicated that most of the monitoring stations were in VOC-limited regimes, which was similar to the results for average FNR at the same grids as the VOC monitoring stations (Figure 9). Furthermore, the distribution trend of RIR-NOx and FNR was relatively consistent, as indicated by a Spearman correlation coefficient of 0.867. The distribution trend of RIR-NOx and FNR were related to the O3 sensitivity. Therefore, when the distribution trends of both methods aligned, both RIR-NOx and FNR methods were effective in diagnosing regional O3 sensitivity.
We assumed that the ground observations and the observation-based model were accurate and used RIR-NOx to evaluate the accuracy of the FNR by TROPOMI Level 2 and Level 3. As shown in Table 2, by Spearman correlation analysis, the correlation between RIR-NOx and Level 2 FNR was significant during episode-S (p < 0.05) but not during episode-A (p > 0.05). As mentioned above, the data quality of the Level 2 HCHOsd was greatly affected by pixel density. The correlation between RIR-NOx and FNR improved significantly from Level 2 (R = 0.429, p = 0.337) to Level 3 data (R = 0.857, p = 0.014) during episode-A. The FNR results obtained by Level 3 (Figure S10) showed that the monitoring stations were mostly in a VOC-limited regime and transition regime during the two sampling periods, except for CHTH during episode-A. CHTH was also the only region where RIR-NOx was much higher than RIR-VOCs. Overall, the Level 3 dataset demonstrated good application performance when the pixel density was low.
In addition, correlations were relatively significant when FNR values were obtained by the Level 2 dataset within 20–40 km radius circles (Table 2). Similar to the comparisons of HCHOgg and HCHOsd, correlations declined significantly when the radius was greater than 50 km. This suggested that the representative range of sampling sites was within 50 km.

4. Discussion

Based on the concentration and spatial distribution characteristics of trace gases derived from ground-based and satellite observations, HCHO exhibited typical spatiotemporal variations and seasonal representativeness during the two selected O3 episodes. High concentrations of HCHO were observed in areas with abundant anthropogenic sources, particularly in the early afternoon, when photochemical processes were active. The regional transport of O3 and HCHO played a significant role in the GBA, and the impact of monsoon on HCHO spatial distribution was evident.
Compared to some previous studies that used seasonal and interannual average HCHO in America [47,57], our study showed that daily column HCHO also had good agreement with surface HCHO in the hotspot GBA. During O3 episodes, the HCHO levels are significantly elevated at the near surface. As a result, there is a stronger correlation between surface HCHO and column HCHO. Moreover, similar to averaging HCHO over time, averaging over space can also improve the accuracy of HCHOsd [82], and our study demonstrated an improved correlation between HCHOgg and HCHOsd by this method. It is effective to target O3 episodes research through daily averages of HCHOsd.
Linear relationships between HCHOgg and HCHOsd differed during episode-S and episode-A, which highlighted the challenge of using column HCHO alone to reconstruct surface HCHO due to differences in the physical characteristics of observation methods. Peak values of HCHO occur at an altitude of 0.2 km rather than at ground level, regardless of whether it is a polluted or non-exceedance day [94]. Despite this limitation, combining the analysis of these two datasets can provide valuable insights into the chemical and meteorological processes that impact HCHO, and HCHOsd demonstrated a good application performance in O3 sensitivity analysis through verification with the RIR method based on ground observations.
Although the two days of grid sampling campaigns have some representativeness in evaluating the ability of column HCHO to characterize surface HCHO, and pixel density has been considered an important factor, other factors such as different wind directions between 0 km and 0.2 km altitude and temperature variation still need to be investigated through further comparative research. Moreover, the potential applications of satellite observations for health and air quality communities depend on the ability of column HCHO to explain patterns and trends of surface HCHO. Surface HCHO could be detected at some VOC monitoring stations in the GBA, which could be used for long-term trend analysis in the future.

5. Conclusions

This study is the first comparative analysis of satellite observations and grid field measurements in a typical photochemical-active megacity cluster GBA. We evaluated the ability of column HCHO to characterize surface HCHO during periods of high HCHO and O3 levels with severe photochemical pollution. Different from the comparison between MAX-DOAS and satellite data at a single point in previous studies, horizontal spatial variations of HCHO were prioritized in this paper.
HCHOsd in the GBA had a good correlation with HCHOgg, especially when cloud coverage was minimal and satellite-to-ground pixel distances were small. Combined with the concept of grid field measurements, comparing the surface HCHOgg and the average HCHOsd within the 20 km × 20 km grid, it was discovered that the pixel density greatly impacted spatial correlation between HCHOgg and average HCHOsd. In order to improve the TROPOMI product, mathematical techniques, including averaging and kriging methods, were tested. Through sensitivity analysis, it was determined that: (1) daily averages within 20 km radius circles were recommended to ensure data credibility, (2) while increasing pixel density by relaxing quality control standards (i.e., CRF) only slightly improved HCHOsd. The original standard of QA_value (CRF < 0.5) is applicable to GBA.
Based on the averaging method, we established a TROPOMI Level 3 HCHO by limiting the pixel density as the quality control procedure. This new dataset can effectively correlate between HCHOgg and HCHOsd on days with low pixel density. Furthermore, we also proposed the potential application of the Level 3 dataset. We found that: (1) the FNR method based on satellite observations and the RIR method based on ground observations exhibited consistent results for O3 sensitivity analysis, and (2) using the Level 3 dataset can effectively screen out areas with poor FNR results and provide data support for O3 issues during photochemical pollution periods in regions without ground-based observations.
Our results provide a reference for photochemical pollution during other time periods and in other urban hotspots. In addition, the correlation between satellite observations and surface air quality will be optimized by spatiotemporal averaging of hourly HCHOsd, given the advent of geostationary satellites. Of course, due to the limited number of days of ground-based observation, further comparative research is needed to demonstrate the influence of meteorological factors on satellite observation capabilities.
Finally, it was considered that the representative range of the VOCs, including HCHO sampling sites in the GBA, was within 50 km. Hence, we suggested establishing VOC monitoring stations within a 50 km radius in the GBA area, which can more effectively assess the potential of column HCHO as a proxy for VOC emissions and oxidations from patterns and trends, and further analyze surface HCHO and O3 problems, leading to more fitting pollution reduction policies.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/2072-4292/15/16/3998/s1, Table S1: Information of the grid field sampling sites; Table S2: Information of the VOCs monitoring stations; Table S3: Ground-based HCHO, related trace gases and meteorological parameters in the GBA during the episode-S; Table S4: Ground-based HCHO, related trace gases and meteorological parameters in the GBA during the episode-A; Figure S1: The location of GBA and the main cities in this area. Figure S2: Meteorological data according to the European Centre for Medium Range Weather Forecasts (ECMWF) ERA5 data; Figure S3: 2021 spring and autumn average HCHO vertical column density from TROPOMI; Figure S4: Pixel distribution of HCHOsd observations during episode-S and episode-A; Figure S5: Cloud information displayed by geostationary satellite at 5:00 UTC (13:00 LT) during episode-S and episode-A; Figure S6: Spatial correlation of surface and column HCHO. Scatterplot of HCHOgg vs. average HCHOsd within 20 km radius circles at 35 sampling sites during episode-S and episode-A; Figure S7: Spatial correlation of surface and column HCHO. Scatterplot of HCHOgg vs. HCHOsd after applying the kriging at 1 km × 1 km spatial sampling at 35 sampling sites during episode-S and episode-A; Figure S8: Sensitivity tests for different radii by spatial slopes between HCHOgg and HCHOsd. HCHOsd were obtained by averaging method; Figure S9: Spatial distribution characteristics of Level 3 FNR and increase in O3 (ΔO3) during episode-S and episode-A; Figure S10: The model-calcul17ated RIR for major O3 precursors groups at satellite overpass time (13:00 to 14:00 LT) (bar chart) during episode-S and episode-A and corresponding Level 3 FNR for the 9 sampling sites (line chart).

