Ambient BTEX Concentrations during the COVID-19 Lockdown in a Peri-Urban Environment (Orl é ans, France)

: During the period from 17 March to 10 May 2020, France saw dramatic shifts in domestic, industrial and transport activities as a national lockdown was introduced. So far, studies have generally focused on urban settings, by contrast, this work reports data for a peri-urban location. Air samples were collected and analyzed using a fully automated GC-MS-FID system in an air quality monitoring station situated in the suburbs of Orl é ans, France. Average concentrations of BTEX (benzene, toluene, ethylbenzene, and xylenes) before, during, and after lockdown, were 402 ± 143, 800 ± 378 and 851 ± 445 pptv, respectively. Diurnal variation in BTEX and correlations between each of its components were analyzed to determine its various sources. The toluene/benzene (T/B) and m,p-xylene/ethylbenzene (MP/E) ratios, photochemical ages were used to explore whether the BTEX were from local or more distant sources. Together with a host of complementary measurements including NOx, O 3 , black carbon, meteorological parameters, and anthropogenic activities, we were able to make some inferences on the sources of BTEX. The results suggest that although anomalous local anthropogenic activity can lead to signiﬁcant changes in BTEX concentrations, pollution levels in Orl é ans are mostly dependent on meteorological conditions, speciﬁcally whether the winds are coming from the Paris region. It appears, based on these measurements, that the pollution in the Orl é ans area is very much tied to the nearby megacity of Paris, this may be true for other peri-urban sites with implications for city planning and pollution mitigation strategies.


Introduction
The COVID-19 pandemic has resulted in reduced emissions into the atmosphere worldwide as a result of the actions taken to limit or even stop certain human activities, which are generally linked to transportation and industrial activities. In some locations, these measures led to substantial reductions in the emissions of important atmospheric pollutants such as nitrogen oxides (NO x ), particulate matter (PM), and volatile organic compounds (VOCs) as reported by a number of studies around the globe. For example, Barré et al. [1] and Guevara et al. [2] reported emissions dropping by up to 60% for NO x , and up to 15% for non-methane volatile organic compounds (NMVOC) during the March/April 2020 lockdown in Europe while Shi and Brasseur [3] showed that the surface concentrations of PM 2.5 and NO 2 in China were reduced by 35% and 60%, respectively. It is expected that the NO x levels decreased in urban areas during lockdown periods since

Sampling Site
As shown in Figure 1, the sampling site is positioned within the campus of the Centre National de la Recherche Scientifique (CNRS) about 10 km south from Orléans city center (47 • 50 17" N, 1 • 56 39" E). Orléans is the capital of the Centre-Val de Loire region, located in central France about 120 kilometers southwest of Paris, with a population of 300 thousand inhabitants. The site is close to a forest belt, and farms. There are no obstructing buildings around the sampling site within 50 m. The measurements reported in this study were made during the period from 26 February to 31 May in 2020.

BTEX Measurements
An automated online gas chromatographic system (GC-FID, AirmoVOC C6-C16 Chromatotec ® , Val-deVirvée -France) was used for in situ BTEX measurements with a time resolution of 40 min. Ambient air was sampled continuously from an inlet mounted at a height of 4 m above the ground through a Teflon tube (length 2.5 m, diameter 1/4 inch) at a 16 L min −1 bypass flow, and then entered the trap at a flow rate of 45 mL min -1 through a heated stainless-steel tube (length ~1.5 m, diameter 1/8 inch). This sampling setup could ensure a short residence time in the sampling line and improve the sampling efficiency of the C6-C16 VOCs. The volume of air samples was 1350 mL for a 30 min sampling period. The trap was heated to 380 °C within 4 min. VOCs were thermally desorbed and injected into capillary columns (analytical column: MXT 30 CE, film thickness: 1 µm, id: 0.28 mm, length: 30 m) with H2 around 3-4 mL min -1 for separation before detection. The temperature ramp of the capillary column (a plot of temperature vs. time is provided in Figure S1 in the supporting information) was set as follows: increased from 38 to 50 °C at a rate of 2 °C min −1 ; increased to 80 °C in 3 min; increased to 220 °C at a rate of 15 °C min −1 , then went to 230 °C at a rate of 2 °C min −1 ; finally reached 269 °C at a rate of 9 °C min −1 and held for 3 min. A flame ionization detector (FID) set at 200 °C was used for quantification.
The accuracy of the system (quantification and identification) was controlled every 12h using an automatic calibration system, which is equipped with four internal permeation tubes containing four internal standards (benzene, n-butane, n-decane, and n-hexane). The detection limits for benzene, toluene, ethylbenzene, m, p-xylene, and o-xylene were ~5 pptv.

