The methodology of the investigation considered the selection of the study site by the identification of natural, fixed, and mobile sources. The analysis period was also chosen due to the PM₁₀ and PM_2.5 particle observations from the National Air Quality Information System (SINAICA) of the INECC. As mentioned in their work [1], the annual arrival trajectories to Mexico pass through the Caribbean, cross the Yucatan Peninsula, and extend through the Gulf of Mexico. Figure 1 shows the area over the Gulf of Mexico.

Figure 1. Gulf of Mexico and selected states. Source: Authors.

The delimitation of the analysis window of the event considered previous studies conducted in Mexico. These studies indicated that during May to July months, there is historically an influx of Saharan dust reaching Mexico. As a second criterion, satellite imagery analysis was performed and showed that during these periods, the transport of the SD was more intense. Finally, this information was corroborated using PM₁₀ and PM_2.5 concentration records obtained from air quality monitoring stations located in the study area [1][8]. The selection for 2018 and 2019 was considered before the peak event in summer 2020, due to the quantity of AQ data in the SINAICA database.

SD from Africa was modeled using the HYSPLIT system in conjunction with backward trajectory analysis, as well as the altitudes of dust transport from the Sahara Desert to Mexico [32]-[34]. The HYSPLIT configuration considered an ensemble with 3 coordinates derived from the visual analysis of the images of NASA® Worldview, the first starting in the Sahara Desert, Latitude 18.787390 and Longitude −14.001159, a second coordinate in the middle of the Atlantic Ocean, and a third coordinate in the Yucatan Peninsula, Latitude 20.274413 and Longitude −87.733412. The configuration with 15 days of backward trajectories and a level high of 500 m AGL, with 315 hours of analysis, meteorology 0000Z Jun 15 2020 GDAS1. In the Yucatan Peninsula, the configuration considered 200 hours, and in both trajectories, the Model Vertical Velocity was used.

Finally, the methodologies were adapted to analyze meteorological phenomena information from the National Meteorological Service of the National Water Commission (SMN-CONAGUA) and satellite images from the NASA® Worldview website. This assessment allowed the estimation of SD transport. The analysis, estimations, graphs, and evaluation of particle concentrations were made in Excel, with the data downloaded and validated from SINAICA, all of this supported by information from the selected air quality information ground stations over the Gulf of Mexico [15][24]-[29].

The current Mexican air quality index, called the “Air and Health Quality Index” (AHQI), considers various regulations from the Ministries of Health and Environment. The Mexican standard NOM-025-SSA1-2014 establishes the reference frameworks for PM₁₀ and PM_2.5pollutants, which are established in accordance with it, as shown in Table 1.

Table 1. PM₁₀ and PM_2.5 calculation bases.

Source: NOM-172-SEMARNAT-2014.

NOM-172-SEMARNAT-2019 establishes the process to calculate the 12-hour weighted moving average for PM₁₀and PM_2.5. The permissible limit values of the Air and Health Quality Index and the related color are shown in Table 2. The formulas are:

where:

C ¯ = weighted moving average concentration;

N = 12;

Σ = data summation;

C_i = average hourly concentration for hour I;

I = consecutive hour of measurement;

W = weighting factor;

W= weight value;

C_max = maximum hourly average concentration over a 12-hour period;

C_min = minimum hourly average concentration over a 12-hour period.

The calculation was performed in Excel to estimate the concentration in µg/m³ and to evaluate the effect of SD particles on air quality values.

Table 2. Air and health quality index (AHQI), permissible limits of PM₁₀ and PM_2.5.

Source: NOM-172-SEMARNAT-2014.

The work focused on the Gulf of Mexico area, including the south, southeast, and northeast of Mexico, considering the states of Yucatan, Veracruz, and Nuevo Leon, due to their proximity to the arrival zone of the SD, which usually comes from the Caribbean annually. Figure 2 shows the selected cities and states in Mexico.

Figure 2. States of the Gulf of Mexico, the Caribbean, and three state cities, where the SD was evaluated in June 2020. Source: Authors.