Author Contributions

Conceptualization, B.W., J.W. and Y.Z.; methodology, Y.Z., J.L., Q.L. and Y.L.; software, Y.Z. and J.L.; validation, Y.Z.; formal analysis, Y.Z. and D.G.; investigation, Y.Z.; resources, B.W. and J.W.; data curation, Y.Z. and X.M.; writing—original draft preparation, Y.Z. and X.M.; writing—review and editing, H.W., D.G., D.W. and X.L.; visualization, Y.Z. and X.M.; supervision, B.W. and J.W.; project administration, B.W.; funding acquisition, B.W. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the National Key R&D Program of China (2022YFC3700604), the National Natural Science Foundation of China (NSFC) Projects (42121004, 42077190, 42205105), and the Science and Technology Program of Guangzhou City (202201010400).

Data Availability Statement

The surface dataset and computer code are available to the community and can be accessed by request from Boguang Wang ([email protected]). The TROPOMI satellite data for NO2 and HCHO are available at https://s5pexp.copernicus.eu/dhus (last access: 27 May 2022).

Acknowledgments

The authors thank Hong Kong Environmental Protection Department, the AECOM company, and all graduate students in JNU-ECI for their help during the sampling campaign. Additionally, we thank ECMWF and HKUST for providing the ERA5 meteorological data and geostationary satellite image information, respectively. We thank Glenn Wolfe at NASA, U.S.A., for providing the original code for F0AM.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Olaguer, E.P.; Rappengluck, B.; Lefer, B.; Stutz, J.; Dibb, J.; Griffin, R.; Brune, W.H.; Shauck, M.; Buhr, M.; Jeffries, H.; et al. Deciphering the Role of Radical Precursors during the Second Texas Air Quality Study. J. Air Waste Manag. Assoc. 2009, 59, 1258–1277. [Google Scholar] [CrossRef]
  2. Cooke, M.C.; Utembe, S.R.; Carbajo, P.G.; Archibald, A.T.; Orr-Ewing, A.J.; Jenkin, M.E.; Derwent, R.G.; Lary, D.J.; Shallcross, D.E. Impacts of formaldehyde photolysis rates on tropospheric chemistry. Atmos. Sci. Lett. 2010, 11, 33–38. [Google Scholar] [CrossRef]
  3. Jia, C.H.; Tong, S.R.; Zhang, X.R.; Li, F.J.; Zhang, W.Q.; Li, W.R.; Wang, Z.; Zhang, G.; Tang, G.Q.; Liu, Z.R.; et al. Atmospheric oxidizing capacity in autumn Beijing: Analysis of the O-3 and PM2.5 episodes based on observation-based model. J. Environ. Sci. 2023, 124, 557–569. [Google Scholar]
  4. Lowe, D.C.; Schmidt, U. Formaldehyde (HCHO) measurements in the nonurban atmosphere. J. Geophys. Res. 1983, 88, 10844–10858. [Google Scholar] [CrossRef]
  5. Wolfe, G.M.; Kaiser, J.; Hanisco, T.F.; Keutsch, F.N.; de Gouw, J.A.; Gilman, J.B.; Graus, M.; Hatch, C.D.; Holloway, J.; Horowitz, L.W.; et al. Formaldehyde production from isoprene oxidation across NOx regimes. Atmos. Chem. Phys. 2016, 16, 2597–2610. [Google Scholar] [CrossRef] [PubMed]
  6. Abbot, D.S.; Palmer, P.I.; Martin, R.V.; Chance, K.V.; Jacob, D.J.; Guenther, A. Seasonal and interannual variability of North American isoprene emissions as determined by formaldehyde column measurements from space. Geophys. Res. Lett. 2003, 30. [Google Scholar] [CrossRef]
  7. Palmer, P.I.; Jacob, D.J.; Fiore, A.M.; Martin, R.V.; Chance, K.; Kurosu, T.P. Mapping isoprene emissions over North America using formaldehyde column observations from space. J. Geophys. Res.-Atmos. 2003, 108. [Google Scholar] [CrossRef]
  8. Millet, D.B.; Jacob, D.J.; Turquety, S.; Hudman, R.C.; Wu, S.L.; Fried, A.; Walega, J.; Heikes, B.G.; Blake, D.R.; Singh, H.B.; et al. Formaldehyde distribution over North America: Implications for satellite retrievals of formaldehyde columns and isoprene emission. J. Geophys. Res.-Atmos. 2006, 111, D24S02. [Google Scholar] [CrossRef]
  9. Zhao, T.L.; Mao, J.Q.; Simpson, W.R.; De Smedt, I.; Zhu, L.; Hanisco, T.F.; Wolfe, G.M.; St Clair, J.M.; Abad, G.G.; Nowlan, C.R.; et al. Source and variability of formaldehyde (HCHO) at northern high latitudes: An integrated satellite, aircraft, and model study. Atmos. Chem. Phys. 2022, 22, 7163–7178. [Google Scholar]
  10. Zhang, H.L.; Li, J.Y.; Ying, Q.; Guven, B.B.; Olaguer, E.P. Source apportionment of formaldehyde during TexAQS 2006 using a source-oriented chemical transport model. J. Geophys. Res. Atmos. 2013, 118, 1525–1535. [Google Scholar] [CrossRef]
  11. Su, W.J.; Liu, C.; Hu, Q.H.; Zhao, S.H.; Sun, Y.W.; Wang, W.; Zhu, Y.Z.; Liu, J.G.; Kim, J. Primary and secondary sources of ambient formaldehyde in the Yangtze River Delta based on Ozone Mapping and Profiler Suite (OMPS) observations. Atmos. Chem. Phys. 2019, 19, 6717–6736. [Google Scholar] [CrossRef]
  12. Wang, B.; Lee, S.C.; Ho, K.F. Characteristics of carbonyls: Concentrations and source strengths for indoor and outdoor residential microenvironments in China. Atmos. Environ. 2007, 41, 2851–2861. [Google Scholar] [CrossRef]
  13. Chi, Y.; Feng, Y.; Wen, S.; Lu, H.; Yu, Z.; Zhang, W.; Sheng, G.; Fu, J. Determination of carbonyl compounds in the atmosphere by DNPH derivatization and LC-ESI-MS/MS detection. Talanta 2007, 72, 539–545. [Google Scholar] [CrossRef] [PubMed]
  14. Zhang, Y.; Xue, L.; Dong, C.; Wang, T.; Mellouki, A.; Zhang, Q.; Wang, W. Gaseous carbonyls in China’s atmosphere: Tempo-spatial distributions, sources, photochemical formation, and impact on air quality. Atmos. Environ. 2019, 214, 116863. [Google Scholar]
