Assessing the Impact of Sulfur Atmospheric Deposition on Terrestrial Ecosystems Close to an Industrial Corridor in the Southeast of Mexico

The main objective of this research work was to diagnose the vulnerability of terrestrial ecosystems to S deposition in Atasta region in Campeche State, Mexico, comprising two simultaneous sampling programs in both, soil and atmospheric deposition on an annual basis during three climatic periods: dry, rainy and cold fronts seasons. From the estimation of soil properties estimation (pH, texture, mineralogy, cationic exchange capacity, and basis saturation %), critical loads and sensitivity classes were assigned to sampled soils based according to the empirical methodology proposed by UNECE. During the dry season, 10 sites fell into sensitivity class 2 (moderately sensitive) and 3 (sensitive). On the other hand, during the rainy season, 8 sites showed a sensitivity class 1 (highly sensitive) and 2 sites presented a sensitivity class 2 (moderately sensitive); whereas along cold fronts season, 12 sites fell into sensitivity class 1 that corresponds to highly sensitive. Sensitivity classes showed a seasonal trend, with a higher sensitivity during rainy and cold fronts seasons; this agrees with the kind of sources influencing on the study area as a result of the prevailing meteorology during these climatic periods. Likewise, S concentration in atmospheric deposition was determined by turbidimetric method, and S deposition fluxes were estimated from surface area of the funnel opening of the sampling device and the sampling period. S deposition fluxes ranged from 0.29 and 14.06 kg S ha⋅yr; with a mean value of 8.57 kg S ha⋅yr. From How to cite this paper: Cerón, R.M., Cerón, J.G., Muriel, M., Rangel, M., Lara, R.C., Tejero, B., Uc, M.P. and Rodríguez, A. (2017) Assessing the Impact of Sulfur Atmospheric Deposition on Terrestrial Ecosystems Close to an Industrial Corridor in the Southeast of Mexico. Journal of Environmental Protection, 8, 1158-1177. https://doi.org/10.4236/jep.2017.810073 Received: August 21, 2017 Accepted: September 23, 2017 Published: September 26, 2017 Copyright © 2017 by authors and Scientific Research Publishing Inc. This work is licensed under the Creative Commons Attribution International License (CC BY 4.0). http://creativecommons.org/licenses/by/4.0/ Open Access