BTEX Measurements
An automated online gas chromatographic system (GC-FID, AirmoVOC C6-C16 Chromatotec ® , Val-deVirvée, France) was used for in situ BTEX measurements with a time resolution of 40 min. Ambient air was sampled continuously from an inlet mounted at a height of 4 m above the ground through a Teflon tube (length 2.5 m, diameter 1/4 inch) at a 16 L min −1 bypass flow, and then entered the trap at a flow rate of 45 mL min −1 through a heated stainless-steel tube (length~1.5 m, diameter 1/8 inch). This sampling setup could ensure a short residence time in the sampling line and improve the sampling efficiency of the C6-C16 VOCs. The volume of air samples was 1350 mL for a 30 min sampling period. The trap was heated to 380 • C within 4 min. VOCs were thermally desorbed and injected into capillary columns (analytical column: MXT 30 CE, film thickness: 1 µm, id: 0.28 mm, length: 30 m) with H 2 around 3-4 mL min −1 for separation before detection. The temperature ramp of the capillary column (a plot of temperature vs. time is provided in Figure S1 in the supporting information) was set as follows: increased from 38 to 50 • C at a rate of 2 • C min −1 ; increased to 80 • C in 3 min; increased to 220 • C at a rate of 15 • C min −1 , then went to 230 • C at a rate of 2 • C min −1 ; finally reached 269 • C at a rate of 9 • C min −1 and held for 3 min. A flame ionization detector (FID) set at 200 • C was used for quantification.
The accuracy of the system (quantification and identification) was controlled every 12h using an automatic calibration system, which is equipped with four internal permeation tubes containing four internal standards (benzene, n-butane, n-decane, and n-hexane). The detection limits for benzene, toluene, ethylbenzene, m, p-xylene, and o-xylene werẽ 5 pptv.  The concentrations of ozone (O 3 ), NO x (include NO and NO 2 ), BC, and particulate matter with a diameter less than 10 µm (PM 10 ) were provided by the air quality monitoring agency at the Centre-Val de Loire region by Lig'Air (http://www.ligair.fr/, last accessed on 15 December 2021). These species were monitored at the same location as the BTEX measurements reported in the present work. Ambient air was sampled continuously from separate Teflon tubes (diameter 1/4 inch) at a height of 1 m above the roof of the container and about 4 m from the ground. BC was measured continuously at 1 min intervals using an aethalometer (AE33 model, Magee Scientific, Berkeley, CA, USA). The O 3 and NO-NO x concentrations were continuously monitored by UV absorption (Thermo Scientific™ Model 49i, Waltham, MA, USA) and chemiluminescence (Thermo Scientific™ Model 42i, Waltham, MA, USA), respectively. A DIGITEL DHA-80 high volume aerosol sampler was also used for PM 10 sampling. Most of the data processing and figures were performed using the R software environment [21] and particularly the Openair package, designed by Carslaw et al. [22].

Meteorological Data
The meteorological data were used from Orléans-Bricy air base station (https://www. infoclimat.fr/, last accessed on 15 December 2021).) due to missing data for several weeks during lockdown from the measurement site.

Data Overview
The measurement period (26 February to 31 May 2020) was divided in three periods (a) before lockdown (26 February-16 March), (b) during lockdown (17 March-10 May), and (c) after lockdown . Average BTEX mixing ratios along with their associated arithmetic mean and interquartile range (IQR) in each period are listed in Table 1. Benzene and ethylbenzene were the most abundant BTEX compounds throughout the study period. On average, the highest concentrations of BTEX were observed during and after the lockdown. The temporal variations of BTEX, NO, NO 2 , O 3 , PM 10 , BC, temperature at this site, and meteorological parameters (wind speed in m s −1 and wind direction in a color-coded way) during the measurement period are illustrated in Figure 2. Furthermore, the relative changes in daily regional mobility data from the Google Community Mobility Reports (https://www.google.com/covid19/mobility/, last accessed on 15 June 2020)) were also used for this analysis (the bottom of Figure 2). The reports chart movement trends over time by geography, across different categories of places such as retail and recreation, groceries and pharmacies, parks, transit stations, workplaces, and residential. These reports are created with aggregated, anonymized sets of data from users who have turned on the Location History setting, which is off by default. This provides a breakdown by category (grocery stores, parks, residential, workplaces, transit stations, and retail) in changes to daily movement at a regional level. The data "percent charge in the grocery and pharmacy" was used to perform the comparisons along the studied period.
Atmosphere 2022, 13, x FOR PEER REVIEW 5 of 15 and pharmacies, parks, transit stations, workplaces, and residential. These reports are created with aggregated, anonymized sets of data from users who have turned on the Location History setting, which is off by default. This provides a breakdown by category (grocery stores, parks, residential, workplaces, transit stations, and retail) in changes to daily movement at a regional level. The data "percent charge in the grocery and pharmacy" was used to perform the comparisons along the studied period. , and temperature (°C) in Orléans during the studied period, with daily "percent change in the grocery and pharmacy", and meteorological parameters (wind speed in m s -1 , wind direction in a color-coded way).