  15. Lu, H.; Cai, Q.Y.; Wen, S.; Chi, Y.; Guo, S.; Sheng, G.; Fu, J. Seasonal and diurnal variations of carbonyl compounds in the urban atmosphere of Guangzhou, China. Sci. Total Environ. 2010, 408, 3523–3529. [Google Scholar] [CrossRef] [PubMed]
  16. Szulejko, J.E.; Kim, K.-H. Derivatization techniques for determination of carbonyls in air. TrAC Trends Anal. Chem. 2015, 64, 29–41. [Google Scholar]
  17. Chance, K.; Palmer, P.I.; Spurr, R.J.D.; Martin, R.V.; Kurosu, T.P.; Jacob, D.J. Satellite observations of formaldehyde over North America from GOME. Geophys. Res. Lett. 2000, 27, 3461–3464. [Google Scholar] [CrossRef]
  18. Veefkind, J.P.; Aben, I.; McMullan, K.; Forster, H.; de Vries, J.; Otter, G.; Claas, J.; Eskes, H.J.; de Haan, J.F.; Kleipool, Q.; et al. 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. Environ. 2012, 120, 70–83. [Google Scholar] [CrossRef]
  19. Kim, J.; Jeong, U.; Ahn, M.H.; Kim, J.H.; Park, R.J.; Lee, H.; Song, C.H.; Choi, Y.S.; Lee, K.H.; Yoo, J.M.; et al. New Era of Air Quality Monitoring from Space: Geostationary Environment Monitoring Spectrometer (GEMS). Bull. Am. Meteorol. Soc. 2020, 101, E1–E22. [Google Scholar]
  20. Zoogman, P.; Liu, X.; Suleiman, R.M.; Pennington, W.F.; Flittner, D.E.; Al-Saadi, J.A.; Hilton, B.B.; Nicks, D.K.; Newchurch, M.J.; Carr, J.L.; et al. Tropospheric emissions: Monitoring of pollution (TEMPO). J. Quant. Spectrosc. Radiat. Transf. 2017, 186, 17–39. [Google Scholar] [CrossRef]
  21. Courreges-Lacoste, G.B.; Sallusti, M.; Bulsa, G.; Bagnasco, G.; Veihelmann, B.; Riedl, S.; Smith, D.J.; Maurer, R. The Copernicus Sentinel 4 mission—A Geostationary Imaging UVN Spectrometer for Air Quality Monitoring. In Sensors, Systems, and Next-Generation Satellites XXI; SPIE Remote Sensing: Warsaw, Poland, 2017; Volume 10423. [Google Scholar]
  22. Lui, K.H.; Ho, S.S.H.; Louie, P.K.K.; Chan, C.S.; Lee, S.C.; Hu, D.; Chan, P.W.; Lee, J.C.W.; Ho, K.F. Seasonal behavior of carbonyls and source characterization of formaldehyde (HCHO) in ambient air. Atmos. Environ. 2017, 152, 51–60. [Google Scholar] [CrossRef]
  23. Ho, K.F.; Ho, S.S.H.; Huang, R.J.; Dai, W.T.; Cao, J.J.; Tian, L.; Deng, W.J. Spatiotemporal distribution of carbonyl compounds in China. Environ. Pollut. 2015, 197, 316–324. [Google Scholar] [CrossRef] [PubMed]
  24. Wang, C.; Huang, X.-F.; Han, Y.; Zhu, B.; He, L.-Y. Sources and Potential Photochemical Roles of Formaldehyde in an Urban Atmosphere in South China. J. Geophys. Res. Atmos. 2017, 122, 11–934. [Google Scholar]
  25. Yuan, B.; Chen, W.; Shao, M.; Wang, M.; Lu, S.; Wang, B.; Liu, Y.; Chang, C.-C.; Wang, B. Measurements of ambient hydrocarbons and carbonyls in the Pearl River Delta (PRD), China. Atmos. Res. 2012, 116, 93–104. [Google Scholar]
  26. Cheng, Y.; Lee, S.C.; Huang, Y.; Ho, K.F.; Ho, S.S.H.; Yau, P.S.; Louie, P.K.K.; Zhang, R.J. Diurnal and seasonal trends of carbonyl compounds in roadside, urban, and suburban environment of Hong Kong. Atmos. Environ. 2014, 89, 43–51. [Google Scholar] [CrossRef]
  27. Lü, H.; Cai, Q.-Y.; Wen, S.; Chi, Y.; Guo, S.; Sheng, G.; Fu, J.; Antizar-Ladislao, B. Carbonyl compounds in the ambient air of hazy days and clear days in Guangzhou, China. Atmos. Res. 2009, 94, 363–372. [Google Scholar] [CrossRef]
  28. Zhang, Y.; Mu, Y.; Meng, F.; Li, H.; Wang, X.; Zhang, W.; Mellouki, A.; Gao, J.; Zhang, X.; Wang, S.; et al. The pollution levels of BTEX and carbonyls under haze and non-haze days in Beijing, China. Sci. Total Environ. 2014, 490, 391–396. [Google Scholar] [CrossRef]
  29. Shim, C.; Wang, Y.H.; Choi, Y.; Palmer, P.I.; Abbot, D.S.; Chance, K. Constraining global isoprene emissions with Global Ozone Monitoring Experiment (GOME) formaldehyde column measurements. J. Geophys. Res.-Atmos. 2005, 110. [Google Scholar] [CrossRef]
  30. Fu, T.M.; Jacob, D.J.; Palmer, P.I.; Chance, K.; Wang, Y.X.X.; Barletta, B.; Blake, D.R.; Stanton, J.C.; Pilling, M.J. Space-based formaldehyde measurements as constraints on volatile organic compound emissions in east and south Asia and implications for ozone. J. Geophys. Res.-Atmos. 2007, 112. [Google Scholar] [CrossRef]
  31. Stavrakou, T.; Muller, J.F.; De Smedt, I.; Van Roozendael, M.; van der Werf, G.R.; Giglio, L.; Guenther, A. Global emissions of non-methane hydrocarbons deduced from SCIAMACHY formaldehyde columns through 2003–2006. Atmos. Chem. Phys. 2009, 9, 3663–3679. [Google Scholar] [CrossRef]
  32. Marais, E.A.; Jacob, D.J.; Kurosu, T.P.; Chance, K.; Murphy, J.G.; Reeves, C.; Mills, G.; Casadio, S.; Millet, D.B.; Barkley, M.P.; et al. Isoprene emissions in Africa inferred from OMI observations of formaldehyde columns. Atmos. Chem. Phys. 2012, 12, 6219–6235. [Google Scholar] [CrossRef]
  33. Barkley, M.P.; De Smedt, I.; Van Roozendael, M.; Kurosu, T.P.; Chance, K.; Arneth, A.; Hagberg, D.; Guenther, A.; Paulot, F.; Marais, E.; et al. Top-down isoprene emissions over tropical South America inferred from SCIAMACHY and OMI formaldehyde columns. J. Geophys. Res.-Atmos. 2013, 118, 6849–6868. [Google Scholar] [CrossRef]
