Sensitivity Analysis for Dry Deposition and PM2.5-bound Content of PCDD/Fs in the Ambient Air

This study mainly involved conducting an atmospheric sensitivity analysis of the dry deposition and PM2.5-bound content of total PCDD/Fs-WHO2005-TEQ, respectively. The results for Fuzhou and Xiamen cities showed that the total PCDD/F mass concentration was the factor most positively correlated to the dry deposition flux: When ΔP/P ranged from –50% to 0%, ΔS/S ranged from –66.0% to 0%, but when ΔP/P increased from 0% to +50%, ΔS/S increased from 0% to +66.0%, respectively. The second factor positively correlated with the deposition flux was the PM2.5 concentration: When ΔP/P ranged from –50% to 0%, ΔS/S ranged from –63.3% to 0%; when ΔP/P increased from 0% to +50% and +100%, ΔS/S ranged from 0% to +20.8 and –0.9%, respectively. Ambient air temperature was found to be less sensitive to dry deposition fluxes in total PCDD/FsWHO2005-TEQ: When ΔP/P ranged from –50% to –17% and 0%, ΔS/S ranged from –17.0% to +5.6% and 0%; when ΔP/P increased from 0% to +50%, ΔS/S increased from 0% to –84.5%, respectively. The sensitivity analysis for PM2.5-bound total PCDD/Fs-WHO2005-TEQ content had similar results to those for dry deposition flux. In addition, in 2018, 2019, and 2020, the annual average PM2.5-bound total PCDD/Fs-WHO2005-TEQ content at Fuzhou and Xiamen was 0.430, 0.127, 0.303, and 0.426 ng-WHO2005-TEQ g–1 in the spring, summer, autumn and winter, respectively, which showed that summer had the lowest content, while spring and winter had the highest. The results of this study provided useful information for gaining a deeper understanding of both dry deposition and particle-bound of PCDD/Fs in the ambient air.


INTRODUCTION
Dry deposition is one of the main ways for air pollutants to enter the ecosystem (Mi et al., 2012). The dry deposition of PCDD/Fs is the sum of deposition in the gas phase and particle phase . Ambient air temperature, wind direction, particle size and other factors also affect the atmospheric deposition process (Wu et al., 2009;Chi et al., 2011).
PCDDs, PCDFs, and PCBs are classified as persistent organic pollutants (POPs) due to their toxicity, carcinogenicity and persistence in the human body (Anezaki et al., 2021). Among them, polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs) are collectively known as PCDD/Fs (Wang et al., 2020). They have been identified as pollutants and are present in almost every component of the global ecosystem, including air, aquatic and marine sediments, fish, wildlife, and human adipose tissue and blood (Goldman et al., 1989;Safe et al., 1993). PCDD/Fs are generally produced through waste combustion (Qiu et al., 2020) or are impurities in the production of agricultural chemicals placed side by side in the environment, in turn causing pollution (Safe et al., 1990); for example, the highly toxic 2,3,7, 8-tetrachlorodibenzo-p-dioxin (TCDD) is the main pollutant in herbicides (Kimbrough et al., 1984). A complex mixture of halogenated aromatic hydrocarbons in food and in the environment is the main source of human exposure to PCDD/Fs. The current risk assessment for these compounds is based on toxicity equivalent factors (TEFs), in which each of the 17 compounds is referred to as the highest toxicity 2,3,7,8-TCDD, which has a reference value equal to one. Toxicity is affected by the position of the chlorine atom in the molecule, where 2,3,7 and 8 are the positions where the highest toxicity is presumed. Each additional chlorine atom at these locations usually reduces the potential toxicity by a factor of several times (Ahlborg et al., 1992;Viluksela et al., 1998;Tuomisto et al., 2012;Schröder et al., 2021).
Particulate matter (PM) comprises suspended particles in the atmosphere in the form of aerosols (Ghosh et al., 2014). According to the aerodynamic diameter, PM can be divided into TSP (0-100 µm), PM10 (0-10 µm), and PM2.5 (0-2.5 µm). The harmful effects of air pollutants on the human body are different. In addition to affecting air visibility and affecting global climate change, PM2.5 is also linked to cardiovascular disease and lung cancer Yin et al., 2019). PM2.5 is ubiquitous in the environment and increases the risk of heart, lung, and respiratory diseases. Also, exposure to these particles can cause short-term health effects such as inflammation of the eyes, nose, throat, and lungs (Polezer et al., 2018;Maciejczyk et al., 2018). PM2.5 is produced through combustion processes and other human activities, such as those taking place in power plants, waste incineration facilities, and the steel manufacturing industry (Li et al., 2018;Mari et al., 2016). In addition, PM2.5 has a large surface area and easily adsorbs toxic compounds such as polycyclic aromatic hydrocarbons (PAHs), polybrominated diphenyl ethers (PBDEs), polychlorinated biphenyls (PCBs), polychlorinated dioxins/furans (PCDD/Fs) (Chung et al., 2019). In the environment, PM2.5 can be used as the propagation medium of PCDD/Fs for long-distance transmission, which seriously harms human health. Therefore, it is of great significance to study the PM2.5-bound content of total PCDD/Fs-WHO2005-TEQ in the atmosphere.
The results of sensitivity analyses indicate the degree of response of some parameters to the state quantity of a system. In the study of the distribution of pollutants in air, sensitivity analyses can quantitatively reflect the influence of atmospheric environmental factors on the concentration of toxic pollutants, so as to determine the dominant factors controlling air pollution .
In this study, from 2018 to 2020, a sensitivity analysis of dry deposition and PM2.5-bound content of total PCDD/Fs-WHO2005-TEQ in two cities (Fuzhou and Xiamen) in the south of China was carried out, compared, and discussed.
Fuzhou is located in the eastern part of China, in the eastern part of Fujian, the lower reaches of the Minjiang River and the coastal areas, at an east longitude of 118°08′-120°31′ and a north latitude of 25°15′-26°39′. It has a subtropical monsoon climate. The average annual temperature in Fuzhou ranges from 20°C-25°C; the average annual sunshine is 1700-1980 hours, and the average annual precipitation is 900-2100 mm. The annual relative humidity is about 77%. The dominant wind direction in Fuzhou is northeast, with southerly winds dominating in summer. The weather is hot from July to September, which is the period with the most typhoon activity, and two typhoons directly land in the city, on average, every year.
Xiamen is located in the southeast of Fujian Province in East China, at an east longitude of 117°53'-118°26' and a north latitude of 24°23'-24°54'. It has a subtropical maritime monsoon climate. The average annual temperature in Xiamen is about 21°C, which is neither severely cold in winter nor extremely hot in summer. The annual average rainfall is about 1200 mm, and the period from May to August has the most rainfall every year. The wind is generally at a level of 3-4 m s -1 , often to the dominant wind for the northeast wind. It is affected by 4-5 typhoons, on average, every year, and most of them occur from July to September.