The states mentioned above are topographically composed of coastal plains and high mountain areas belonging to the Sierra Madre Oriental. The Sierra Madre Oriental extends parallel to the Gulf of Mexico to the south of the State of Veracruz; possibly, this condition promoted the displacement of SD toward the northeast and north of Mexico.

The cities in the states of Yucatan, Veracruz, and Nuevo Leon were selected according to INEGI as metropolitan zones (MZ), as shown in Table 3. These cities have road, air, and rail transport in constant movement. There are industrial sources such as a refinery in Minatitlan, a gas processing plant, and oil fields in Poza Rica, all constant sources of particles in the state of Veracruz. In the city of Monterrey, there are industrial and transportation sources because it is the largest MZ of the selected cities. The city of Mérida, due to its constant annual arrival of dust, is the only reference for the intrusions of SD in Mexico.

Table 3. Metropolitan areas with selected air quality stations.

Source: Prepared by the authors and adapted with information from SINAICA.

Another issue considered in this work was the forest fires that occur in the areas surrounding the selected metropolitan areas, as indicated in the National Forestry Commission (CONAFOR, Spanish acronym) reports from past years. In the weekly report on forest fires, 120 fires were reported in the state of Veracruz, 44 in the state of Nuevo León, and 26 in the state of Yucatan [35][36].

The database information collected for the analysis of the study area was correlated with forest fire data from the National Council of Forests (CONAFOR, Spanish acronym) and the Secretary of Natural and Marine Resources (SEMARNAT, Spanish acronym) [35][36]. There were six forest fires in the Veracruz state in mountain regions, not in the SD trajectory in EMCAV1 or EMCAV2.

INEGI’s demographic data and pollutant information are from SINAICA. The information was used to assess background particles and the air quality data from ground stations or NASA images. In the prior days, the average ranges registered were from 5.1 to 13.4 mg/m³ in PM_2.5 and from 7.7 to 15.9 mg/m³ in PM₁₀ in Poza Rica. In Minatitlan, the averages in PM₁₀ were from 12.4 to 16.0 mg/m³.

For the analysis of satellite images, photos available from the NASA Worldview site [31] were downloaded from June 20 to 30 and until July 12. Figure 3.

The NOAA HYSPLIT trajectory tracking software [32][33] was used in selected cities during this event. The system is available online at https://www.ready.noaa.gov/HYSPLIT.php. The HYSPLIT options allow choosing between independent points or spatial arrangements for areas under evaluation. For this process, the cities of Poza Rica and Minatitlan in the state of Veracruz were selected. Mérida in Yucatan and Monterrey in Nuevo León, initially considering the starting coordinates of the retro trajectory. For this purpose, a point was chosen on the coast of Africa in the Sahara desert, an intermediate point over the Atlantic Ocean was chosen due to assessment of the satellite images, and a point over the Yucatan Peninsula was selected to see how far the SD arrived in the states of the Gulf of Mexico in June 2020.

To assess the synoptic meteorological conditions in the country during the event, the SMN-CONAGUA, the reports corresponding to forecasts 343, 345, 347, 349, 351, 353, and 355 were reviewed. The Meteorological Bulletins of the Regional Hydrometeorological Center in Veracruz, Mexico, summarize the meteorological systems that affected the area. The documents described meteorological events or conditions that could have driven the transport of SD to the northern areas of Mexican territory. Hence, the cities in the state of Veracruz were selected.

Meteorological parameters, when interacting with air quality parameters, create synergistic effects that increase or decrease pollutant concentrations [22][37]-[41]. These parameters were downloaded from meteorological databases, selecting the variables relative humidity (RH), precipitation (PP), wind speed (WS), and

Figure 3. Monitoring the SD event from June 23-26, 2020, over Mexico by means of satellite images from [33]https://worldview.earthdata.nasa.gov/. (a) The arrival of the SD cloud to the southern region of the island of Cuba on June 23, approaching the Caribbean Sea; (b) Image from June 24, the cloud is seen arriving in the Mexican Caribbean over the coasts of Quintana Roo and Central America; (c) On June 25, the SD cloud is seen crossing the Yucatan Peninsula over the states of Quintana Roo, Yucatan, and Campeche, entering the southern Gulf of Mexico; finally, (d) The dust cloud is observed on June 26 over the northeastern part of Mexico over the states of Veracruz, Tamaulipas, and Nuevo León, as well as in part of the Gulf of Mexico on the south-southeastern coast of the United States.

ambient temperature (TA). Some of the most important variables reviewed in this work were correlated with particles and some events that cause disturbances in the results. For the development of the selected parameters of this analysis, the values of mean TA, maximum TA, minimum TA, accumulated PP, RH, and average WS in the evaluated period were considered.