Introduction
The qualitative and quantitative assessing of atmospheric deposition is essential to establish a baseline or estimate the background levels present in a given site, to propose critical load values, to assess the vulnerability of ecosystems, to identify sensitive zones with a potential to exceed critical loads, to assess seasonal and inter-annual trends, and to identify spatial and temporal distribution patterns [1].Based on this, it is possible to propose protection and conservation measures, and finally, in a medium term, to demonstrate the efficacy of environmental policies applied in matter of emissions reduction [2] [3].
N and S deposition fluxes have reached levels that cause alterations in the structure and function of several ecosystems.Therefore, the potential ecological effects derived from atmospheric deposition, constitute one of the main concerns of society [4].The detailed study of this phenomenon, as well as the assessment of its spatial and temporal distribution in a given region, allows decision-makers develop public policies focused to preserve and protect natural resources that could be vulnerable to suffer damages as a result of atmospheric deposition.In this regard, it is necessary to take in account, that ecosystems differ in their sensitivity to acid deposition depending on the buffer capacity of the soils, of soil composition, and of the sensitivity of organisms that inhabit them to the changes in solution or chemical reactions occurring in them [5].
An approach that allows to quantify the amount of pollutants deposition that can damage ecosystems is the estimation of critical loads.Critical loads constitute quantitative estimations about sensitivity of the ecosystems.These values were developed as a tool in 80's and 90's decades through Nordic Cooperation, being obtained in Europe and United States of America to investigate the potential impact of acid deposition in different ecosystems.A critical load can be defined as the quantitative estimation of exposure to one or more pollutants, below which, significant detrimental effects for the environment do not occur [6].
In general, critical loads can be estimated at three levels of complexity: level 0, level 1, and level 2. At "level 0", critical loads are empirically assigned to soils with different sensitivity using existing data.A critical load categorized as "level Journal of Environmental Protection 1" corresponds to a method using a mass balance calculation as steady state models (SMB model); whereas, a critical load with a "level 2" is derived from methods that use dynamic models that requires detailed information [7].
In the case of Mexico, methods "level 1" and "level 2" to derive critical loads are not useful since, empirical loads are not available at a national scale, and due to the limited information of data; therefore, a "level 0" approach must be applied as a first approximation to estimate critical loads.
Besides this, to establish critical load values, and to estimate their exceedances in a given site, atmospheric deposition measurements are required [2] [8] [9].However, to determine the current inputs of nitrogen and sulfur to the ecosystems, the most of deposition maps are based on results of tridimensional chemical transport models that require to be validated by comparison with field measurements.This kind of study constitutes a valuable opportunity for decision-makers to identify potential impacts associated with different emission sources and, to assess their geographical and temporal distribution.The mapping of sensitivity and the critical load approach, constitute methods that allow to analyze these risks.
Additionally, several methods have been developed to calculate and to map critical loads not only at a regional scale, even at national and global scale [6].
The global assessments reported in other research works include maps and data, that give a preliminary diagnosis of the sensitive areas to acid deposition in regions of Europe, Canada, United Sates, China, Japan and Australia [1] [9] [10] [11].However, unfortunately, in relation to another places in the world, as the case of developing countries, there is not enough information [2] [3] [12].Particularly, in Mexico, excepting some threshold values reported for pine forests [9] [13] [14], there are not any data or maps available that show N and S deposition fluxes distribution for vulnerable zones.Since, reference values are not available for Mexican ecosystems, the few studies that have been done in this regard, usually compare the N and S current deposition fluxes with critical load values reported for Europe and United States.It is relevant to investigate in which extent, the empirical critical load approach can be applied in developing countries with ecosystems and climate very different, and where required data are scarce or are not available [2].
However, it is necessary to consider that deposition patterns, the biodiversity and the response of sensitive organisms in tropical countries as Mexico, can be different to those existing in temperate regions in mid-latitudes.Considering the difficulties in the application of this methodology in areas outside Europe, this work applied an empirical method to diagnose the sensitivity of terrestrial ecosystems to S atmospheric deposition in the Atasta region, by estimating a critical load "level 0", by comparing with the estimated current deposition fluxes to calculate their exceedances.Finally, the findings were mapped to identify critical zones, and the role of mangrove vegetation in attenuation process of these effects was discussed.

Study Area
This study was carried out in the region of Atasta in Campeche State, Mexico (Figure 1).Atasta is located in the western part of Terminos Lagoon, that comprises three communities (Xicalango, Puerto Rico and Atasta) and two important ecological regions (Terminos Lagoon and the Lagoon system named Pom-Atasta).
Atasta region is a coastal plain that runs parallel to the coastal line, located in Atasta peninsula, with an approximate surface of 190 km 2 , a length of 50 km, and average depth of 2.7 m.The main characteristics in this system are highly dependent on geomorphology, the tide influence, the exchange of water mass with Terminos Lagoon, and the influence of climatic phenomena.Dominant vegetation is mangrove.
Climate in this region is sub-humid warm with rains occurring along summer and autumn seasons.The annual average rainfall is 1300 mm and the annual average temperature is 27˚C.Prevailing winds come from NE from March to October, when this region is under the influence of cold fronts named "Nortes", and from SE during the rest of the year, when the region is influenced by tropical maritime air as a result of trade winds.In addition, this region is influenced by earth and sea breezes as a result of differential heating between land and sea.This area constitutes a hotspot in Mexico due to the following reasons: 1) its proximity to offshore platforms area in Campeche Sound; and 2) its proximity to Natural protected named "Terminos Lagoon".

Sampling
A multiple transects design with 13 sampling points (P1-P13) was utilized.Sampling points location is presented in Figure 1.Sampling period included three

Soil Sampling
A Transects were constituted by two quadrants separated one meter away.Each quadrant was composed by 4 points with 0.40 m of separation.Sampling was carried out by using a core sampler with a volume of 193.3 cm 3 at 0.30 m depth, obtaining a composite sample for each sampling point for three climatic period.
Composite samples were dried at ambient temperature (25˚C) during 15 days, grounded and sieved (500 μm).Finally, dried samples were stored in sealed and tagged plastic bags until their analysis.

