Study area

The study area is located in the Orizaba Valley, in the center of Veracruz, Mexico, including the cities of Orizaba and Córdoba (Figure 1). This region is characterized by its high agricultural activity and many industrial sources, as well as the emissions associated with vehicular sources from the 150D highway (one of the most important highways in Mexico). Additionally, the Orizaba Valley is located downwind from important urban and industrial zones, where oil and oil products are the main productive activity. The sampling areas were distributed as follows: sites S1 to S5 were located in the City of Orizaba and sites S6 to S10 were located in the City of Córdoba (Figure 1).

Figure 1 Study area and sampling area locations in the Orizaba Valley, Veracruz, Mexico. 

Sampling and analysis procedure

The experimental design was a simple classification comparison, with three treatments (climatic season: dry, rainy and cold front seasons) and 10 experimental units (sites S1 to S5 in the City of Orizaba and sites S6 to S10 in the City of Córdoba). The response quantitative variables were classified based on four-month and total flows (1 year) of the N and S atmospheric deposit in the Orizaba Valley region. The collector was made of a funnel attached to a PVC tube, connected to a column that contained 30 g of RII (Amberlite^TM IRN 150). The column was sealed using fiberglass at the top (as a filter) and at the bottom (as a support platform). The funnel was covered with a mesh to prevent solid matter (leaves and bugs) from entering. The column was placed inside a PVC tube to protect the resin from direct solar radiation. The bottom of the column was connected to an open PVC standard valve, allowing the flow to drain. The sampling devices were left outdoors. The funnel sent the solution of the collected samples through the RII column, where ions were retained (^{Fenn and Poth, 2004}; ^{Root et al., 2013}) (Figure 2).

Figure 2 Sampling device used for N and S passive collection. 

The sampling was carried out from January 1 to December 31, 2015, and it was divided into three sampling sub-periods (4 months each). This included the three main seasonal periods: cold fronts (nortes), dry, and rainy seasons. At the end of each sampling sub-period, ions were recovered using a KCl 2 N solution and specifically designed devices. RII columns were extracted in two sequences, using 100mL of KCl 2 N. Those extracts were analyzed, in order to determine NO3-, SO42- and NH4+. Control (extracts from non-exposed RII column) was analyzed and the average was subtracted from the average of total recovered ions for each resin column. The kg ha^-1 per sub-period flow in the N deposit (such as NO3-+NH4+) and in the S deposit (such as SO42-) and the kg ha^-1 year^-1 full-period flow (2015) were obtained dividing the mass of solute by the funnel surface area and the exposition period (total and four-monthly, respectively). In order to determine the absorption and recovery efficiency, a lab test was carried out, using six resin columns, prepared with a synthetic solution, with three replicates. The absorption of NO3-, NH4+ and SO42- contained by the synthetic solution was 98 %. The RII columns recovery efficiency ranged from 95.3 % to 96.8 % (NO3-), from 97.6 % to 99.5 % (SO42-), and from 98.2 % to 100 % (NH4+).

The NH4+ ion was determined through a molecular absorption spectrometry using the indophenol blue method (^{Fresenius et al., 1988}). The SO42- ion was quantified using the NMX-AA-074-SCFI-2014 turbidimetric analysis (^{Secretaría de Economía, 2015}). The NO3- ion was analyzed using the NMX-AA-079-SCFI-2001 brucine method (^{Secretaría de Economía, 2000}). In the three cases, a UV HACH (DR2800 model) spectrometer was used.

Meteorological data

In order to trace the origin of air masses during the study, the air mass trajectories were estimated for the 96 previous hours for the selected sites and dates in the Orizaba Valley. The U.S. NOAA’s HYSPLIT model (Hybrid Single Particle Lagrangian Integrated Trajectory Model) was used to carry out this task (https://ready.arl.noaa.gov/HYSPLIT_traj.php).

Flow of de NH4+posits

The flows of NH4+ deposits were higher during the cold front season (Figure 3a). These belonged to sites S7, S10, and S2. Sites S2 and S6 located in rural areas, outside the cities of Orizaba and Córdoba (Figure 3b). The soils of this area can be considered as agricultural soils and, as a result, the extensive use of fertilizers may influence the NH4+ flows during cold front and dry seasons. However, because this area is surrounded by the 150 D highway (Mexico City-Veracruz), vehicular emissions could also contribute to the NH4+ background levels. NH₃ emissions from gasoline-fuelled vehicles were reported by ^{Pierson and Brachaczek (1983)}. Vehicles that produce NO as a result of their catalytic converters will store enough H₂ to reduce NO to NH₃. The highway crosses the Orizaba Valley and the NH₃ emissions might have contributed to NH4+ levels of this region. Additionally, mining and cement industries are established around the City of Orizaba. Particulate matter, NO_x, SO₂, CO, and CO₂ and lower amounts of NH₃ and NH4+ are the primary emissions of cement production. Agro-industrial development and fertilizer production are carried out in the Córdoba region. Sites S7 and S10 are found in the vicinity of those industries and are adjacent to the 150D highway. Those vehicular emissions might also have contributed to the levels in this area. NH4+ levels showed significant differences between sampling stations. This suggests that rainfall conditions the mechanism that removes the reduced N from the atmosphere (such as NH₃ or NH4+). There were not significant differences between sampling sites, indicating that NH4+ levels were evenly distributed. The flows of NH4+ deposits were higher in the sampling sites with soils used for agricultural and industrial purposes. This matches previous discussions about the subject (Figure 3C). Therefore, the relative contribution of sources can be categorized as follows: industrial ˃ agricultural activities ˃ vehicular sources.

Figure 3 Flows of NH4+ deposits (kg ha^-1 per each four-month period) in the Orizaba Valley during 2015: A) per season, B) per sampling site; and C) per soil use. 

Flows of NO3- deposits

NO3- had a seasonal pattern, with higher values during dry season (Figure 4A). This matches the local character of atmospheric NO₂ (the atmospheric residence time of NO₂ is approximately). As a result of their life, NO₂ and NO3- are found in high concentrations near their sources (^{Noone, 2012}). Additionally, significant vehicle emissions of NO_x can be generated due to daily traffic during the rush hours of a city. The NO3- levels obtained were caused by a mixture of local sources. Figures 4 (B) and 4 (C) indicate that the flows of deposits were higher in areas where utilization of soils are industrial and agricultural. Therefore, the relative contribution of sources in this region can be categorized as follows: industrial ˃ agricultural activities ˃ vehicular sources. levels were significantly different between seasons. This indicates that rainfall conditions the way in which NO3- is removed from the atmosphere. There were no significant differences between sampling sites. Therefore, NO3- levels were evenly distributed.

Figure 4 Flows of NO3- deposits ((kg ha^-1 per each four-month period) in the Orizaba Valley during 2015: A) per season, B) per sampling site, and C) per soil use. 

Flows of SO42- deposits

The ion had a seasonal pattern, with higher levels during rainy and cold front seasons (Figure 5A). This behavior matches the SO₂ regional character in the atmosphere (^{Seinfeld and Pandis, 1998}). Since SO42- dry deposit is slow, rainfall essentially removes the atmospheric fraction of SO42-. Junge (1960) reported that the residency time of atmospheric ranges from 2.5 to 5 days; in contrast, ^{Finlayson-Pitts and Pitts (1986)} indicated that it lasts from 60 hours to 2.5 days. The results are consistent with the transport time as a wind speed function during the cold fronts, rainy, and dry seasons: 1, 2-3 d, and 4-5 d, respectively (Figure 5A).

Figure 5 Flows of SO42- deposits ((kg ha^-1 per each four-month period) in the Orizaba Valley during 2015: A) per season, B) per sampling site, and C) per soil use. 

In addition to the large-scale transport of SO₂, industrial sources in the region significantly contributed to pollution (Figure 5B and C). However, background levels of sulfate in the Orizaba Valley influenced the whole region, as a result of large-scale transport (Figure 5A). The relative contribution of regional sources can be categorized as follows: industrial ˃ agricultural activities ˃ vehicular sources.

SO42- levels had significant differences between dry, rainy and cold fronts seasons. Therefore. rainfall and large-scale transport -associated with weather conditions during these two seasons- conditioned the way in which is removed from the atmosphere. In contrast, there were no significant differences betwwn sampling sites. Therefore, SO42- levels were evenly distributed.

Effect of regional emissions transported at large-scale over the S deposit in the Orizaba Valley

Large-scale atmospheric circulation

According to the last Mexican domestic emissions inventory (http:/sinea/semarnat.gob.mx/sinea.php), SO₂ emissions (t year^-1) were 576 247.83, 15 104.43, and 167 448.54, for Campeche, Tabasco, and Veracruz, respectively. Therefore, the three states jointly contributed 33.85 % to the total SO₂ emissions in the country. Campeche contributes 25.71 %. The Coatzacoalcos-Minatitlán industrial corridor -where oil refineries and chemical and petrochemical industries are located- is found east-southeast of the Orizaba Valley. The Secretaría de Medioambiente y Recursos Naturales (SEMARNAT) considers it as an atmospheric pollutant critical zone. Numerous land oil facilities are found in Tabasco -including exploration and production facilities, shipping and storage terminals, and gas processing plants. Additionally, offshore oil and gas exploration and production is carried out in the coast of Campeche. Most of the facilities are located in the Sonda de Campeche and include a sour gas recompression plant (in the Atasta Peninsula), which receives and processes the gas sent from the offshore platform. Therefore, high SO42- levels in the Orizaba Valley can be a consequence of the large-scale transport of the SO₂ emissions released by these upwind sources.

In certain weather scenarios SO₂ emissions are transported to the Orizaba Valley. SO₂ emissions generated in offshore platforms (Figures 6A, B, and C) and in land gas, fuel, oil, and oil products facilities (Figure 6F: Coatzacoalcos-Minatitlán, Atasta and oil fields in Tabasco) affected the Orizaba Valley during this study. During the cold front season, air masses were transported from northern Mexico. As a result, sources from that direction might have contributed to the sulfate levels in the Orizaba Valley (Figures 6 D and E). Thermoelectric plants and oil facilities can be found to the North of the cities of Altamira (Tamaulipas) and Poza Rica (Veracruz). Along with the cities of Tampico and Matamoros (Tamaulipas), Monterrey (Nuevo León), and Brownsville (Texas), they might have contributed additional pollutants. These results match the findings of ^{Kahl et al. (2007)}, who proved a large-scale transport from these two areas to northern Veracruz.

Figure 6 Typical air mass trajectories for the Orizaba Valley (2015). 

^{Kahl et al. (2007)} studied the large-scale transport towards El Tajín. They used weather resources such as global wind field re-analysis, air mass trajectory analysis, and local sources measurement. In this way, the emission and transport of precursor chemicals of acids from potential significant sources were identified, such as offshore oil fields in the south of the Gulf of Mexico and oil facilities in Astata and Dos Bocas, in Campeche and Tabasco. Based on the air mass trajectories during the 5 d since their arrival, ^{Kahl et al. (2007)} established two weather regimes that control the transport to northern Veracruz. One was the eastern flow that dominates the transport during the summer (from June to August), when air mass trajectories go through the Caribbean Sea, crossing the Yucatan Peninsula and the south of the Gulf of Mexico (the Sonda de Campeche, where the offshore platforms are located). The second was a north-northwest flow, associated with the North American anticyclonic circulation. This flow affects the atmosphere during winter or cold front seasons, favoring the transport from central and southern USA -with air mass trajectories that go through Texas, the North of the Gulf of Mexico (where oil facilities can also be found), and northern Mexico- before it arrives to El Tajín. Additionally, ^{Kahl et al. (2007)} reported that, from June to August -when the transport direction during several days (5 d) come from the East-, the air remains at a low altitude (200-600 msnm). Under these conditions, regional sources make a significant contribution to pollution.

The straight distance between El Tajín and the offshore platforms of the Sonda de Campeche (approximately 500 km) can be compared with the distance between the Orizaba Valley and the platforms in the State of Campeche (approximately 450 km East). The estimate transport from the oil platforms in the Sonda de Campeche to the Orizaba Valley takes 1 to 5 days (as a wind speed function). Therefore, both flows probably also affect the air mass transport when they arrive to Orizaba Valley, during rainy and cold fronts seasons (^{Kahl et al., 2007}). Regarding the influence of local phenomena, mountain-valley breeze systems created by regional topography can contribute to the variation of the pollutant concentration between areas. To support this hypothesis, a daily sampling protocol (with a daily resolution) must be applied. During the development of this study, air quality monitoring stations were not available in the Orizaba Valley.

The regional nature of sulfate and its precursors

The SO42- can directly be formed in the atmosphere from SO₂ oxidation, and its oxidation rate determines how long it remains in the atmosphere. It can also originates from H₂SO₄, which, at the same time, is formed from SO_x and produces SO42- particles. However, even in clean atmospheres (zones without urban and industrial sources), significant levels of SO42- particles have been quantified. Therefore, the conclusion is that SO42- levels are related to the atmospheric reactions to anthropogenic SO₂ and that they can be the consequence of regional large-scale transport processes (^{Khoder, 2002}). The dry SO₂ to oxidation rate and the wet oxidation rate in atmospheric conditions is lower than those of NO_x. SO₂ and its oxidative products stay longer in the atmosphere than NO₂. While SO₂ stays from 2 to 5 d in the atmosphere (^{Seinfeld and Pandis, 1998}), NO₂ only stays 1 d. Therefore, compared to NO_x, SO₂ can be transported at a greater distance by the movement of air masses. Consequently, SO42- and its precursor gas are known as regional pollutants, while NO3- is a local pollutant. This fact supports the hypothesis that high levels of SO42- in the Orizaba Valley are originated in relatively distant sources, as a result of the large-scale transport of SO₂ regional emissions during rainy and cold fronts seasons, when the Orizaba Valley was subjected to meso-scale meteorological phenomena. However, quality air monitoring data for Tabasco, Campeche, and Veracruz were not available, and SO₂ concentrations in this region went unrecorded.

Flows of deposit and reference values

Critical loads have been estimated for several regions of the world. Alpine ecosystems -which are more sensitive than lowlands ecosystems- showed a critical load of 5 kg N ha^-1 year^-1 (^{Fenn and Geiser, 2011}), while Nuevo Mexico and California have reported reference values of 3.8 kg N ha^-1 year^-1 and 4-7 kg N ha^-1 year^-1, respectively (^{Grennfelt and Nilsson, 1988}). A critical load of 3 kg S ha^-1 year^-1 and 2-5 kg S ha^-1 year^-1 have been proposed for very sensitive areas and for natural forests, respectively. In Mexico, no critical load values are available and there are very few studies about this subject (mainly focused in pine forests). Atmospheric inputs of 15 kg ha^-1 year^-1 were reported for pine groves in Desierto de los Leones, on the outskirts Mexico City (^{Fenn et al., 2002}), while in Zoquiapan -a place located to the east and upwind from Mexico City- N and S inputs of 5.5 and 8.8 kg ha^-1 year^-1, respectively, were reported (^{Pérez-Suárez et al., 2008}).

The magnitude of the S flows in our study was similar to those found in some temperate mountain regions in Mexico. A research carried out in central Veracruz about several types of vegetation covers and soil uses reports 6-27 kg ha^-1 year^-1 and 2-4 kg ha^-1 year^-1 inputs for S and N, respectively (^{Ponette-González et al., 2010}). Flows of S obtained by ^{Ponette-González et al. (2010)} can be compared with those found in forests around Mexico City that receive from 9 to 20 kg S ha^-1 year^-1 (Fenn et al., 1999). According to ^{Ponette-González et al. (2010)}, certain zones in central Veracruz face an acidification risk and that uplands are downwind from the more important industrial zones in Mexico, where facilities that emit significant amounts of SO₂ are concentrated. Those facilities include thermoelectric plants, gas and oil extraction facilities and refineries, petrochemical plants, paper industry facilities, and sugar refineries. The findings of ^{Ponette-González et al. (2010)} matched those reported by ^{Parungo et al. (1990)}, who observed that upwind anthropogenic sources increase S deposits in mountain ecosystems in Sierra Madre Oriental (State of Veracruz). The results found by ^{Parungo et al. (1990)}, ^{Kahl et al. (2007)}, and ^{Ponette-González et al. (2010)} support the hypothesis established in our research. In our study, the mean flows of deposits estimated for N (as N-NH4++N-NO3-) and for S (as SO42-) in the Orizaba Valley were 1.42 kg N ha^-1 year^-1 and 27.25 kg S ha^-1 year^-1, respectively. The mean flows of N deposits did not surpass the critical load values proposed for sensitive ecosystems. However, the mean flows of S exceeded almost sixfold the threshold limit values proposed for natural forests and sensitive areas. Additionally, the mean flows of S deposits obtained in this study exceeded three times the values reported for Zoquiapan, Estado de Mexico (^{Pérez-Suárez et al., 2008}). The annual mean flow estimated in this study for S deposits (27.25 kg ha^-1 year^-1) can be compared with the maximum value obtained by ^{Ponette-González et al. (2010)} in sites located in agro-forestry systems in central Veracruz. It is important to emphasize that the flows of S deposits in the Orizaba Valley had a seasonal variability: 9.75 kg ha^-1 per each four-month period during dry season and 33.8 and 38.24 kg ha^-1 per each four-month period during rainy and cold front seasons, respectively. Unfortunately, this seasonal behavior could not be compared and discussed with the data obtained by ^{Ponette-González et al. (2010)}, because they did not evaluate seasonality. However, comparing only the flows of S deposits during the dry season (9.75 kg ha^-1 per a four-month period), the resulting values are very similar to those reported for Zoquiapan (^{Pérez-Suárez et al., 2008}).

^{Cerón et al. (2015)} and ^{Cerón et al. (2016)} reported flows of N deposits of 0.8 and 2.15 kg ha^-1 per each four-month period for Astata and Ciudad del Carmen, respectively. Both cities are located in Campeche. These levels are not significantly different from those obtained in our study, suggesting a local origin in all cases. Nevertheless, for S deposit, ^{Cerón et al. (2015)} reported 9.22 and 4.7 kg ha^-1 per each four-month period for Astata and Ciudad del Carmen, respectively. These values are lower than the ones obtained in our study. It is important to mention that Ciudad del Carmen is located upwind from Atasta and that, in turn, Atasta is located upwind from the Orizaba Valley. Both zones are located east of the Orizaba Valley and the more important emission points are located in Campeche. Since SO₂ is a regional pollutant, its deposits do not affect the surroundings of its emission point. As a result of its spatial and temporal variability -as well as its residence time in the atmosphere-, it becomes a sulfate deposit in downwind sites which are relatively distant from its source.

Levels of S tend to increase in areas located west of the Sonda de Campeche (Figure 7). SO₂ stays in the atmosphere approximately 2-5 d; this is enough time for the air masses to cross southeastern and northern Mexico, with a West-Southwest component (during rainy and cold front seasons). Along this trajectory, SO₂ is transformed and deposited as SO42-. Its values are higher in the western-southwestern direction.

Figure 7 Trend of the S deposit in kg ha^-1 year^-1 for Ciudad del Carmen (Campeche), Atasta (Campeche), and the Orizaba Valley, as the air masses that travel from the East-Northeast direction to the West-Southwest direction. 

The studies about the flows of S deposits are scarce in Mexico and there are not enough data to evaluate this variation. Additionally, there are not environmental monitoring networks in southern Mexico that measure meteorological variables -such as speed and direction of the wind, and relative humidity. Therefore, the SO₂ levels in this region remain unknown and they cannot be correlated with the flows of S deposits.