  34. Zhu, L.; Jacob, D.J.; Mickley, L.J.; Marais, E.A.; Cohan, D.S.; Yoshida, Y.; Duncan, B.N.; Abad, G.G.; Chance, K.V. Anthropogenic emissions of highly reactive volatile organic compounds in eastern Texas inferred from oversampling of satellite (OMI) measurements of HCHO columns. Environ. Res. Lett. 2014, 9, 114004. [Google Scholar] [CrossRef]
  35. Bauwens, M.; Stavrakou, T.; Muller, J.F.; De Smedt, I.; Van Roozendael, M.; van der Werf, G.R.; Wiedinmyer, C.; Kaiser, J.W.; Sindelarova, K.; Guenther, A. Nine years of global hydrocarbon emissions based on source inversion of OMI formaldehyde observations. Atmos. Chem. Phys. 2016, 16, 10133–10158. [Google Scholar] [CrossRef]
  36. Kwon, H.A.; Park, R.J.; Oak, Y.J.; Nowlan, C.R.; Janz, S.J.; Kowalewski, M.G.; Fried, A.; Walega, J.; Bates, K.H.; Choi, J.; et al. Top-down estimates of anthropogenic VOC emissions in South Korea using formaldehyde vertical column densities from aircraft during the KORUS-AQ campaign. Elem.-Sci. Anthr. 2021, 9, 00109. [Google Scholar] [CrossRef]
  37. Wang, Y.X.; Lin, N.; Li, W.; Guenther, A.; Lam, J.C.Y.; Tai, A.P.K.; Potosnak, M.J.; Seco, R. Satellite-derived constraints on the effect of drought stress on biogenic isoprene emissions in the southeastern US. Atmos. Chem. Phys. 2022, 22, 14189–14208. [Google Scholar] [CrossRef]
  38. Mendoza-Rodriguez, C.A.; Cardenas, C.R.; Espinosa-Ponce, C. Long term (2005–2016) study of formaldehyde (HCHO) columns from satellite data in two regions in the south of Mexico. Evidence of the impact of agricultural activity. Remote Sens. Appl.-Soc. Environ. 2023, 29, 100894. [Google Scholar] [CrossRef]
  39. Liao, J.; Hanisco, T.F.; Wolfe, G.M.; St Clair, J.; Jimenez, J.L.; Campuzano-Jost, P.; Nault, B.A.; Fried, A.; Marais, E.A.; Abad, G.G.; et al. Towards a satellite formaldehyde—In Situ hybrid estimate for organic aerosol abundance. Atmos. Chem. Phys. 2019, 19, 2765–2785. [Google Scholar]
  40. Wolfe, G.M.; Nicely, J.M.; Clair, J.M.S.; Hanisco, T.F.; Liao, J.; Oman, L.D.; Brune, W.B.; Miller, D.; Thames, A.; Abad, G.G.; et al. Mapping hydroxyl variability throughout the global remote troposphere via synthesis of airborne and satellite formaldehyde observations. Proc. Natl. Acad. Sci. USA 2019, 116, 11171–11180. [Google Scholar] [CrossRef] [PubMed]
  41. Zhu, Q.D.; Laughner, J.L.; Cohen, R.C. Combining Machine Learning and Satellite Observations to Predict Spatial and Temporal Variation of near Surface OH in North American Cities. Environ. Sci. Technol. 2022, 56, 7362–7371. [Google Scholar] [CrossRef]
  42. Zhang, Y.Z.; Wang, Y.H.; Crawford, J.; Cheng, Y.; Li, J.F. Improve observation-based ground-level ozone spatial distribution by compositing satellite and surface observations: A simulation experiment. Atmos. Environ. 2018, 180, 226–233. [Google Scholar]
  43. Cheng, Y.; Wang, Y.H.; Zhang, Y.Z.; Crawford, J.H.; Diskin, G.S.; Weinheimer, A.J.; Fried, A. Estimator of Surface Ozone Using Formaldehyde and Carbon Monoxide Concentrations Over the Eastern United States in Summer. J. Geophys. Res.-Atmos. 2018, 123, 7642–7655. [Google Scholar] [CrossRef]
  44. Duncan, B.N.; Yoshida, Y.; Olson, J.R.; Sillman, S.; Martin, R.V.; Lamsal, L.; Hu, Y.T.; Pickering, K.E.; Retscher, C.; Allen, D.J.; et al. Application of OMI observations to a space-based indicator of NOx and VOC controls on surface ozone formation. Atmos. Environ. 2010, 44, 2213–2223. [Google Scholar] [CrossRef]
  45. Jin, X.M.; Fiore, A.M.; Murray, L.T.; Valin, L.C.; Lamsal, L.N.; Duncan, B.; Folkert Boersma, K.; De Smedt, I.; Abad, G.G.; Chance, K.; et al. Evaluating a Space-Based Indicator of Surface Ozone-NOx-VOC Sensitivity over Midlatitude Source Regions and Application to Decadal Trends. J. Geophys. Res.-Atmos. 2017, 122, 10231–10253. [Google Scholar]
  46. Jin, X.M.; Fiore, A.; Boersma, K.F.; De Smedt, I.; Valin, L. Inferring Changes in Summertime Surface Ozone-NOx-VOC Chemistry over US Urban Areas from Two Decades of Satellite and Ground-Based Observations. Environ. Sci. Technol. 2020, 54, 6518–6529. [Google Scholar] [CrossRef]
  47. Zhu, L.; Jacob, D.J.; Keutsch, F.N.; Mickley, L.J.; Scheffe, R.; Strum, M.; Abad, G.G.; Chance, K.; Yang, K.; Rappengluck, B.; et al. Formaldehyde (HCHO) As a Hazardous Air Pollutant: Mapping Surface Air Concentrations from Satellite and Inferring Cancer Risks in the United States. Environ. Sci. Technol. 2017, 51, 5650–5657. [Google Scholar] [CrossRef] [PubMed]
  48. Su, W.J.; Hu, Q.H.; Chen, Y.J.; Lin, J.A.; Zhang, C.X.; Liu, C. Inferring global surface HCHO concentrations from multisource hyperspectral satellites and their application to HCHO-related global cancer burden estimation. Environ. Int. 2022, 170, 107600. [Google Scholar] [CrossRef] [PubMed]
  49. Kleinman, L.I. The dependence of tropospheric ozone production rate on ozone precursors. Atmos. Environ. 2005, 39, 575–586. [Google Scholar] [CrossRef]
  50. Martin, R.V.; Parrish, D.D.; Ryerson, T.B.; Nicks, D.K.; Chance, K.; Kurosu, T.P.; Jacob, D.J.; Sturges, E.D.; Fried, A.; Wert, B.P. Evaluation of GOME satellite measurements of tropospheric NO2 and HCHO using regional data from aircraft campaigns in the southeastern United States. J. Geophys. Res.-Atmos. 2004, 109. [Google Scholar] [CrossRef]
  51. Lamsal, L.N.; Duncan, B.N.; Yoshida, Y.; Krotkov, N.A.; Pickering, K.E.; Streets, D.G.; Lu, Z.F. U.S. NO2 trends (2005–2013): EPA Air Quality System (AQS) data versus improved observations from the Ozone Monitoring Instrument (OMI). Atmos. Environ. 2015, 110, 130–143. [Google Scholar]