S Deposition Fluxes Sampling
Automatic deposition collectors are expensive and require to fulfill with specific requirements for their installation and operation, whereas passive samplers as through fall devices due to their low cost allow to increase the density of the sampling grid to assess spatial and temporal patterns in a given site.For this reason, in this research work, sampling devices used to collect S atmospheric deposition fluxes were those reported by [5] [15].These sampling devices are based in exchange resin columns with a mixed bed (Amberlite IRN 150) to collect specific ions that are retained along the resin column.This resin column is attached to a funnel (as collection surface), and has glass fiber at the top (as a filter to avoid the fall of leaves or insects) and at the bottom (as a resin support).
Once, the hydrologic flux is collected by the funnel, it is channeled through the resin column, where ions are retained.At the end of each sampling period (climatic season), ions retained (S as sulfate) were extracted by using extraction solutions to recover them (2 N KCl solution) and analyzed by turbidimetric method.

S Atmospheric Deposition Fluxes
S deposition was estimated as sulfate, therefore, once extracted, atmospheric deposition samples obtained from passive sampling devices (collectors type through fall), were analyzed by turbidimetric method to obtain sulfate concentrations [17].Extraction efficiency was calculated as the percentage of loaded ions on the column to the recovered in sequential extractions, obtaining an effi-ciency of approximately of 96.8%.Finally, the surface area of the funnel opening and the sampling period were used to obtain S deposition fluxes for each sampling site as S Kg per land area per year (kg S ha −1 ⋅yr −1 ).

Skokloster Method
To assess the vulnerability of terrestrial ecosystems in Atasta region to S atmospheric deposition, the empirical method proposed by Kulylenstierna et al. [2] based on Skokloster methodology was applied.According to Skokloster workshop [18], the rate of chemical weathering of minerals is the main factor that determines critical loads for forest soils.
From this, soil materials are divided in five classes basing on dominant mineralogy.Consequently, critical load ranges are assigned to these classes, considering cationic exchange capacity (CIC) and saturation percentage of bases (% SB).Kuyenstierna et al. [2] modified Skokloster method to use CIC and % SB to assign sensitivity classes (Table 1), and as consequence, to assign critical load values for S (Table 2).This same methodology has been used by Duan et al. [19] to map critical loads of deposition level "0" for soils in China.
This method assesses the relative sensitivity by using representative variables of mineral weathering of soils, many of which are difficult to measure, calculate or map.Saturation percentage of bases reflects the weathering rate of soils, and it is a variable easy to measure, that constitutes the net result between the gain and the loss of basic ions from soils.As CIC and % SB increases, sensitivity of ecosystems to S atmospheric deposition decreases.

Data Interpolation and Mapping
Point measurements obtained in this work were interpolated by using kriging technique, therefore, to obtain S deposition fluxes maps, a geostatistical procedure was used to obtain field measurements into a continuous pattern (isolines) in the sampling grid of the study area (SURFER v.10.0).Once isolines were generated, data deposition was mapped for each climatic season to assess their spatial and temporal distribution.In addition, sensitivity classes were mapped for each climatic season to identify vulnerable zones.Garrido [20] as a function of CIC and % SB (Table 3, Table 4).

Results and Discussion
CIC values found during dry season ranged from 8.78 to 23.25 meq/100 g, with the lowest value (8.78 meq/100 g) for site labeled as "P13" that corresponds to a soil classified as Gleysol, which is confirmed by permanently saturated soils in water (Table 5).On the other hand, the site that showed a greater value for CIC (23.25 meq/100 g) was the site labeled as "P10" that corresponds to a soil classified as Solonchak, characterized by presenting salts accumulation and not have a good drainage system (Table 5).However, in general, soils in the study area during the dry season showed a CIC from low to medium, so that, can be considered as soils from medium to poor with a requirement of organic matter, excepting by site "P13" that could be considered as very poor with a significant  deficiency in organic matter.
With respect to bases saturation (% SB), their values were between 18% and 88%.Site "P12" presented the highest value, whereas site "P10" has the lowest values (Table 5).Both sites correspond to soils classified as Solonchak; with salts accumulation, a bad drainage and a clay content below of 35%.It could explain the reason why site "P10" presented the highest value of CIC, attributed to the salts accumulation process.Seven sampling points presented values below of 50%, suggesting that, soils in the study area have important deficiencies in nutrients.The remaining six sites, presented values between 50% and 90%, suggesting that they are medium soils whose wealth depend on total CIC values.
On the other hand, from Table 5, it can be observed that eight sampling sites reached sensitivity classes between 2 and 3, that correspond to sensitivity categories of moderately sensitive and sensitive, respectively.Site "P10" obtained a sensitivity class of 1 (highly sensitive); whereas sampling sites labeled as "P11" and "P12", reached sensitivity classes of 4 and 5 (Poorly Sensitive and Not Sensitive, respectively).Site "P10" is an industrial site with a wet and sandy soil, the high content of humidity in this site results in a low % SB (18%), typical of soils classified as Solonchak, with a bad drainage, causing that exchangeable ions be leached from the soil, resulting in a low buffering capacity and a high sensitivity class (Class 1).
Otherwise, sites "P11" and "P12" are industrial sites with a low sensitivity although they are located near to a gas sour re-compression plant, since they have sandy and compact soils with a low content of humidity, result in % SB values between 60.55% and 88%, enhancing the buffering capacity and resulting in soils poorly sensitive to S atmospheric deposition during this seasonal period.

