A case study of heavy PM2.5 secondary formation by N2O5 nocturnal chemistry in Seoul, Korea in January 2018: Model performance and error analysis
Introduction
Heavy pollution of PM2.5 (Particulate Matter with the aerodynamic diameter of less than 2.5 μm), particularly secondary PM2.5 is one of the most urgent societal issues in Northeast Asia, and national measures to improve PM2.5 air quality have been implemented in South Korea (Kim et al., 2017a; Kim et al., 2017b). The South Korean government has established a new standard crisis management manual for PM2.5, comprising four different levels (attention, caution, alert, and serious) of alert standards and an associated response system in the event of heavy PM2.5, defined as daily average concentrations consecutively exceeding 50 μg/m3 on multiple days. Because atmospheric processes relevant to heavy PM2.5 are highly complex in urban areas (due to meteorological/chemical uncertainties, urban morphology, etc.), a basic understanding of secondary particle formation has become more important (Lee et al., 2019) and is still a challenge for reliable operational PM2.5 air quality forecasting.
As themain components of air quality forecasting system, meteorology and chemistry are two major factors in determining the PM2.5 concentrations. These two processes are interrelated and sometimes act in a compensatory direction in the comprehensive regional air quality model structure, and thus, numerical sensitivity simulations for reliable predictions are prerequisite. Numerous numerical sensitivity results documented by the National Institute of Environmental Research (NIER) based on failed PM2.5 forecasts have shown that the cause of the forecast uncertainty originating from meteorological variables over the 5 years (2015–2019) could be attributed to the wind speed overpredictions (of more than 50%), as well as other factors, such as planetary boundary layer height, wind direction and temperature (NIER, 2018). In particular, wind speed overprediction has frequently occurred in long-range transport cases, and it also becomes important in in urban-scale air quality predictions for high concentrations in stagnant atmospheric conditions in the Seoul Metropolitan Area (SMA), Korea (Park and Kim, 1999; Kim and Ghim, 2002; Park et al., 2004).
The planetary boundary layer (PBL) refers to the lower atmospheric layer, with a depth that is generally less than 2 km, in which human activities take place. PBL has been reported to play an important role in the vertical distribution of air pollutants (Stull, 1988). Previous studies have shown the differences and inconsistencies originating from the PBL parameterization scheme used in modeling studies (Madala et al., 2015; Kim et al., 2015; Banks and Baldasano, 2016; Mohan and Gupta, 2018; Sarkar et al., 2019; Yang et al., 2021). Moreover, the importance and improvement of PBL simulations by turbulence have been highlighted for air pollution modeling in numerous previous studies (Madala et al., 2015; Kim et al., 2015; Mohan and Gupta, 2018; Yang et al., 2021). For example, Kim et al. (2015) studied the sensitivities of the vertical dispersion of pollutants to different PBL schemes using offline meteorology (WRF) and chemistry-transport (Polair3D/Polyphemus) models, and showed that they influence the PM vertical distributions, not only because they influence vertical mixing (PBL height and eddy diffusion coefficient), but also the horizontal wind fields and humidity. Mohan and Gupta (2018) found that PBL parameterization schemes can also have a significant impact on exploring the physical mechanism of the pollution process and predicting the dynamic variations of pollutants, while PBL-PM2.5 coupled studies were also performed to examine PBL sensitivity (Su et al., 2018; Lee et al., 2019a; Lee et al., 2019b; Li et al., 2021).
On the other hand, the analysis of errors originating from chemical mechanisms is also important. For example, secondary nitrate (NO3−) is an important chemical component of PM2.5, and is recognized to be one of the most highly uncertain factors in predicting PM2.5. In the SMA, Korea, the mean concentrations of nitrate have generally been higher than those of sulfate (SO42−) in recent years (NIER, 2017; Jo et al., 2020; Kim et al., 2021), while secondary inorganic aerosol (SIA) species are becoming dominant in PM2.5 (Pathak et al., 2009; Khan et al., 2010; Squizzato et al., 2012; Shin et al., 2016; Seo et al., 2017). Recently, PM2.5 forecasting has also explored considerable uncertainties in the nighttime heterogeneous dinitrogen pentoxide (N2O5) chemistry (Prabhakar et al., 2017).
Nevertheless, the error quantifications of general biases induced by chemical processes (i.e., secondary organic/inorganic formation process) are yet to be assessed and remain highly uncertain in PM2.5 forecasting. In this context, sensitivity analysis of inorganic species, together with observational studies to validate the simulation, are needed to improve the nitrate formation mechanism over Northeast Asia for the improvement of PM2.5 predictions. This is because nitrate is a major component in urban areas, and inaccurate representation of nitrate PM2.5 formation chemistry directly results in severely failed PM2.5 forecasts.
Nitrate aerosols are formed mainly by two atmospheric pathways: (1) the reaction of OH with NO2 in the daytime and (2) the N2O5 heterogeneous hydrolysis during nighttime (Finlayson-Pitts and Pitts Jr., 1997; Ravishankara, 1997; Finlayson-Pitts and Pitts Jr., 2000; Brown et al., 2006a; Brown et al., 2006b). The product of these two (1) and (2) reactions, HNO3, is a limiting reagent and/or will thermodynamically partition to the aerosol phase, depending on the ammonia or other inorganic gaseous species (Franchin et al., 2018; Ibikunle et al., 2020; Nenes et al., 2020).
Many previous studies have pointed out that nitrate formation via N2O5 heterogeneous hydrolysis is important in producing high PM2.5 concentrations, especially during winter, because of the longer nighttime length; thus, the N2O5 uptake coefficient (γN2O5) is an uncertain, but important parameter in N2O5 heterogeneous hydrolysis (Brown et al., 2006a; Baasandorj et al., 2017; Wang et al., 2018). Nevertheless, the extremely high PM2.5 (i.e., up to the alert or serious levels of predictions) induced purely by nighttime N2O5 formation in urban areas, did not frequently occur in SMA in Korea. However, the heavy PM2.5 wintertime episode of January 13–15, 2018, was found to be mainly induced by nighttime N2O5 heterogeneous reaction, inferring from the modeling-based estimation of reaction rates for HNO3 formation and the relevant measurement (NIER, 2018). This made it possible to assess quantitatively the operational PM2.5 predictions system capabilities and uncertainties on nitrate formation originating from the N2O5 uptake process in SMA.
In this study, we carried out a series of numerical simulations to (1) confirm the importance of the N2O5 heterogeneous hydrolysis process in the SMA during nighttime and (2) compare the uncertainty ranges originating from two factors: the N2O5 uptake process as an uncertain factor in the chemistry, and the parameterization of PBL height as an uncertain factor in meteorology. We first investigated PM2.5 concentrations and weather conditions and selected stagnant winter days for a case study to minimize the wind speed uncertainty in the SMA. N2O5 experiments were conducted under the same framework of Jo et al. (2019), and four PBL schemes (YSU, ACM2, MYJ, and QNSE) were employed to quantify the PBL bias, while numerical tests were carried out to evaluate the ranges in PBL errors in the WRF-CMAQ model.
Section snippets
Modeling system and domain
To conduct sensitivity analyses associated with air quality, we adopted the Weather Research and Forecast model (WRF, https://www.mmm.ucar.edu/weather-research-and-forecasting-model) and the United States Environmental Protection Agency's (US EPA's) Community Multi-scale Air Quality model (CMAQ v5.0.2, https://www.cmascenter.org/cmaq/). The WRF (v3.6.1) was used to provide input meteorological fields for the CMAQ (ver. 5.0.2), using the grid nudging technique as a data assimilation method (
Base case simulation
The simulated meteorological and chemical variables were compared with the measurements at the Bulgwang site, as shown in Fig. 3. For the meteorological variables, 2 m temperature, RH, 10 m wind speed, and PBL height were extracted at the nearest grid point to the Bulgwang site. In Fig. 3, the overall simulated meteorological variables generally showed similar temporal variations with the measurements, with the exception of the underestimated PBL height. The resulting index of agreement (IOA)
Summary and conclusions
This study analyzed the sensitivity of major meteorological and chemical uncertainty factors that greatly affect the accuracy of air quality models. The WRF-CMAQ was applied to simulate experiments on the PBL as a meteorology uncertainty factor and N2O5 uptake coefficient as a chemical uncertainty factor to identify the effect on the results of PM2.5 numerical simulation in winter. The study period covered January 13–15, 2018, during which the measured NO3− concentrations indicated that the
Author statement
H.-Y. Jo, G. Heo, and C.-H. Kim: Conceptualization, M. Lee, J-.A. Kim, and T. Lee; Data curation M.-S. Park, Y.-H. Lee and S.-W. Kim.; Formal analysis, H.-J. Lee. and Y.-J. Jo.; Investigation, G. Heo.; Methodology, H.-Y. Jo and H.-J. Lee; Visualization, H.-Y. Jo, G. Heo, and C.-H. Kim; Writing—original draft preparation, H.-Y. Jo and C.-H. Kim; Writing—Review and Editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2020R1A6A1A03044834 and 2020R1I1A1A01072998).
References (65)
- et al.
Impact of WRF model PBL schemes on air quality simulations over Catalonia, Spain
Sci. Total Environ.
(2016) - et al.
A quantitative analysis of grid nudging effect on each process of PM2.5 production in the Korean Peninsula
Atmos. Environ.
(2015) - et al.
Nocturnal fine particulate nitrate formation by N2O5 heterogeneous chemistry in Seoul Metropolitan Area, Korea
Atmos. Res.
(2019) - et al.
Changes in inorganic aerosol compositions over the Yellow Sea area from impact of Chinese emissions mitigation
Atmos. Res.
(2020) - et al.
Characterization of PM2.5, PM2.5-10 and PM10 in ambient air, Yokohama, Japan
Atmos. Res.
(2010) - et al.
Effects of the density of meteorological observations on the diagnostic wind fields and the performance of photochemical modeling in the Greater Seoul area
Atmos. Environ.
(2002) - et al.
Influence of an urban canopy model and PBL schemes on vertical mixing for air quality modeling over greater Paris
Atmos. Environ.
(2015) - et al.
Mesoscale atmospheric flow field simulations for air quality modeling over complex terrain region of Ranchi in eastern India using WRF
Atmos. Environ.
(2015) - et al.
Aerosol mass spectrometric measurements of stable crystal hydrates of oxalates and inferred relative ionization efficiency of water
J. Aerosol Sci.
(2011) - et al.
Sensitivity of PBL parameterizations on PM10 and ozone simulation using chemical transport model WRF-Chem over a sub-tropical urban airshed in India
Atmos. Environ.
(2018)