  52. Ma, J.Z.; Beirle, S.; Jin, J.L.; Shaiganfar, R.; Yan, P.; Wagner, T. Tropospheric NO2 vertical column densities over Beijing: Results of the first three years of ground-based MAX-DOAS measurements (2008–2011) and satellite validation. Atmos. Chem. Phys. 2013, 13, 1547–1567. [Google Scholar]
  53. Kramer, L.J.; Leigh, R.J.; Remedios, J.J.; Monks, P.S. Comparison of OMI and ground-based in situ and MAX-DOAS measurements of tropospheric nitrogen dioxide in an urban area. J. Geophys. Res.-Atmos. 2008, 113. [Google Scholar] [CrossRef]
  54. Cersosimo, A.; Serio, C.; Masiello, G. TROPOMI NO(2)Tropospheric Column Data: Regridding to 1 km Grid-Resolution and Assessment of their Consistency with In Situ Surface Observations. Remote Sens. 2020, 12, 2212. [Google Scholar]
  55. De Smedt, I.; Stavrakou, T.; Hendrick, F.; Danckaert, T.; Vlemmix, T.; Pinardi, G.; Theys, N.; Lerot, C.; Gielen, C.; Vigouroux, C.; et al. Diurnal, seasonal and long-term variations of global formaldehyde columns inferred from combined OMI and GOME-2 observations. Atmos. Chem. Phys. 2015, 15, 12519–12545. [Google Scholar]
  56. Wang, Y.P.; Wang, Z.F.; Yu, C.; Zhu, S.Y.; Cheng, L.X.; Zhang, Y.; Chen, L.F. Validation of OMI HCHO Products Using MAX-DOAS observations from 2010 to 2016 in Xianghe, Beijing: Investigation of the Effects of Aerosols on Satellite Products. Remote Sens. 2019, 11, 203. [Google Scholar]
  57. Wang, P.D.; Holloway, T.; Bindl, M.; Harkey, M.; De Smedt, I. Ambient Formaldehyde over the United States from Ground-Based (AQS) and Satellite (OMI) Observations. Remote Sens. 2022, 14, 2191. [Google Scholar] [CrossRef]
  58. Yang, X.; Xue, L.; Wang, T.; Wang, X.; Gao, J.; Lee, S.; Blake, D.R.; Chai, F.; Wang, W. Observations and Explicit Modeling of Summertime Carbonyl Formation in Beijing: Identification of Key Precursor Species and Their Impact on Atmospheric Oxidation Chemistry. J. Geophys. Res. Atmos. 2018, 123, 1426–1440. [Google Scholar]
  59. Yang, X.; Xue, L.; Yao, L.; Li, Q.; Wen, L.; Zhu, Y.; Chen, T.; Wang, X.; Yang, L.; Wang, T.; et al. Carbonyl compounds at Mount Tai in the North China Plain: Characteristics, sources, and effects on ozone formation. Atmos. Res. 2017, 196, 53–61. [Google Scholar]
  60. Kanaya, Y.; Pochanart, P.; Liu, Y.; Li, J.; Tanimoto, H.; Kato, S.; Suthawaree, J.; Inomata, S.; Taketani, F.; Okuzawa, K.; et al. Rates and regimes of photochemical ozone production over Central East China in June 2006: A box model analysis using comprehensive measurements of ozone precursors. Atmos. Chem. Phys. 2009, 9, 7711–7723. [Google Scholar]
  61. Huang, J.; Feng, Y.; Li, J.; Xiong, B.; Feng, J.; Wen, S.; Sheng, G.; Fu, J.; Wu, M. Characteristics of carbonyl compounds in ambient air of Shanghai, China. J. Atmos. Chem. 2009, 61, 1–20. [Google Scholar]
  62. Shen, L.; Jacob, D.J.; Zhu, L.; Zhang, Q.; Zheng, B.; Sulprizio, M.P.; Li, K.; De Smedt, I.; Abad, G.G.; Cao, H.S.; et al. 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. 2019, 46, 4468–4475. [Google Scholar]
  63. Wang, H.; Wu, Q.Z.; Guenther, A.B.; Yang, X.C.; Wang, L.N.; Xiao, T.; Li, J.; Feng, J.M.; Xu, Q.; Cheng, H.Q. A long-term estimation of biogenic volatile organic compound (BVOC) emission in China from 2001–2016: The roles of land cover change and climate variability. Atmos. Chem. Phys. 2021, 21, 4825–4848. [Google Scholar]
  64. Pu, D.C.; Zhu, L.; De Smedt, I.; Li, X.C.; Sun, W.F.; Wang, D.K.; Liu, S.; Li, J.; Shu, L.; Chen, Y.Y.; et al. Response of Anthropogenic Volatile Organic Compound Emissions to Urbanization in Asia Probed with TROPOMI and VIIRS Satellite Observations. Geophys. Res. Lett. 2022, 49, 99470. [Google Scholar] [CrossRef]
  65. Zhang, X.Y.; Zhao, L.M.; Cheng, M.M.; Chen, D.M. Estimating Ground-Level Ozone Concentrations in Eastern China Using Satellite-Based Precursors. IEEE Trans. Geosci. Remote Sens. 2020, 58, 4754–4763. [Google Scholar] [CrossRef]
  66. Jin, X.M.; Holloway, T. Spatial and temporal variability of ozone sensitivity over China observed from the Ozone Monitoring Instrument. J. Geophys. Res.-Atmos. 2015, 120, 7229–7246. [Google Scholar]
  67. Wang, W.N.; Ronald, V.; Ding, J.Y.; van Weele, M.; Cheng, T.H. Spatial and temporal changes of the ozone sensitivity in China based on satellite and ground-based observations. Atmos. Chem. Phys. 2021, 21, 7253–7269. [Google Scholar] [CrossRef]
  68. Ren, J.; Guo, F.F.; Xie, S.D. Diagnosing ozone-NOx-VOC sensitivity and revealing causes of ozoneincreases in China based on 2013–2021 satellite retrievals. Atmos. Chem. Phys. 2022, 22, 15035–15047. [Google Scholar] [CrossRef]
  69. Zheng, J.Y.; Zhong, L.J.; Wang, T.; Louie, P.K.K.; Li, Z.C. Ground-level ozone in the Pearl River Delta region: Analysis of data from a recently established regional air quality monitoring network. Atmos. Environ. 2010, 44, 814–823. [Google Scholar] [CrossRef]
  70. Liu, X.F.; Wang, N.; Lyu, X.P.; Zeren, Y.Z.; Jiang, F.; Wang, X.M.; Zou, S.C.; Ling, Z.H.; Guo, H. Photochemistry of ozone pollution in autumn in Pearl River Estuary, South China. Sci. Total Environ. 2021, 754, 141812. [Google Scholar] [CrossRef]