Rainy Season
CIC values during the rainy season ranged from 0.06 to 22.07 meq/100 g, with the lowest values in sites "P1" and "P10", respectively; that corresponds to soils type Arenosol and Solonchak.On the other hand, the site which presented the highest value for CIC was the site labeled a "P6" (22.07 meq/100 gr), with a soil classified as Arenosol, characterized by being depth and have a high salts concentration (Table 6).
In sensitive), three sites had a sensitivity class 2 (moderately sensitive), two sites presented sensitivity class 3 (sensitive) and only one site did not show sensitivity (Table 6).Along this season, the following sites exhibited high sensitivities: P2, P3, P4, P6, P7, P8 y P12, with % SB values very low because of leaching of interchangeable cations, which is characteristic at this time of year.
Besides the characteristics of the soil, the vulnerability can be exacerbated by the potential impact of regional emissions of SO 2 that is long-transported from distant sources during rainy and cold fronts seasons.Both, sulfate ion and its precursor (SO 2 ), are considered as regional pollutants.For this reason, when S atmospheric deposition fluxes are measured in sites potentially impacted by regional emissions, it is difficult to find significant differences between sites, since the most of measured sulfate comes from regional sources located upwind, whose transport is more significant during these seasons in comparison with the dry season, due to prevailing meteorological phenomena that favors this transport process.Finally, site labeled as "P5" was not sensitive with a sensitivity class of 5, and corresponds to mangrove soils.It suggests that mangrove vegetation could play an important role in the attenuation of ecological effects of S deposition.

Cold Fronts Season
CIC values during cold front season were between 9.08 and 25.45 meq/100 g (Table 7), with the highest values for sampling site labeled as "P13", that corresponds to soils type Solonchak, characterized by salts accumulation.
On the other hand, sampling site "P1" presented the lowest values of CIC (9.08 meq/100 g), this sampling site has a soil type Arenosol, (sandy and depth),  9).Sampling site labeled as "P5" was less sensitive, and corresponds to a soil with mangrove vegetation, highlighting the important role that mangrove ecosystem plays in the attenuation process of effects attributed to atmospheric Deposition.

Cold Fronts Season
Eleven of a total of thirteen sampling sites showed exceedance percentages major than 50%.Site labeled as "P13" showed an exceedance of 22.93%, and only one site (site P5) did not show exceedance (Table 10).In comparison with dry and rainy seasons, exceedance percentages obtained along this season increased progressively.Likewise, sampling site labeled as "P5" was the less sensitive, which corresponds to mangrove ecosystem.

Mapping Sensitivity Class and S Deposition Fluxes
Once sensitivity classes were obtained, software SURFER v.10.0 was applied to obtain maps showing both, sensitivity classes and S deposition fluxes, to identify spatial trends, temporal patterns and critical zones with a high vulnerability.

Dry Season
In Figure 2, it is possible to observe that sensitivity did not show any spatial trend, and it is related with inherent soil characteristics that prevail in each sampling site.Site "P10" which presents a vulnerability class of "1" (highly sensitive) corresponds to sour gas recompression plant, suggesting that, the observed vulnerability is more influenced by local sources and disturbance degree in the studied soils.In Figure 3, it is possible to identify the eastern edge of Atasta Peninsula and the surroundings of sour gas recompression plant as critical zones,  presenting the highest S deposition fluxes.