Another meteorological phenomenon known as El Niño Southern Oscillation (ENSO) annually causes changes in global temperature. The 2020 ENSO could have caused a particularly warm year, with increases in global temperatures above the sea and the air. Hence, it could be a driver of dust transport from the Sahara Desert to the Caribbean and the Gulf of Mexico. Simulations by Zhou et al. [2] describe SST anomalies in the central-eastern equatorial Pacific, including strong anomalous southeasterly winds over the southeastern tropical Pacific and the related subsurface thermal anomalies from 1870 to 2021.

A necessary phase to account for the probable dust particles from the Sahara consists of downloading information from the SINAICA air quality databases of the INECC, considering the download of PM_2.5 and PM₁₀ information in the reference periods chosen due to the availability of information and temporal proximity to the event. The years 2018, 2019, and 2020 were considered during the months of the SD transport events above the Gulf of Mexico. The database is available at https://sinaica.inecc.gob.mx/.

One phase addressing SD involves downloading data from the SINAICA air quality databases of INECC [42], including PM_2.5 and PM₁₀ measurements for selected reference periods, based on data availability and temporal proximity to the event. The analysis considers the years 2018, 2019, and 2020, focusing on the months when SD transport events occurred over the Gulf of Mexico.

NOAA’s HYSPLIT trajectory tracking software [32][33] modeled trajectories with multiple starting points. The results of the dust transport over the Mexican territory during the study period are shown in Figure 4.

During the study period, SMN-CONAGUA reported in forecast bulletins 343, 345, 347, 349, 351, 353, and 355 a low-pressure meteorological event related to tropical waves 6, 7, 8, and cold fronts from the north out of season, which are shown in Figures 5(a)-(d). It was observed that these favored the transport of SD over the coasts of the Gulf of Mexico and the national territory. Table 4 shows that during the event in the cities of the states of Yucatan and Veracruz, a low-pressure system was associated and impelled the particles from the Atlantic Ocean towards Mexican territory. High temperatures were also recorded during this period. These temperatures may have favored transport events due to convection of surrounding air from cold fronts and from the out-of-season north winds that arrived at the Gulf of Mexico coast, and warm temperatures that promoted winds, with dust transport to the incidence area, as observed in Figures 5(a)-(d).

Table 4 summarizes some meteorological parameters of the cities where the study was carried out, due to air quality measurement stations installed, which include measurements of meteorological variables. The parameters of relative humidity (RH), maximum ambient temperature (TA max), minimum (TA min),

Figure 4. HYPLIT trajectories: (a) Transport from the Sahara Desert, and (b) Considering the Yucatan Peninsula during the analysis period in June 2020. Source: Authors.

Table4. Average monthly weather data for selected cities for the month of June 2020.

Source: [43]https://www.meteored.mx/.

and average (TA average), wind speed (WS), and precipitation (PP) are important in this dust particle transport phenomenon. In the day-by-day assessment, these parameters undergo variations that cause a decrease in particle concentration, accelerate or remain suspended, or also allow speed variation of the SD transport in the atmosphere. As mentioned in its forecast by SMN-CONAGUA, high temperatures were observed in the daytime environment on the Veracruz coast.

The assessment of the information began with the search for the monthly periods

Figure 5. Weather forecast for Mexico from June 20 to 26, daily reports in SMN-CONAGUA, available at https://www.uv.mx/personal/ngonzalez/data_ojap/.

during which the Sahara Desert dust transport phenomena had an impact in reference years from 2018 to 2020. The information on air quality and meteorological parameters was downloaded from the SINAICA website. The AQ measuring stations were selected, and the years 2018, 2019, and 2020 were chosen. The selection of the measuring stations considered the probable transport previously estimated with the HYSPLIT® software, which described the trajectories towards some of the comparison points over the Gulf of Mexico. It is important to mention that the existence of AQ monitoring stations in Veracruz and Yucatan is limited compared to the city of Monterrey. Information from the selected receptor cities and AQ stations is shown in Table 5.

The downloaded files in CSV format were reviewed and validated using the procedure of the INECC atmospheric monitoring manuals [44], the NOM NOM-172-SEMARNAT-2014 and NOM-025-SSA1-2014, criteria for the elimination of negative data from the observations, and verification of whether they cover the standard percentage of at least 75% sufficiency of information from the observations of the month.

Table5. Selected air quality monitoring stations and receptor coordinates.

Source: SINAICA.

The results from the day-by-day analysis of the SD transport event show the recorded particle concentrations from the air quality monitoring stations selected for this work. They are presented below in Figures 6(a)-(d).

In Figure 6(a) at the EMCAY1 station corresponding to the city of Mérida, the entry of SD into Mexico and the concentrations in µg/m³ were observed during the evaluation period from June 20 to 30; considering the meteorological conditions of the period, the effect of the low-pressure channel and the southern winds coming from the Caribbean area were observed. In general, this area receives the SD from the Caribbean in the Mexican Republic; concentrations of 40.08 and 52.87 µg/m³ differences on average were observed in the period; there are maximum observations of up to 164.0 µg/m³ on June 22, compared to those of 100.0

Figure 6. Comparison of concentrations of PM_2.5 observations at air quality monitoring stations in the cities of (a) Merida (EMCAY1), (b) Minatitlan (EMCAV2), (c) Poza Rica (EMCAV1), and (d) Monterrey (EMCAM1), 2018, 2019 and 2020.

µg/m³ in 2019. In general, in 2019, it did not exceed 100.0 µg/m³ of PM_2.5, unlike in 2020, when in multiple observations it exceeded 100.0 µg/m³. As mentioned in previous subtopics, factors such as forest fires, vehicular traffic, and proximity to the maritime zone generated various types of particles that may affect the background observations of the area not considered in this work.

Figure 6(b), which corresponds to the EMCAV2 station, shows average observations from 2018 to 2020 of 20.33, 23.65, and 25.33 µg/m³, respectively; maximum values of 101.0, 93.0, and 112.1 µg/m³ of PM_2.5 concentration in 2018, 2019, and 2020; and 80% of the values for 2019 and 2018 were less than 80.0 µg/m³. This station is in an area where particulate matter typically comes from the industrial zone of “Pemex” and petrochemical companies. It is situated in a metropolitan area with aerosol presence due to its proximity to the Gulf of Mexico coast. As observed in the Hysplit backward trajectory analysis, this area also forms a current crossing the Mexican territory from the Gulf of Mexico to the Pacific Ocean through the Isthmus of Tehuantepec in Oaxaca. This likely facilitated the transport of some particles from the Gulf through this region. However, at altitudes below 4000 meters above ground level (AGL), the trajectory moves toward the northern part of the state, which is associated with the images from the SMN-CONAGUA forecast, as shown in Figure 4 and Figure 5.

Figure 6(c), corresponding to the EMCAV1 station, clearly shows an increase in PM_2.5 particles compared to the years 2018 and 2019; values less than 40.0 µg/m³ were observed in the years 2018 and 2019, compared to the year 2020, which presents values higher than 60.0 µg/m³, reaching maximun of up to 110.0 µg/m³. In this station, during the analysis period, a meteorological system was detected with high temperatures and a cyclonic system that promoted transport over the MZ of Poza Rica during the period from June 26.0 to 30.0 in the period. The maximun concentration values between 105.0 and 110.0 µg/m³, as shown in the framed area. Within the MZ of Poza Rica, “Petroleos Mexicanos” burns gases from the cryogenic plant’s gas processing process in a pit burner for long periods of the year, generating concentration value peak readings in the same month of the analysis, but which, in contrast to other years, did not exceed the values reached with dust from the Sahara.

Finally, in Figure 6(d) EMCAM1 station, a decrease in PM_2.5and concentration readings was observed, compared to 2020 readings; these were below those in 2019, with lower concentrations than 20.0 µg/m³. This data could be due to a low-pressure channel in southern, as seen in Figure 5 and Table 1. Texas, with data precipitation, and a jet stream rotating above the Gulf of Mexico, promoting the dispersion of particles as seen in Figure 5.

The movement of SD from the Yucatan Peninsula, whose PM_2.5 concentration values are usually similar, is notable since it is the access point for this annual transport. It also shows how the particles moved towards the state of Veracruz through the Coatzacoalcos-Minatitlan area, and upon their arrival in the northern area in the MZ of Poza Rica, they remained for a few days. The observed PM_2.5 concentrations continued their way and entered the northern states, such as Nuevo León. It was observed that they did not arrive in the same concentration; this could have been due to meteorological phenomena coming from the south of the U.S. and associated with the Gulf of Mexico. Figure 7(a) shows that in 2019 and 2020, there were differences between the average PM_2.5 concentrations of 2.0 µg/m³.

It was observed that despite being a city with non-industrial activity, maximum values were exceeded in both years. In 2020, values up to 164.0 µg/m³ were reached. The differences were not significant between the two periods, as seen in the graph, which suggests the recurrent annual entry of SD into the peninsula. Table 6 shows the station’s statistics.

In Figure 7(b), the same behavior was observed in 2020, with higher PM_2.5 concentration values in the maximum and average values for the years 2018, 2019, and 2020, respectively: 20.0, 23.0, and 24.0 µg/m³ average. There is a significant difference; the EMCAV2 presents the increasing behavior of PM_2.5, with values corresponding to 1.0 to 4.0 µg/m³. This represents an estimated 29% increase when considering the observations of the year 2018 and 7% compared to the year 2019. See Table 6 for the station statistics calculated in Excel.

Figure 7(c) of EMCAV1 records and shows, according to the observations in Figure 5, that the SD occurred between June 25 and 28, reaching higher values in June 2020 as seen in Figure 7(c); the percentage differential values were estimated between 9% and 14% in the 2018 and 2019 concentrations respectively, with very similar mean values but with very large standard deviations in 2018; it is observed that there were a couple of events before the entry of SD, and these values may be due to a particle emission process that is recurrently carried out by the gas processing

Figure 7. Comparisons of PM2.5 observations for June in 2018, 2019, and 2020 at SEMARNAT air quality stations in the states of Yucatan (EMCAY1), Veracruz (EMCAV1 AND 2), and Nuevo León (EMCAM1).

complex that emits BC particles or to some unidentified factor. It also shows how, before and after the event, concentration changes abruptly. Table 6 shows the station’s statistics.

Figure 7(d) of the EMCAM1 station shows concentrations lower than those in 2019. This, as discussed above, could have been due to the meteorological event described above, which reduced PM_2.5 concentration in the air and promoted its dispersion. See Table 6.

Table 6. Summary of concentration statistics at stations in Yucatan, Veracruz, and Nuevo Leon.

As can be seen, the PM_2.5 concentration values between the stations reviewed show how, as the SD was transported through the states used to monitor the phenomenon in the day-to-day analysis, the PM_2.5 concentrations increased and consequently affected air quality in its path, especially in the states of Yucatan and Veracruz. Table 7 shows how the index value changed at station EMCAV1 as it passed through the cities of Poza Rica and Minatitlan.

Table 7 shows the average daily values of PM₁₀ and PM_2.5 in the MZ of Minatitlan and Poza Rica, where it can be observed how, within the study interval from June 25 to 30 and the first 5 days of July, the average concentrations of these parameters increase in such a way that they modify the values of air quality in the comparative cities in the state of Veracruz.

Table 7. Average daily concentrations of PM₁₀ and PM_2.5 in µg/m³ in two cities in the state of Veracruz with SEMARNAT air quality monitoring systems.

Note: (1) and (3) ZM Poza Rica, (2) ZM Minatitlán. Source: INECC SINAICA data extracted in 2022.

Figure 8 shows the average daily concentrations in the time series from June 1 to July 4, 2020.

Figure 8. Time series of the average daily concentration of PM_2.5 and PM₁₀ from June 1 to July 5, 2020, during the intrusion of SD in the study area. Source: Authors.