  71. Zhao, M.; Zhang, Y.N.; Pei, C.L.; Chen, T.S.; Mu, J.S.; Liu, Y.H.; Wang, Y.J.; Wang, W.X.; Xue, L.K. Worsening ozone air pollution with reduced NOx and VOCs in the Pearl River Delta region in autumn 2019: Implications for national control policy in China. J. Environ. Manag. 2022, 324, 116327. [Google Scholar] [CrossRef]
  72. Wang, X.Q.; Zhang, T.S.; Pei, C.L.; Chen, D.H.; Lu, L.H.; Xiang, Y. Monitoring of Vertical Distribution of Ozone Using Differential Absorption Lidar in Guangzhou. Chin. J. Lasers-Zhongguo Jiguang 2019, 46, 20220579. [Google Scholar]
  73. Hong, Y.; Chen, C.; Bao, H.; Shen, J. Sources and Sensitivity Analysis of Ozone in Spring over the Southwestern Part of Pearl River Delta Region. Ecol. Environ. Sci. 2021, 30, 984–994. [Google Scholar]
  74. Fan, J.; Ju, T.; Wang, Q.; Gao, H.; Huang, R.; Duan, J. Spatiotemporal variations and potential sources of tropospheric formaldehyde over eastern China based on OMI satellite data. Atmos. Pollut. Res. 2021, 12, 272–285. [Google Scholar]
  75. Louie, P.K.K.; Ho, J.W.K.; Tsang, R.C.W.; Blake, D.R.; Lau, A.K.H.; Yu, J.Z.; Yuan, Z.; Wang, X.; Shao, M.; Zhong, L. VOCs and OVOCs distribution and control policy implications in Pearl River Delta region, China. Atmos. Environ. 2013, 76, 125–135. [Google Scholar]
  76. Wang, S.; Zheng, J.; Fu, F.; Yin, S.; Zhong, L. Development of an emission processing system for the Pearl River Delta Regional air quality modeling using the SMOKE model: Methodology and evaluation. Atmos. Environ. 2011, 45, 5079–5089. [Google Scholar]
  77. Wan, Z.; Ji, S.; Liu, Y.; Zhang, Q.; Chen, J.; Wang, Q. Shipping emission inventories in China’s Bohai Bay, Yangtze River Delta, and Pearl River Delta in 2018. Mar. Pollut. Bull. 2020, 151, 110882. [Google Scholar] [CrossRef]
  78. Wu, Z.; Zhang, Y.; He, J.; Chen, H.; Huang, X.; Wang, Y.; Yu, X.; Yang, W.; Zhang, R.; Zhu, M.; et al. Dramatic increase in reactive volatile organic compound (VOC) emissions from ships at berth after implementing the fuel switch policy in the Pearl River Delta Emission Control Area. Atmos. Chem. Phys. 2020, 20, 1887–1900. [Google Scholar]
  79. Zeren, Y.; Guo, H.; Lyu, X.; Jiang, F.; Wang, Y.; Liu, X.; Zeng, L.; Li, M.; Li, L. An Ozone “Pool” in South China: Investigations on Atmospheric Dynamics and Photochemical Processes Over the Pearl River Estuary. J. Geophys. Res. Atmos. 2019, 124, 12340–12355. [Google Scholar]
  80. Zheng, J.; Zhang, L.; Che, W.; Zheng, Z.; Yin, S. A highly resolved temporal and spatial air pollutant emission inventory for the Pearl River Delta region, China and its uncertainty assessment. Atmos. Environ. 2009, 43, 5112–5122. [Google Scholar]
  81. Mo, X.J.; Gong, D.C.; Liu, Y.F.; Li, J.Y.; Zhao, Y.M.; Zhao, W.L.; Shen, J.; Liao, T.; Wang, H.; Wang, B.G. Ground-based formaldehyde across the Pearl River Delta: A snapshot and meta-analysis study. Atmos. Environ. 2023, 43, 1352–2310. [Google Scholar]
  82. De Smedt, I.; Pinardi, G.; Vigouroux, C.; Compernolle, S.; Bais, A.; Benavent, N.; Boersma, F.; Chan, K.L.; Donner, S.; Eichmann, K.U.; et al. Comparative assessment of TROPOMI and OMI formaldehyde observations and validation against MAX-DOAS network column measurements. Atmos. Chem. Phys. 2021, 21, 12561–12593. [Google Scholar]
  83. Franco, B.; Marais, E.A.; Bovy, B.; Bader, W.; Lejeune, B.; Roland, G.; Servais, C.; Mahieu, E. Diurnal cycle and multi-decadal trend of formaldehyde in the remote atmosphere near 46A degrees aEuro N. Atmos. Chem. Phys. 2016, 16, 4171–4189. [Google Scholar] [CrossRef]
  84. Sun, K.; Zhu, L.; Cady-Pereira, K.; Miller, C.C.; Chance, K.; Clarisse, L.; Coheur, P.F.; Abad, G.G.; Huang, G.Y.; Liu, X.; et al. A physics-based approach to oversample multi-satellite, multispecies observations to a common grid. Atmos. Meas. Tech. 2018, 11, 6679–6701. [Google Scholar] [CrossRef]
  85. Jenkin, M.E.; Young, J.C.; Rickard, A.R. The MCM v3.3.1 degradation scheme for isoprene. Atmos. Chem. Phys. 2015, 309, 11433–11459. [Google Scholar] [CrossRef]
  86. Wang, Y.; Liu, T.; Gong, D.C.; Wang, H.; Guo, H.; Liao, M.P.; Deng, S.; Cai, H.; Wang, B.G. Anthropogenic Pollutants Induce Changes in Peroxyacetyl Nitrate Formation Intensity and Pathways in a Mountainous Background Atmosphere in Southern China. Environ. Sci. Technol. 2023, 57, 6253–6262. [Google Scholar] [CrossRef]
  87. Saunders, S.M.; Jenkin, M.E.; Derwent, R.G.; Pilling, M.J. Protocol for the development of the Master Chemical Mechanism, MCM v3 (Part A): Tropospheric degradation of non-aromatic volatile organic compounds. Atmos. Chem. Phys. 2003, 3, 161–180. [Google Scholar]
  88. Xue, L.K.; Wang, T.; Gao, J.; Ding, A.J.; Zhou, X.H.; Blake, D.R.; Wang, X.F.; Saunders, S.M.; Fan, S.J.; Zuo, H.C.; et al. Ground-level ozone in four Chinese cities: Precursors, regional transport and heterogeneous processes. Atmos. Chem. Phys. 2014, 14, 13175–13188. [Google Scholar]
  89. Chen, T.S.; Xue, L.K.; Zheng, P.G.; Zhang, Y.N.; Liu, Y.H.; Sun, J.J.; Han, G.X.; Li, H.Y.; Zhang, X.; Li, Y.F.; et al. Volatile organic compounds and ozone air pollution in an oil production region in northern China. Atmos. Chem. Phys. 2020, 20, 7069–7086. [Google Scholar]
