**3.1 Effect of meteorological variables on particulate matter and ozone concentrations**

Multiple linear regression between meteorological variables (temperature, relative humidity RH, wind speed, and wind direction) and PM (10 and 2.5) and ozone was performed to assess the effect of meteorological variables on PM and O3 concentrations. In 90% of the linear regression models, wind speed and wind direction were statistically significant variables, whereas temperature and RH were not always statistically significant. The effect of meteorological variables on PM and on ozone concentrations was analyzed using the quantile regression at 75% between wind speed and pollutant concentration [21].

**Table 1** indicates that there is a negative relation between PM10 in urban areas and wind speed. Similarly, a negative relation was found between PM2.5 in urban areas and wind speed. In semi-urban areas, a positive relation between PM2.5 and wind speed was found. Moreover, in rural and semi-urban areas, O3 was positively related to wind speed. Higher wind speeds in urban areas cause a higher dilution of pollutants in the Planetary Boundary Layer (PBL) decreasing PM concentrations. However, higher winds in rural and semi-urban areas enhance soil erosion, producing dust and increasing PM concentrations.


*Neg: negative coefficient, Pos: positive coefficient, NS: not statistically significant (p < 0.05).*

#### **Table 1.**

*Results of quartile regression between pollutants (PM10, PM2.5, and O3) and wind speed.*

#### **3.2 Pollutant concentration patterns**

#### *3.2.1 Hourly averages*

Hourly mean pollution concentrations for stations in urban, semi-urban, and rural stations are shown in **Figure 2** for June to January and **Figure 3** for February to May.

During the first hours of the day (0:00 to 7:00 LT), the median PM10 concentration in the semi-urban area is lower than that recorded in the rural and urban areas. Between 9:00 h and 15:00 LT, the median PM10 concentration in the urban area increases and exceeds that of PM10 in other regions. At night (18:00–22:00 h LT), the PM10 concentration in the city becomes homogenous. This variability occurs during the dry and wet seasons and is more pronounced during the dry season (**Figure 2**). The hourly behavior of PM2.5 is similar to that of PM10. However, the difference in PM2.5 concentration between the urban area and the semi-urban and rural areas is more pronounced from 9:00 to 12:00 LT (**Figure 3**).

The diurnal variation of ozone is mainly due to photochemical reactions due to solar radiation. Therefore, the ozone concentration increases from 9:00 local time to a maximum of around 15:00 local time. At night (20:00 to 7:00 LT), the ozone concentration is the lowest in the urban area compared to other regions (**Figures 2** and **3**).

**Figure 2.**

*Diurnal cycle of PM10, PM2.5, and O3 concentrations in urban, semi-urban, and rural areas in GMC during the dry season (February-May). Data from 2012 to 2022. O3 concentrations in ppb and PM (10 and 2.5) in μg/m3 .*

*Diurnal cycle of PM10, PM2.5, and O3 concentrations in urban, semi-urban, and rural areas in GMC during the wet season (June-January) for the period 2012–2022. O3 concentrations in ppb and PM (10 and 2.5) in μg/m3 .*

## *3.2.2 Monthly averages of pollution concentrations in GMC*

**Figure 4** shows the monthly average concentrations of pollutants in urban, semiurban, and rural areas of the GMC.

Monthly ozone concentrations increased during the dry season (February-May) in urban, semi-urban, and rural areas during 2012–2022. The rainy season begins in late May in central and southern Mexico. The rain removes pollutants from the atmosphere, resulting in a decrease in tropospheric ozone concentration beginning in June. Monthly ozone concentrations are higher in rural and semi-urban areas than in urban areas (**Figure 4**). This urban-rural gradient has also been observed in other cities around the world and originates from the ratio of NOx to VOCs [22, 23].

Similarly, PM10 concentrations increase during the dry season. However, during this period, PM10 concentrations are higher in semi-urban areas than in rural areas. Furthermore, during winter (October-January), PM10 concentrations do not differ significantly between rural, urban, and semi-urban areas. The seasonal behavior of PM2.5 is similar to that of PM10. The median PM2.5 concentration is 25 μg/m3 from November to May.

#### **3.3 Trend analysis**

The trend analysis shows that O3 in urban areas in the wet season has no trend change in the period 2012–2022. Also, PM2.5 concentrations in the dry season in rural

**Figure 4.** *Monthly ozone (O3), PM10, and PM2.5 concentrations in urban, semi-urban, and rural areas of Mexico City for the period 2012—2022. O3 concentrations in ppb and PM (10 and 2.5) in μg/m3 .*

and semi-urban areas have the same trend in the period 2012–2022. In contrast, PM2.5 concentrations in urban areas in dry and wet seasons changed six and four times in the period 2012–2022, respectively. O3 concentrations in rural areas in dry and wet seasons changed five times in the period 2012–2022.

In 2020, there are significant (p < 0.05) trend changes in PM2.5 concentrations in urban areas during the dry season and O3 concentrations in rural and semi-urban areas during the wet season.

#### **3.4 Concentration difference**

According to the boxplots shown in **Figure 5**, the concentration of PM10 is decreasing in urban, semi-urban, and rural areas compared to previous years. However, the largest decrease was found in rural areas in 2020. A decrease of 38.87% (U = 17,156; p < 5.02e-68) was found in 2020 compared to 2019. PM2.5 concentrations in urban areas decreased on average by 9.7% compared to 2019. PM2.5 in rural and semi-urban areas decreased on average by 11.68% and 14.84%, respectively (U = 17,156; p > 5.02e-68 and U = 58,648; p = 0.002). In contrast, ozone concentrations increased on average by 3% in urban, 14% in rural, and 16% in semiurban stations in 2020 compared to 2019, due to the chemical reaction explained in detail by Peralta et al. [9].

Average increases or changes of PM (10 and 2.5) and O3 were calculated using data from the whole year 2020. In order to assess the immediate effect of a major intervention on PM and ozone concentrations in urban, rural, and semi-urban areas of the