Rainy Season
From Figure 4, it can be observed that sensitivity showed a slight spatial variation, with higher sensitivity classes in the Eastern edge of Atasta Peninsula, since this region is more exposed to air-masses coming from East as a result of meteorological conditions prevailing along this climatic season.In addition, a sensitivity class of "1" was found in the surroundings of sour-gas recompression plant.From Figure 5, it can be observed that higher S deposition fluxes were found in the surroundings of sour gas recompression plant (site P10), just where sensitivity classes were higher.

Cold Fronts Season
Sensitivity showed a complete uniformity, with a sensitivity class of "1" (highly sensitive) in most sampling sites (Figure 6), suggesting a regional influence as    In addition, S deposition fluxes were higher in those sampling sites which presented a sensitivity class of "1" (highly sensitive).It suggests that, there may be two factors acting in synergy to exacerbate the potential effects on ecosystems of the region: soil characteristics and exposition pattern to regional pollutants during this season.

Conclusion
Results obtained in this study constituted the first approximation in the Southern of Mexico to establish critical loads level "0", giving a preliminary assessment about the vulnerability based on criteria proposed by Kuylenstierna et al. [2].It can be concluded that sensitivity of terrestrial ecosystems in Atasta-Xicalango region has a seasonal component well defined, with the following trend: Cold Fronts > Rainy > Dry.This same seasonal trend was obtained from S deposition fluxes maps, with the highest fluxes when the studied zone was subject to the influence of meteorological phenomena occurring at mesoscale.During the dry season, regional contribution was minimal, limited to background levels with a local origin.Both factors, soil characteristics and prevailing meteorology act in synergy to intensify the potential ecological effects associated to S Atmospheric deposition.However, it was demonstrated that soils with mangrove vegetation present a minor sensitivity, highlighting the importance of conserving this kind of ecosystems in this region.

Figure 2 .
Figure 2. Sensitivity Class Map for the studied area during the dry season.

Figure 3 .
Figure 3. S Deposition Fluxes Map for the studied area during the dry season.

Figure 4 .
Figure 4. Sensitivity Class Map for the studied area during the rainy season.

Figure 5 .
Figure 5. S Deposition Fluxes Map for the studied area during the rainy season.

Figure 6 .
Figure 6.Sensitivity Class Map for the studied area during the cold fronts season.

Figure 7 .
Figure 7. S Deposition Fluxes Map for the studied area during the cold fronts season.

Table 2 .
Critical load relative to sensitivity class assigned.

3.1. Sensitivity Class Assignment 3.1.1. Dry Season
CIC levels in soils show the ability to retain cations, availability and amount of nutrients for plants.Soils with low values of CIC are poor in organic matter and indicate a low availability to retain nutrients.This capacity of soils allows to retain the necessary elements to give nutrients to plants, which otherwise would be in solution readily available for their leaching.Therefore, as CIC increases, the natural fertility increases.Soils are classified according to criteria established by

Table 3 .
Soil classification as a function of CIC.

Table 4 .
Soil classification as a function of % SB.

Table 5 .
Sensitivity classes assigned to sampling sites during dry season.

Table 6 .
general, soils in the study area during rainy season presented CIC values Sensitivity classes assigned to sampling sites during rainy season.
It is necessary to consider that site "P5" is type Solonchak, with a high salts content, reflected by the highest CIC values found.On the other hand, site "P4" corresponds to a soil classified as Arenosol, characterized by being depth and have a high sand content.For this reason, during rainy season, the high grade of leaching resulted in % SB values very low.Eight sampling sites of a total of thirteen, presented values below of 50%; four sites showed % SB values between 50 and 90%; and only one site had values greater than 90%; suggesting that they have soils with deficiencies in nutrients and could be vulnerable or sensitive to S atmospheric deposition, as is confirmed in the following section.In comparison with those values obtained for the dry season, CIC and % SB values during rainy season decreased progressively.Seven sites presented a sensitivity class 1 (highly

Table 7 .
Sensitivity classes assigned to sampling sites during Cold Fronts season.

Table 8 .
Critical loads and exceedances for dry season.

Table 9 .
Critical loads and exceedances for rainy season.

Table 10 .
Critical loads and exceedances for cold fronts season.