  90. Liu, T.T.; Hong, Y.W.; Li, M.R.; Xu, L.L.; Chen, J.S.; Bian, Y.H.; Yang, C.; Dan, Y.B.; Zhang, Y.N.; Xue, L.K.; et al. Atmospheric oxidation capacity and ozone pollution mechanism in a coastal city of southeastern China: Analysis of a typical photochemical episode by an observation-based model. Atmos. Chem. Phys. 2022, 22, 2173–2190. [Google Scholar]
  91. Wang, J.; Zhang, Y.; Wu, Z.; Luo, S.; Song, W.; Wang, X. Ozone episodes during and after the 2018 Chinese National Day holidays in Guangzhou: Implications for the control of precursor VOCs. J. Environ. Sci. 2022, 114, 322–333. [Google Scholar]
  92. Lin, Y.L.; Hong, Y.W.; Ji, X.T.; Xu, K.; Shao, Z.Q.; Yu, R.L.; Chen, J.S. Pollution Characteristics of Atmospheric Formaldehyde (HCHO) and its Environmental Effects in Autumn in a Coastal City. China Environ. Sci. 2022, 43, 52–60. [Google Scholar]
  93. GB 3095-2012; Ambient Air Quality Standards. Ministry of Environmental Protection of the People’s Republic of China: Beijing, China, 2012. (In Chinese)
  94. Hong, Q.Q.; Zhu, L.B.; Xing, C.Z.; Hu, Q.H.; Lin, H.; Zhang, C.X.; Zhao, C.H.; Liu, T.; Su, W.J.; Liu, C. Inferring vertical variability and diurnal evolution of O-3 formation sensitivity based on the vertical distribution of summertime HCHO and NO2 in Guangzhou, China. Sci. Total Environ. 2022, 827, 154045. [Google Scholar] [CrossRef] [PubMed]
  95. Wang, Y.; Shen, J.; Wang, H.; Wu, G.C.; Chen, Y.Q.; Liu, T.; Gong, D.C.; Ou, J.; Shi, Y.K.; Zhang, T.; et al. Unexpected seasonal variations and high levels of ozone observed at the summit of Nanling Mountains: Impact of Asian monsoon on southern China. Atmos. Environ. 2021, 253, 118378. [Google Scholar]
  96. Kuhlmann, G.; Hartl, A.; Cheung, H.M.; Lam, Y.F.; Wenig, M.O. A novel gridding algorithm to create regional trace gas maps from satellite observations. Atmos. Meas. Tech. 2014, 7, 451–467. [Google Scholar]
  97. Levelt, P.F.; Van den Oord, G.H.J.; Dobber, M.R.; Malkki, A.; Visser, H.; de Vries, J.; Stammes, P.; Lundell, J.O.V.; Saari, H. The Ozone Monitoring Instrument. IEEE Trans. Geosci. Remote Sens. 2006, 44, 1093–1101. [Google Scholar] [CrossRef]
  98. Chen, Z.Y.; Zhang, R.; Zhang, T.H.; Ou, C.Q.; Guo, Y.M. A kriging-calibrated machine learning method for estimating daily ground-level NO2 in mainland China. Sci. Total Environ. 2019, 690, 556–564. [Google Scholar] [CrossRef] [PubMed]
  99. Guan, J.; Jin, B.H.; Ding, Y.Z.; Wang, W.; Li, G.X.; Ciren, P. Global Surface HCHO Distribution Derived from Satellite Observations with Neural Networks Technique. Remote Sens. 2021, 13, 4055. [Google Scholar] [CrossRef]
  100. Liu, T.T.; Hu, B.Y.; Xu, X.B.; Hong, Y.W.; Zhang, Y.R.; Wu, X.; Xu, L.L.; Li, M.R.; Chen, Y.T.; Chen, X.Q.; et al. Characteristics of PM2.5-bound secondary organic aerosol tracers in a coastal city in Southeastern China: Seasonal patterns and pollution identification. Atmos. Environ. 2020, 237, 117710. [Google Scholar]
  101. Lin, H.T.; Wang, M.; Duan, Y.S.; Fu, Q.Y.; Ji, W.H.; Cui, H.X.; Jin, D.; Lin, Y.F.; Hu, K. O-3 Sensitivity and Contributions of Different NMHC Sources in O-3 Formation at Urban and Suburban Sites in Shanghai. Atmosphere 2020, 11, 295. [Google Scholar] [CrossRef]
Remotesensing 15 03998 g001
Figure 1. Average HCHO vertical column density from TROPOMI in 2021. GBA is designated as three regions for analysis. The grid sampling sites of HCHO are labeled as circles. The general monitoring stations are labeled as black points. The VOC monitoring stations, which are used for O3 sensitivity analysis, are labeled as triangles.
Figure 1. Average HCHO vertical column density from TROPOMI in 2021. GBA is designated as three regions for analysis. The grid sampling sites of HCHO are labeled as circles. The general monitoring stations are labeled as black points. The VOC monitoring stations, which are used for O3 sensitivity analysis, are labeled as triangles.
Remotesensing 15 03998 g001
Remotesensing 15 03998 g002
Figure 2. Spatial distribution characteristics of ground-based HCHO, CO, O3 and NO2 during the morning and early afternoon of two O3 episodes. Episode-S results are shown on the left (a–h) and episode-A results are shown on the right (i–p).
Figure 2. Spatial distribution characteristics of ground-based HCHO, CO, O3 and NO2 during the morning and early afternoon of two O3 episodes. Episode-S results are shown on the left (a–h) and episode-A results are shown on the right (i–p).
Remotesensing 15 03998 g002
Remotesensing 15 03998 g003
Figure 3. Spatial distribution characteristics of HCHOsd during (a) episode-S and (b) episode-A.
Figure 3. Spatial distribution characteristics of HCHOsd during (a) episode-S and (b) episode-A.
Remotesensing 15 03998 g003
Remotesensing 15 03998 g004
Figure 4. HCHOgg and HCHOsd from different regions during (a) episode-S and (b) episode-A. The HCHOsd are the average values of pixels in the grids corresponding to the sampling sites in the WGBA, PRE, and EGBA.
Figure 4. HCHOgg and HCHOsd from different regions during (a) episode-S and (b) episode-A. The HCHOsd are the average values of pixels in the grids corresponding to the sampling sites in the WGBA, PRE, and EGBA.
Remotesensing 15 03998 g004
Remotesensing 15 03998 g005