Diurnal Variation and Correlations
The study of diurnal variations of air pollutants could provide valuable information about the sources, transport, and chemical formation/destruction of such pollutants. The diurnal variations of the BTEX, BC, PM10, NOx, and O3 during the present study are shown in Figure 3 (a series of diurnal variations in benzene, toluene, ethylbenzene, m,p-xylene , and temperature ( • C) in Orléans during the studied period, with daily "percent change in the grocery and pharmacy", and meteorological parameters (wind speed in m s −1 , wind direction in a color-coded way).

Diurnal Variation and Correlations
The study of diurnal variations of air pollutants could provide valuable information about the sources, transport, and chemical formation/destruction of such pollutants. The diurnal variations of the BTEX, BC, PM 10 , NO x , and O 3 during the present study are shown in Figure 3 (a series of diurnal variations in benzene, toluene, ethylbenzene, m,p-xylene and o-xylene mixing ratios before, during and after lockdown are provided in the supporting information, Figure S2). and o-xylene mixing ratios before, during and after lockdown are provided in the supporting information, Figure S2). BTEX concentrations have been observed to be generally higher during the night than daytime. The concentrations-time profiles show a decrease of the total BTEX from around 6 am to around 3 pm in the three periods. This is likely to result from a combination of effects such as photochemical reactions, an increase in the atmospheric mixing depth, and/or a reduction in emission rates [23]. During the afternoon, an increase in concentration is observed around 3 pm. This increase is observed during the period before the lockdown, which could be attributed to traffic emissions and home heating fuels consumed in winter in the late afternoon. Some subtle changes in the diurnal pattern of BC concentrations can be observed before, during, and after lockdown. It appears, for example, that pre-lockdown, there were small concentrations throughout the day, increasing steadily BTEX concentrations have been observed to be generally higher during the night than daytime. The concentrations-time profiles show a decrease of the total BTEX from around 6 am to around 3 pm in the three periods. This is likely to result from a combination of effects such as photochemical reactions, an increase in the atmospheric mixing depth, and/or a reduction in emission rates [23]. During the afternoon, an increase in concentration is observed around 3 pm. This increase is observed during the period before the lockdown, which could be attributed to traffic emissions and home heating fuels consumed in winter in the late afternoon. Some subtle changes in the diurnal pattern of BC concentrations can be observed before, during, and after lockdown. It appears, for example, that prelockdown, there were small concentrations throughout the day, increasing steadily into the night, which could reflect people heating their houses during the cold hours of the evening. During lockdown, BC is present in higher concentrations in the early to mid-morning period, decreasing over mid-day. This could potentially indicate changes in living habits as people are working from home and are needing to heat. This is also seen post-lockdown, since many people remained working from home during that period, although the levels were lower probably as a consequence of the warmer ambient conditions. The highest correlations between BC and PM 10 were observed during the lockdown, which may indicate a common, possibly wood-burning source. The diurnal profiles of NOx were not very distinct before lockdown, and show typical diurnal variability during and after lockdown, which be a consequence of a difference in wind direction before lockdown. No clear influence of lockdown was observed in the diurnal profiles of ozone.
Moreover, a strong correlation coefficient (r = 0.83) was obtained for ethylbenzene/xylene isomers before the lockdown (Figure 4a). This fact suggests that both compounds could be emitted from common sources [24]. Toluene is also widely used as a solvent in paint manufacture, the production of adhesives and glues, while benzene emissions could be mainly due to wintertime domestic heating combustion in France [25]. During the lockdown (Figure 4b), the high correlations among all the BTEX suggests that the contributions are likely to be emitted by the same sources. The correlation becomes even stronger after the lockdown period as shown by Figure 4c, which indicates that the sources of BTEX were becoming increasingly homogenous.
into the night, which could reflect people heating their houses during the cold hours of the evening. During lockdown, BC is present in higher concentrations in the early to midmorning period, decreasing over mid-day. This could potentially indicate changes in living habits as people are working from home and are needing to heat. This is also seen post-lockdown, since many people remained working from home during that period, although the levels were lower probably as a consequence of the warmer ambient conditions. The highest correlations between BC and PM10 were observed during the lockdown, which may indicate a common, possibly wood-burning source. The diurnal profiles of NOx were not very distinct before lockdown, and show typical diurnal variability during and after lockdown, which be a consequence of a difference in wind direction before lockdown. No clear influence of lockdown was observed in the diurnal profiles of ozone.
Moreover, a strong correlation coefficient (r = 0.83) was obtained for ethylbenzene/xylene isomers before the lockdown (Figure 4a). This fact suggests that both compounds could be emitted from common sources [24]. Toluene is also widely used as a solvent in paint manufacture, the production of adhesives and glues, while benzene emissions could be mainly due to wintertime domestic heating combustion in France [25]. During the lockdown (Figure 4b), the high correlations among all the BTEX suggests that the contributions are likely to be emitted by the same sources. The correlation becomes even stronger after the lockdown period as shown by Figure 4c, which indicates that the sources of BTEX were becoming increasingly homogenous.

BTEX Ratios Assessment
The Toluene/Benzene ratio (T/B) is often used in BTEX source apportionment studies [26]. It is a tool for characterizing the distance from (and age of) vehicular emission sources. While toluene emissions are associated with traffic and industries, benzene from these sources has been reduced as a consequence of gasoline regulations. The principal source of benzene in France is therefore likely to be residential heating [27].
It can be observed in Figure 5 that the T/B ratio falls below 1 in general, with some higher values observed in the first week of lockdown. This suggests that there were no major changes in BTEX sources during our study period, which is surprising, given the decrease in transport activities and the change in the population's daily routines during lockdown. When the location of the measurement site in Orléans is considered in relation to potential BTEX sources in Paris, the T/B ratios < 1 that we observe are broadly consistent with the measurements of Salameh [28], who show that in suburban sites such as Melun and other locations on the outskirts of Paris, T/B ratios are close to 1. Nevertheless, it is possible that a significant contribution of the benzene concentration comes from residential heating as mentioned above. Temperature was not a strong determinant of T/B, which suggests that evaporation rates were not the major control on this ratio in our study.

BTEX Ratios Assessment
The Toluene/Benzene ratio (T/B) is often used in BTEX source apportionment studies [26]. It is a tool for characterizing the distance from (and age of) vehicular emission sources. While toluene emissions are associated with traffic and industries, benzene from these sources has been reduced as a consequence of gasoline regulations. The principal source of benzene in France is therefore likely to be residential heating [27].
It can be observed in Figure 5 that the T/B ratio falls below 1 in general, with some higher values observed in the first week of lockdown. This suggests that there were no major changes in BTEX sources during our study period, which is surprising, given the decrease in transport activities and the change in the population's daily routines during lockdown. When the location of the measurement site in Orléans is considered in relation to potential BTEX sources in Paris, the T/B ratios < 1 that we observe are broadly consistent with the measurements of Salameh [28], who show that in suburban sites such as Melun and other locations on the outskirts of Paris, T/B ratios are close to 1. Nevertheless, it is possible that a significant contribution of the benzene concentration comes from residential heating as mentioned above. Temperature was not a strong determinant of T/B, which suggests that evaporation rates were not the major control on this ratio in our study. In this study, the photochemical age was used to estimate the origin of the air masses. Its definition depends on the fact that the more reactive chemicals are consumed faster than the less reactive ones, so that as an air mass ages there are systematic changes in concentration ratios [29]. The T/B ratio in the sampled air as described by Roberts [30]: where ktoluene and kbenzene are the rate coefficients for the reaction with OH (ktoluene = 5.63 × 10 -12 and kbenzene = 1.22 × 10 -12 at 298 K in cm 3 molecule -1 s -1 [31]. [OH] is the average concentration of the OH radical (2.1 × 10 6 molecule cm -3 , [32]). The initial emission ratio of T/B was set to 1.69 (initial value of B/T = 0.59), taken from Airparif station, which is located on the roof of the main Airparif building, close to busy roads, in the Paris city center [33]. It can be seen from Figure 5 that when the T/B ratio increases, the photochemical age decreases. For example, on the evening of 13 March and the early morning of 16 March, it is as low as 1.37 and 1.14 days, respectively. This suggests that the pollutants are fresher and the source of the pollution is relatively close to the site (100~300 km). In most of the observation period, the photochemical age is more than 3 days, indicating that long-distance air transport could contribute to the air pollution in this region (> 500 km).
The m,p-xylene to ethylbenzene (MP/E) ratio could also be used to evaluate the age of air parcels and as an indicator for the photochemical age of the VOCs in the atmosphere. The xylenes are more reactive towards the OH radical than ethylbenzene so the low ratio of MP/E could be used as an indication of an aged air parcel. Relatively constant MP/E ratios ranging from 2.8 to 4.6 with a mean value of 3.5 due to near traffic exhaust emissions have been previously reported [34]. MP/E ratio along the studied period is presented in Figure 5. MP/E ratios average shows a quite small but noticeable difference before (0.87), during (0.72), and after (0.59) the lockdown. The values in the three periods suggest a rather aged air parcel, and the decreasing trend is an expected effect due to the increase in solar radiation as the spring season approaches. During and after lockdown the variation in the MP/E ratio reaches its maximum value before sunrise and its minimum at midafternoon ( Figure 5), in agreement with what is expected for a photochemically aged air parcel [35]. On the other hand, before the lockdown, this cyclical variation in the MP/E In this study, the photochemical age was used to estimate the origin of the air masses. Its definition depends on the fact that the more reactive chemicals are consumed faster than the less reactive ones, so that as an air mass ages there are systematic changes in concentration ratios [29]. The T/B ratio in the sampled air as described by Roberts [30]: where k toluene and k benzene are the rate coefficients for the reaction with OH (k toluene = 5.63 × 10 −12 and k benzene = 1.22 × 10 −12 at 298 K in cm 3 molecule −1 s −1 [31]. [OH] is the average concentration of the OH radical (2.1 × 10 6 molecule cm −3 , [32]). The initial emission ratio of T/B was set to 1.69 (initial value of B/T = 0.59), taken from Airparif station, which is located on the roof of the main Airparif building, close to busy roads, in the Paris city center [33]. It can be seen from Figure 5 that when the T/B ratio increases, the photochemical age decreases. For example, on the evening of 13 March and the early morning of 16 March, it is as low as 1.37 and 1.14 days, respectively. This suggests that the pollutants are fresher and the source of the pollution is relatively close to the site (100~300 km). In most of the observation period, the photochemical age is more than 3 days, indicating that long-distance air transport could contribute to the air pollution in this region (> 500 km).
The m,p-xylene to ethylbenzene (MP/E) ratio could also be used to evaluate the age of air parcels and as an indicator for the photochemical age of the VOCs in the atmosphere. The xylenes are more reactive towards the OH radical than ethylbenzene so the low ratio of MP/E could be used as an indication of an aged air parcel. Relatively constant MP/E ratios ranging from 2.8 to 4.6 with a mean value of 3.5 due to near traffic exhaust emissions have been previously reported [34]. MP/E ratio along the studied period is presented in Figure 5. MP/E ratios average shows a quite small but noticeable difference before (0.87), during (0.72), and after (0.59) the lockdown. The values in the three periods suggest a rather aged air parcel, and the decreasing trend is an expected effect due to the increase in solar radiation as the spring season approaches. During and after lockdown the variation in the MP/E ratio reaches its maximum value before sunrise and its minimum at mid-afternoon ( Figure 5), in agreement with what is expected for a photochemically aged air parcel [35]. On the other hand, before the lockdown, this cyclical variation in the MP/E ratio between day and night is not clearly observed, this may be due to the different emission sources during winter ( Figure 6). Figure 2 shows that the prevailing wind direction changed significantly after the first period. This is likely to have also a strong effect, transporting air from different locations with different distance.
Atmosphere 2022, 13, x FOR PEER REVIEW 9 of 15 ratio between day and night is not clearly observed, this may be due to the different emission sources during winter ( Figure 6). Figure 2 shows that the prevailing wind direction changed significantly after the first period. This is likely to have also a strong effect, transporting air from different locations with different distance.

Impact From Community Mobility and Meteorological Effects on the Measured BTEX
In order to get a better understanding of the BTEX sources we made a series of supporting measurements including meteorological information, O3, NO, NO2, PM10, and BC. For this purpose, we investigated correlations over four episodes taken from the whole period of investigation: before lockdown (26 February-12 March), declared lockdown (13 ,15 and 16 March), lockdown (17 March-10 May), and after the lockdown (11 May-31 May). Figure 7 shows correlation coefficient between BTEX, O3, NO, NO2, PM10, BC, and temperature (T) for each period.
BTEX average concentrations were relatively low before lockdown, with an average mixing ratio around 402 ± 143 pptv. Nevertheless, on 13 March and 15-16 March, two higher concentration peaks of BTEX were detected, with about 820 and 1040 pptv, respectively. To better understand the possible causes and identify the main contribution sources for these particular events, correlation coefficients were calculated using the corresponding data of these particular days as shown in Figure 7b. From this figure, it can be seen that there was a high positive correlation (r = 0.90) between BC and NO2, but BTEX did not correlate very well with either of them (r = 0.51 with NO2 and r = 0.63 with BC), suggesting that it could be another source besides fuels and/or biomass burning [36].
During the same period, the warm weather would suggest that it is less likely that the burning of wood for home heating represented a significant source, however, given the extremely different daily schedules of people during this timeframe, it remains possible that residential heating was a contributing factor. Considering these facts, BTEX concentration increase during these episodes could be related to transportation emissions rather than other sources. This may be linked to the announcement of official measures to be implemented in order to prevent the spread of COVID-19 on 12 March. Similarly, for 15 and 16 March with the order from the French Government to close all non-essential public places and businesses, including restaurants, cafes, shops, and entertainment venues from 14 March in order to limit social movements. To support this hypothesis, the Google Community Mobility Reports data for Friday 13, Sunday 15, and Monday 16 of March were also used. From these data, displayed in Figure 2, it is possible to observe a substantial increase in the mobility of people (percent change) on the same dates where the maximum concentrations of BTEX were observed. Considering this, it could be speculated that most of the emissions detected during 13, 15, and 16 March were strongly related to the local use of transport vehicles on an elevated scale to visit pharmacies and grocery stores.

Impact from Community Mobility and Meteorological Effects on the Measured BTEX
In order to get a better understanding of the BTEX sources we made a series of supporting measurements including meteorological information, O 3 , NO, NO 2 , PM 10 , and BC. For this purpose, we investigated correlations over four episodes taken from the whole period of investigation: before lockdown (26 February-12 March), declared lockdown (13,15 and 16 March), lockdown (17 March-10 May), and after the lockdown (11 May-31 May). Figure 7 shows correlation coefficient between BTEX, O 3 , NO, NO 2 , PM 10 , BC, and temperature (T) for each period.
BTEX average concentrations were relatively low before lockdown, with an average mixing ratio around 402 ± 143 pptv. Nevertheless, on 13 March and 15-16 March, two higher concentration peaks of BTEX were detected, with about 820 and 1040 pptv, respectively. To better understand the possible causes and identify the main contribution sources for these particular events, correlation coefficients were calculated using the corresponding data of these particular days as shown in Figure 7b. From this figure, it can be seen that there was a high positive correlation (r = 0.90) between BC and NO 2 , but BTEX did not correlate very well with either of them (r = 0.51 with NO 2 and r = 0.63 with BC), suggesting that it could be another source besides fuels and/or biomass burning [36].
During the same period, the warm weather would suggest that it is less likely that the burning of wood for home heating represented a significant source, however, given the extremely different daily schedules of people during this timeframe, it remains possible that residential heating was a contributing factor. Considering these facts, BTEX concentration increase during these episodes could be related to transportation emissions rather than other sources. This may be linked to the announcement of official measures to be implemented in order to prevent the spread of COVID-19 on 12 March. Similarly, for 15 and 16 March with the order from the French Government to close all non-essential public places and businesses, including restaurants, cafes, shops, and entertainment venues from 14 March in order to limit social movements. To support this hypothesis, the Google Community Mobility Reports data for Friday 13, Sunday 15, and Monday 16 of March were also used. From these data, displayed in Figure 2, it is possible to observe a substantial increase in the mobility of people (percent change) on the same dates where the maximum concentrations of BTEX were observed. Considering this, it could be speculated that most of the emissions detected during 13, 15, and 16 March were strongly related to the local use of transport vehicles on an elevated scale to visit pharmacies and grocery stores. During the lockdown, social mobility decreased significantly as it can be seen from the relative changes in the Mobility Reports (Figure 2). At the same time, NO concentrations also showed a progressive decrease, which is likely to be associated with the reduction of local traffic emissions. However, BTEX concentrations showed multiple elevated episodes during the lockdown period. From the correlation coefficients shown in Figure  7, it is possible to observe that this increase does not correlate with the NO concentration (r = 0.02), is weakly correlated with NO2 (r = 0.29), and only shows a moderate correlation with BC (r = 0.45). Moreover, from Figure 2, it is possible to observe that wind direction changed markedly in the region, from southwest before lockdown to northeast during the lockdown.
To consider a possible effect of these variables, polar plots of each pollutant and temperature against wind direction were constructed in Figure 8, where each factor is plotted as a function of wind speed and wind direction. The dominant wind direction came from the northeast 43.1% of the time, which was most frequent during the lockdown. Other wind directions were north 15%, east 8.7%, southeast 5.2%, south 8.1%, southwest 12% west 5.8%, and northwest 2%. This is consistent with trajectory results from HYSPLIT (https://www.ready.noaa.gov/HYSPLIT_traj.php, accessed on 25 October 2021). Briefly as shown in Figure 9, before the lockdown (panel a), air masses mainly originated from the west, from the Atlantic Ocean. This represents a relatively clean air masses, in line During the lockdown, social mobility decreased significantly as it can be seen from the relative changes in the Mobility Reports (Figure 2). At the same time, NO concentrations also showed a progressive decrease, which is likely to be associated with the reduction of local traffic emissions. However, BTEX concentrations showed multiple elevated episodes during the lockdown period. From the correlation coefficients shown in Figure 7, it is possible to observe that this increase does not correlate with the NO concentration (r = 0.02), is weakly correlated with NO 2 (r = 0.29), and only shows a moderate correlation with BC (r = 0.45). Moreover, from Figure 2, it is possible to observe that wind direction changed markedly in the region, from southwest before lockdown to northeast during the lockdown.
To consider a possible effect of these variables, polar plots of each pollutant and temperature against wind direction were constructed in Figure 8, where each factor is plotted as a function of wind speed and wind direction. The dominant wind direction came from the northeast 43.1% of the time, which was most frequent during the lockdown. Other wind directions were north 15%, east 8.7%, southeast 5.2%, south 8.1%, southwest 12%, west 5.8%, and northwest 2%. This is consistent with trajectory results from HYSPLIT (https://www.ready.noaa.gov/HYSPLIT_traj.php, accessed on 25 October 2021). Briefly, as shown in Figure 9, before the lockdown (panel a), air masses mainly originated from the west, from the Atlantic Ocean. This represents a relatively clean air masses, in line with lower BTEX levels observed before the lockdown (Figure 8). However, during and after the lockdown (panels b and c, respectively), air masses have a more continental origin with various directions with more southerly and westerly components which is likely to pick up pollution plumes from built-up areas such as Paris which helps to explain the higher BTEX levels observed during these periods.
Atmosphere 2022, 13, x FOR PEER REVIEW 11 of 15 with lower BTEX levels observed before the lockdown (Figure 8). However, during and after the lockdown (panels b and c, respectively), air masses have a more continental origin with various directions with more southerly and westerly components which is likely to pick up pollution plumes from built-up areas such as Paris which helps to explain the higher BTEX levels observed during these periods. For NO and NO2, polar plot representations clearly show that maximum concentrations were observed for winds blowing from the southwest. The southwest of the site is located near a traffic intersection, gas station, and provincial road, which would indicate very local sources.
Another area with high NO2 concentration was found in the north, which has a similar situation for BC and PM10. It is noted that the site is located in the south of Orléans (distance 6 km from city center) and south west of Paris (distance 90 km), and that the major source of volatile organic compounds during wintertime in the Paris region is wood burning [33]. Hence, impact of biomass burning emissions, especially during the lockdown, was clearly evidenced on NO2, BC, and PM10 atmospheric concentrations.
Concerning the BTEX mixing ratios, the probability to observe high concentrations decreased with the wind speed. Higher concentrations were associated with a speed ranging between 0 and 4 m s -1 , lower concentrations of all pollutants were observed when the wind speed was greater than 12 m s -1 , intuitively, this result indicates that the high wind speed plays a role in "cleaning up" the air over the measurements site area. For NO and NO 2 , polar plot representations clearly show that maximum concentrations were observed for winds blowing from the southwest. The southwest of the site is located near a traffic intersection, gas station, and provincial road, which would indicate very local sources.
Another area with high NO 2 concentration was found in the north, which has a similar situation for BC and PM 10 . It is noted that the site is located in the south of Orléans (distance 6 km from city center) and south west of Paris (distance 90 km), and that the major source of volatile organic compounds during wintertime in the Paris region is wood burning [33]. Hence, impact of biomass burning emissions, especially during the lockdown, was clearly evidenced on NO 2 , BC, and PM 10 atmospheric concentrations.
Concerning the BTEX mixing ratios, the probability to observe high concentrations decreased with the wind speed. Higher concentrations were associated with a speed ranging between 0 and 4 m s −1 , lower concentrations of all pollutants were observed when the wind speed was greater than 12 m s −1 , intuitively, this result indicates that the high wind speed plays a role in "cleaning up" the air over the measurements site area. Atmosphere 2022, 13, x FOR PEER REVIEW 12 of 15

Conclusions
The ambient levels and possible sources of atmospheric BTEX during the lockdown period in Orléans are reported in this study. The variation of the mean BTEX concentration suggests that besides source strength, the seasonal and diurnal variations of atmospheric BTEX in peri-urban areas also strongly depend on meteorological conditions and photochemical activity.
T/B ratios are generally low (<1) at the site during all three periods. In contrast to the observations of Salameh [28], which focused on the Paris area, we find that temperature is not a strong determinant of T/B, which suggests that evaporation rates are not the major control on this ratio in our study, and that photochemical age may be playing a more important role. Previous observations in Parisian suburbs such as Melun [28] find T/B ratios of approximately 1, which is similar to our own observations and suggests that we are observing a more aged air mass in line with the overall findings of the MEGAPOLI campaign [28].
Surprisingly, the overall BTEX mixing ratios were in general at their highest (~ 1 ppb) during the lockdown period, a time when we would expect to have the lowest volumes of traffic, which is supported by the Google Community mobility reports. Some of this effect is likely to be caused by meteorological conditions, where a change in the prevailing wind direction from SW to NE occurred near the onset of lockdown, suggesting that Orléans during this time could be more affected by plumes from the Paris megacity. It is noted that the overall concentrations are lower than those observed in Paris, as would be expected from photochemical losses as well as mixing and diffusion during the transport process. Changes to daily routines that could result in increases of residential heating

Conclusions
The ambient levels and possible sources of atmospheric BTEX during the lockdown period in Orléans are reported in this study. The variation of the mean BTEX concentration suggests that besides source strength, the seasonal and diurnal variations of atmospheric BTEX in peri-urban areas also strongly depend on meteorological conditions and photochemical activity.
T/B ratios are generally low (<1) at the site during all three periods. In contrast to the observations of Salameh [28], which focused on the Paris area, we find that temperature is not a strong determinant of T/B, which suggests that evaporation rates are not the major control on this ratio in our study, and that photochemical age may be playing a more important role. Previous observations in Parisian suburbs such as Melun [28] find T/B ratios of approximately 1, which is similar to our own observations and suggests that we are observing a more aged air mass in line with the overall findings of the MEGAPOLI campaign [28].
Surprisingly, the overall BTEX mixing ratios were in general at their highest (~1 ppb) during the lockdown period, a time when we would expect to have the lowest volumes of traffic, which is supported by the Google Community mobility reports. Some of this effect is likely to be caused by meteorological conditions, where a change in the prevailing wind direction from SW to NE occurred near the onset of lockdown, suggesting that Orléans during this time could be more affected by plumes from the Paris megacity. It is noted that the overall concentrations are lower than those observed in Paris, as would be expected from photochemical losses as well as mixing and diffusion during the transport process. Changes to daily routines that could result in increases of residential heating emissions could also explain elevated BTEX mixing ratios during lockdown, although it is noted that the meteorological conditions during this time were generally warmer than the period that preceded lockdown. This hypothesis is supported on some days by the higher loading of BC and PM 10 observed during lockdown, which could be attributable to a wood smoke source.
In addition to the BTEX measurements, we present a suite of supporting measurements including NO x , O 3 , PM 10 , BC, and meteorological data. In general, we did not observe strong correlations between these parameters and BTEX, suggesting that their respective sources could be different.
Even though we can be sure of changes in emission sources during the lockdown period, we find that the highest ambient pollution levels occurred during a time of national inactivity. Given that BTEX are considered to be dominated by traffic [25], and that local traffic volumes were reduced during the lockdown, we conclude therefore that ambient pollution levels in Orléans are strongly connected with the neighboring Paris region, which may be true of other suburban and peri-urban sites around the globe. Furthermore, if the higher BTEX episodes observed during lockdown are indeed related to more local sources of residential heating, then this implies that emission control of wood burning should be prioritized. From the perspective of pollution mitigation, we recognize that some of these conclusions remain tentative. Larger, more detailed datasets which would allow positive matrix factorization for example, would allow a more definitive source apportionment, and ultimately aid in emission control strategies.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/atmos13010010/s1, Figure S1: the temperature ramp of the capillary column. Figure S2: a series of diurnal variations in benzene, toluene, ethylbenzene, m,p-xylene and o-xylene mixing ratios before, during and after lockdown provided in the supporting information.