PCDD/F Concentration
In the absence of measured data, the PCDD/F concentration can be simulated using a regression analysis. Two regression analysis equations were selected, and the results were averaged. The two analysis equations are as follows (Huang et al., 2011;Lee et al., 2016): where Yl and Y2 represent the total PCDD/F concentration, and X represents the PM10 concentration in the urban atmosphere.
The goodness-of-fit for the regression equation is R 2 = 0.9855 (Wang et al., 2010;Suryani et al., 2015). The results indicate good forecasting reliability and goodness of fit. In this study, a regression was used to solve the total PCDD/F concentration. The total PCDD/F mass concentration was obtained from the mean values of Y1 and Y2, and the PCDD/Fs were analyzed and discussed respectively by combining the meteorological data for the cities of interest.

Dry Deposition Flux
The dry deposition flux is composed of the diffusion of gaseous matter and the deposition of particulate matter.
where FT represents the total dry deposition flux (pg WHO2005-TEQ m -2 month -1 ); Fg represents the diffusion of gaseous matter producing dry deposition flux (pg WHO2005-TEQ m -2 month -1 ); Fp represents the gravitational settling of particulate matter contributing to dry deposition flux (pg WHO2005-TEQ m -2 month -1 ); CT represents the total PCDD/F concentration in the atmosphere (pg m -3 ); Vd,T represents the dry deposition velocity of PCDD/Fs (gas+particle phases), 0.42 cm s -1 (Shih et al., 2006); Cg represents the calculated PCDD/F concentration in the gas phase (pg m -3 ); Vd,g represents the dry deposition velocity of PCDD/Fs in the gas phase, 0.01 cm s -1 (Sheu et al., 1996); Cp represents the calculated PCDD/F concentration in the particle phase (pg m -3 ), and Vd,p represents the dry deposition velocity of PCDD/Fs in the particle phase (cm s -1 ).

Sensitivity Analysis
In a sensitivity analysis, the influence value σi is used to evaluate the degree of influence of changes in the parameters on the output value, where a higher level of sensitivity indicates a greater influence on the system state variables, and vice versa.
where P0 and P represent the initial and modified value of a parameter; ΔP represents the amount of the parameter to add or subtract; S0 and S represent the initial and predicted values of the parameters, and ΔS represents the response value for each parameter (Zhao et al., 2018a).

Sensitivity Analysis of Dry Deposition Flux
In this study, the sensitivity analysis of the dry deposition flux for total PCDD/Fs-WHO2005-TEQ was mainly focused on the total PCDD/F mass concentration, the PM2.5 concentration, and the ambient air temperature.
In Fuzhou, the sensitivity analysis work was found to be dependent on the initial total PCDD/F mass concentration = 0.5231 pg m -3 , PM2.5 = 23 µg m -3 and an ambient air temperature = 21.5°C. The sensitivity analysis for Xiamen was dependent on the initial total PCDD/F mass concentration values = 0.5231 pg m -3 , PM2.5 = 22 µg m -3 , and ambient air temperature = 22.5°C. The parametric sensitivity for the dry deposition flux of total PCDD/Fs-WHO2005-TEQ in Fuzhou and Xiamen are shown in Figs. 1 and 2, respectively.
According to the sensitivity analysis (Figs. 1 and 2), in terms of the total PCDD/F mass concentration, when ΔP/P ranged from -50% to 0%, ΔS/S ranged from -71.1% to 0%; when ΔP/P increased from 0% to +50% and +80%, ΔS/S ranged from 0% to +71.1% and +85.3%, respectively in Fuzhou. In the case of Xiamen, when ΔP/P ranged from -50% to 0%, ΔS/S ranged from -60.8% to 0%; when ΔP/P increased from 0% to +50% and +60%, ΔS/S increased from 0% to +60.8% and +73.0%, respectively. In Fuzhou and Xiamen, the effect of the total PCDD/F mass concentration on the dry deposition fluxes of total PCDD/Fs-WHO2005-TEQ increased with increases in the concentration. The sensitivity analysis of the PCDD/Fs showed that the total PCDD/F mass concentration in the ambient air had a significant influence on the dry deposition flux of total PCDD/Fs-WHO2005-TEQ, where the sensitivity of this factor was obviously higher than that for other factors.
In Fuzhou (Fig. 1), the impact of the PM2.5 concentration on the dry deposition fluxes of total PCDD/Fs-WHO2005-TEQ can be divided into two parts: When ΔP/P ranged from -45% to 0%, ΔS/S ranged from -52.9% to 0%, but when ΔP/P increased from 0% to +45% and +100%, ΔS/S fluctuated from 0% to +17.3% and -10.0%, respectively. In the case of Xiamen (Fig. 2), for the PM2.5  concentration, when ΔP/P ranged from -55% to 0%, ΔS/S ranged from -74.4% to 0%, but when ΔP/P increased from 0% to +55% and +100%, ΔS/S increased from 0% to +24.8% and +8.2%, respectively. PM can thus reflect the PCDD/F mass concentration in the particle phase. In Fuzhou and Xiamen, the effect of the PM2.5 concentration on the dry deposition fluxes of total PCDD/Fs-WHO2005-TEQ increased first and then decreased with increases in the PM2.5 concentration. The above results indicated that the sensitivity of the PM2.5 concentration on the dry deposition flux of total PCDD/Fs-WHO2005-TEQ was weaker when the PM2.5 concentration was higher than 23 µg m -3 , or the total PCDD/F mass concentration was greater than 0.523 pg m -3 .
When the PM2.5 concentration increases (Figs. 1 and 2), the effect of the ambient air temperature on the dry deposition flux of total PCDD/Fs-WHO2005-TEQ increases first and then decreases. For Fuzhou, when ΔP/P ranged from -50% to -17% and 0%, ΔS/S ranged from -8.9% to +3.7% and 0%, but when ΔP/P increased from 0% to +50%, ΔS/S decreased from 0% to -50.6%, respectively. In the case of Xiamen, when ΔP/P ranged from -50% to -17% and 0%, ΔS/S ranged from -25.2% to +7.5% and 0%, and when ΔP/P increased from 0% to +40%, ΔS/S decreased from 0% to -83.2%, respectively. Temperature affects the dry deposition flux of total PCDD/Fs-WHO2005-TEQ by changing the gas-particle distribution of PCDD/Fs. As the temperature increases, the PCDD/Fs will cause a larger amount of particle-phase PCDD/F mass to evaporate into the gas phase. When the temperature is lower than -17.0°C, this parameter has a positive effect on the dry deposition flux of the total PCDD/Fs-WHO2005-TEQ; when the temperature is higher than -17.0°C, the air temperature is negatively correlated with the dry deposition flux of the total PCDD/Fs-WHO2005-TEQ.

Level of Dry Deposition of Total PCDD/Fs-WHO2005-TEQ
The monthly dry deposition fluxes of total PCDD/Fs-WHO2005-TEQ in Fuzhou and Xiamen in 2018, 2019, and 2020 are presented in Fig. 4.
The dry deposition fluxes of total PCDD/Fs-WHO2005-TEQ in Fuzhou and Xiamen were the lowest in summer and the highest in spring. The results showed that the particulate phase PCDD/Fs were easily removed by dry deposition, and the temperature was negatively correlated with the particle phase concentration. At low temperatures, PCDD/Fs mainly exists in the particle phase, and more PCDD/Fs in the particle phase are removed by dry deposition under the force of gravity, so the dry deposition flux increases with a decrease in the temperature.

Sensitivity Analysis of PM2.5-bound Total PCDD/Fs-WHO2005-TEQ Content
A sensitivity analysis can provide a basis for determining some important parameters of PM2.5bound total PCDD/Fs-WHO2005-TEQ content. In this study, the total PCDD/F mass concentration, PM2.5 concentration, and the ambient air temperature may have affected the total PM2.5-bound total PCDD/Fs-WHO2005-TEQ content. In the sensitivity analysis for Fuzhou, the initial values of the total PCDD/F mass concentration = 0.4976 pg m -3 ; the PM2.5 = 22 µg m -3 , and the ambient air temperature = 21.5°C. The parametric sensitivity for the PM2.5-bound total PCDD/Fs-WHO2005-TEQ content in Fuzhou is shown in Fig. 5. In the sensitivity analysis of Xiamen, the initial values of the total PCDD/F mass concentration = 0.5231 pg m -3 , the PM2.5 = 22 µg m -3 , and the ambient air temperature = 22.5°C. The parametric sensitivity for the total PM2.5-bound total PCDD/Fs-WHO2005-TEQ content in Xiamen is shown in Fig. 6.
In regard to the total PCDD/F mass concentration parameters (Figs. 5 and 6), when ΔP/P ranged from -20% to 0%, ΔS/S ranged from -74.3% to 0%, but when ΔP/P increased from 0% to +55% and +100%, ΔS/S increased from 0% to +89.1% and +36.5%, respectively in Fuzhou. In the case of Xiamen, when ΔP/P ranged from -20% to 0%, ΔS/S ranged from -41.3% to 0%; when ΔP/P increased from 0% to +45% and +100%, ΔS/S fluctuated from 0% to +38.2% and -18.0%, respectively. In Fuzhou and Xiamen, the effect of the PCDD/F mass concentration on PM2.5-bound total PCDD/Fs-WHO2005-TEQ content increased first and then decreased with increases in the PCDD/F mass concentration. The results showed that the total PCDD/F mass concentration in the two cities had a significant effect on the total PM2.5-bound total PCDD/Fs-WHO2005-TEQ content. This may have been because the total PCDD/Fs-WHO2005-TEQ concentration depends on the total PCDD/F mass concentration, so a change in the PCDD/F mass concentration has a significant impact on the total PM2.5-bound total PCDD/Fs-WHO2005-TEQ content.
In conclusion, PM2.5-bound total PCDD/Fs-WHO2005-TEQ content in Fuzhou and Xiamen was most sensitive to the total PCDD/F mass concentration, followed by the PM2.5 concentration and then the ambient air temperature.
The results show that the average PM2.5-bound total PCDD/Fs-WHO2005-TEQ content in summer was lower than that in the other three seasons. This is because with the rise in temperature, a large amount of PCDD/Fs evaporates from the particle phase to the gas phase, resulting in a lower content of total PCDD/Fs-WHO2005-TEQ. In addition, the PM2.5-bound total PCDD/Fs-WHO2005-TEQ content in Fuzhou and Xiamen in 2020 was lower than that in 2018-2019, especially in February, and the content in both cities was significantly reduced. This may have been related to COVID-19 in 2020, where control measures significantly improved the air quality.

Total PCDD/Fs-WHO2005-TEQ Concentration
The total PCDD/Fs-WHO2005-TEQ concentration in Fuzhou and Xiamen in the ambient air, from 2018 to 2020 are shown in Fig. 9.
It can be seen that the total PCDD/Fs-WHO2005-TEQ concentrations in the ambient air of Fuzhou exhibit seasonal changes, with the highest concentrations in spring and the lowest concentrations in summer. This may be related to atmospheric inversion in winter. When a temperature inversion occurs due to the weakening of the wind, the atmosphere is in a stable state, and the upper and lower levels of air do not exchange easily. At this time, if a large amount of waste gas is discharged into the atmosphere, it is difficult to diffuse the polluted gas due to interference from the inversion layer, so the concentration of polluted gas in the atmosphere increases, thereby increasing the intensity of the air pollution. In addition, the average total PCDD/Fs-WHO2005-TEQ concentration in 2020 (0.0326 pg-WHO2005-TEQ m -3 ) was 12.7% lower than the average in 2018-2019 (0.0326 pg-WHO2005-TEQ m -3 ), especially in February. In February 2020, Fuzhou formulated control measures, including home quarantine and suspension of work and production, to focus COVID-19 prevention and control. Therefore, the total PCDD/Fs-WHO2005-TEQ concentration in February 2020 was 26.7% lower than that in February of 2018 and 2019, indicating a significant improvement in atmospheric quality due to COVID-19 control measures.