Figure 5. Spatial correlation of surface and column HCHO. Scatterplot of HCHOgg vs. overlapping TROPOMI pixels at 35 sampling sites during (a) episode-S and (b) episode-A. Scatterplot of HCHOgg vs. average HCHOsd at 33 sampling grids during (c) episode-S and (d) episode-A. Spatial correlation coefficient r values and numbers of the available data are shown in the legend.
Figure 5. Spatial correlation of surface and column HCHO. Scatterplot of HCHOgg vs. overlapping TROPOMI pixels at 35 sampling sites during (a) episode-S and (b) episode-A. Scatterplot of HCHOgg vs. average HCHOsd at 33 sampling grids during (c) episode-S and (d) episode-A. Spatial correlation coefficient r values and numbers of the available data are shown in the legend.
Remotesensing 15 03998 g005
Remotesensing 15 03998 g006
Figure 6. Sensitivity tests for different radii by spatial correlation coefficients between HCHOgg and HCHOsd. HCHOsd were obtained by (a) direct averaging method and (b) averaging after kriging method.
Figure 6. Sensitivity tests for different radii by spatial correlation coefficients between HCHOgg and HCHOsd. HCHOsd were obtained by (a) direct averaging method and (b) averaging after kriging method.
Remotesensing 15 03998 g006
Remotesensing 15 03998 g007
Figure 7. Spatial distribution characteristics of HCHOsd improved from TROPOMI Level 2 during (a) episode-S and (b) episode-A to Level 3 during (c) episode-S and (d) episode-A. Level 2 was re-gridded from the TROPOMI spatial resolution to a consistent resolution of 5.5 km × 5.5 km.
Figure 7. Spatial distribution characteristics of HCHOsd improved from TROPOMI Level 2 during (a) episode-S and (b) episode-A to Level 3 during (c) episode-S and (d) episode-A. Level 2 was re-gridded from the TROPOMI spatial resolution to a consistent resolution of 5.5 km × 5.5 km.
Remotesensing 15 03998 g007
Remotesensing 15 03998 g008
Figure 8. Spatial correlation of surface and column HCHO. Scatterplot of HCHOgg vs. Level 3 HCHOsd during (a) episode-S and (b) episode-A. Spatial correlation coefficient r values (R) and the amount of available data (N) during the sampling periods are shown in the legend.
Figure 8. Spatial correlation of surface and column HCHO. Scatterplot of HCHOgg vs. Level 3 HCHOsd during (a) episode-S and (b) episode-A. Spatial correlation coefficient r values (R) and the amount of available data (N) during the sampling periods are shown in the legend.
Remotesensing 15 03998 g008
Remotesensing 15 03998 g009
Figure 9. The model-calculated RIR for major O3 precursors groups in the 9 VOC monitoring stations at satellite overpass time (13:00 to 14:00 LT) (bar chart) during episode-S and corresponding average FNR at the same grids (line chart).
Figure 9. The model-calculated RIR for major O3 precursors groups in the 9 VOC monitoring stations at satellite overpass time (13:00 to 14:00 LT) (bar chart) during episode-S and corresponding average FNR at the same grids (line chart).
Remotesensing 15 03998 g009
Table 1. Summary of spatial correlation coefficient r values (R) and number of available data (N) when using different cloud filters.
Table 1. Summary of spatial correlation coefficient r values (R) and number of available data (N) when using different cloud filters.
Episode-SEpisode-A
CRFNumber (N)Correlation (R)Number (N)Correlation (R)
<0.40350.82330.50
<0.45350.82330.56
<0.50350.82350.56
<0.55350.83350.56
<0.60350.82350.44
Table 2. Summary of Spearman correlation coefficient between FNR and RIR-NOx. FNR values were obtained by different methods (Level 2: overlapping TROPOMI pixels, Grids: averaging Level 2 data within 20 km × 20 km, 10–50 km: averaging Level 2 data within 10–50 km radius circles, Level 3: our new procedure).
Table 2. Summary of Spearman correlation coefficient between FNR and RIR-NOx. FNR values were obtained by different methods (Level 2: overlapping TROPOMI pixels, Grids: averaging Level 2 data within 20 km × 20 km, 10–50 km: averaging Level 2 data within 10–50 km radius circles, Level 3: our new procedure).
Episode-SEpisode-A
NumberCorrelationp-ValueNumber Correlation p-Value
Level 270.8570.01470.4290.337
Grids90.8670.00290.4330.244
10 km90.5000.17080.5950.120
20 km90.7500.02090.6330.067
30 km90.9000.00190.6330.067
40 km90.7830.01390.5000.170
50 km90.6500.05890.2330.546
Level 390.7500.02070.8570.014
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Zhao, Y.; Mo, X.; Wang, H.; Li, J.; Gong, D.; Wang, D.; Li, Q.; Liu, Y.; Liu, X.; Wang, J.; et al. A Comparative Study of Ground-Gridded and Satellite-Derived Formaldehyde during Ozone Episodes in the Chinese Greater Bay Area. Remote Sens. 2023, 15, 3998. https://doi.org/10.3390/rs15163998

AMA Style

Zhao Y, Mo X, Wang H, Li J, Gong D, Wang D, Li Q, Liu Y, Liu X, Wang J, et al. A Comparative Study of Ground-Gridded and Satellite-Derived Formaldehyde during Ozone Episodes in the Chinese Greater Bay Area. Remote Sensing. 2023; 15(16):3998. https://doi.org/10.3390/rs15163998

Chicago/Turabian Style

Zhao, Yiming, Xujun Mo, Hao Wang, Jiangyong Li, Daocheng Gong, Dakang Wang, Qinqin Li, Yunfeng Liu, Xiaoting Liu, Jinnian Wang, and et al. 2023. "A Comparative Study of Ground-Gridded and Satellite-Derived Formaldehyde during Ozone Episodes in the Chinese Greater Bay Area" Remote Sensing 15, no. 16: 3998. https://doi.org/10.3390/rs15163998

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop