STUDY AREA

The atmospheric monitoring was carried out in seven cities: Esteio (ES), Triunfo (TR), Charqueadas (CH), Montenegro (MO), Santo Antônio da Patrulha (SA), Caraá (CA) and Maquiné (MA), all located in the northeastern region of the state of Rio Grande do Sul, southern Brazil (Figure 1).

Figure 1 Location of the seven monitored cities (CA, MA, SA, CH, TR, ES and MO) and the city where lichen samples of Parmotrema tinctorum were collected (TA) to be later exposed to monitor air pollution. Abbreviations: CA = Caraá; MA = Maquiné; SA = Santo Antônio da Patrulha; CH = Charqueadas; TR = Triunfo; ES = Esteio; MO = Montenegro; TA = Tapes. 

These cities have great importance for the state economy hosting many industries (Esteio, Charqueadas and Montenegro), a Petrochemical Complex (Triunfo), and an important agricultural area of family farming (Santo Antônio da Patrulha, Maquiné and Caraá). A description of the main industrial activities, the demographic density and the rural practices of the monitored cities are detailed in ^{Koch et al. (2016)}. They also have a similar climate, elevation, and some of them (Charqueadas, Esteio and Santo Antônio da Patrulha) have air-monitoring systems with automatic and semi-automatic air samplers. The elevation of sampling areas ranges from 10 to 75 m above sea level and the climate is subtropical humid, Cfa type according to the updated Köppen–Geiger classification (^{Peel et al. 2007}).

SAMPLE PREPARATION AND LICHEN EXPOSURE TREATMENTS

The foliose epiphytic lichen Parmotrema tinctorum (Despr. ex Nyl.) Hale was chosen as the biomonitor species. This species has a known tolerance to air pollution, having been used in other monitoring studies in southern Brazil (^{Käffer et al. 2012}) and is abundant throughout the country (^{Spielmann and Marcelli 2009}). The lichen P. tinctorum is naturally found in all the studied cities, but we chose to transplant the thalli from a less polluted site to standardize the exposure and due to its low abundance in some sites. Samples of P. tinctorum were collected mainly on trunks of planted Eucalyptus spp. in Tapes (Figure 1) to be later transplanted into potentially polluted sites. The municipality of Tapes is located at least 80 km distant from large urban centers and thus far from important sources of air pollution, being agriculture and cattle raising the main economic sources of the region. Before being exposed, the thalli of P. tinctorum were left in the laboratory for acclimation during ten days (^{Käffer et al. 2012}).

Lichens were placed in 55 x 35 cm synthetic nettingsin standardized structures ca. 120 cm from the ground and shaded to reduce sunlight exposure (Figure 2). They were placed vertically, which was the same position they were found on the trees. When available, lichen transplants were set close to existing automatic/semi-automatic air samplers. Before exposing the samples of P. tinctorum, we performed chemical control analyses to test for metal and sulphur contents. Prior physiological damages were also assessed in some samples of these thalli.

Figure 2 a. Thalli of Parmotrema tinctorum exposed in the standardized structures. b. Detail of P. tinctorum in the synthetic nettings after three months of exposure. 

Transplants were exposed from March to October 2013 and from April to November 2014. Sub-samples of these materials were taken after two, five and seven months of exposure, in all seven cities monitored, in both years. We gathered data for 14 months and six sub-samples in each city. Maquiné and Caraá were the control sites, based on the characteristics of low urbanization and the absence of large industries in these municipalities. Maquiné was only monitored in 2014, after Caraá, supposed to have lower air pollution, showed some important pollutants concentration in 2013.

PHYSIOLOGICAL ANALYSES

Each of the six lichen sub-samples were tested for physiological responses to atmospheric pollution, and analyzed prior to exposure, for control. All pieces of bark were removed and the lichen thalli were cleaned with a brush. Part of the thalli was macerated and the percentages of live, dead and plasmolysed algal cells were counted under the microscope using the neutral-red test at 5% (^{Le Blanc 1971}, ^{Calvelo and Liberatore 2004}). With the same macerate samples we analyzed Chlorophyll a and b according to the method proposed by ^{Boonpragob (2002)}, which consisted of test tubes with 10 mL of 96% ethanol and 5 cm² of lichens. Absorbance in the wavelengths of 649 and 665 nm was measured with spectrophotometer Digimed DME-21. Chlorophyll contents and the percentage of live, dead and plasmolysed algal cells are methods for analyzing physiological damages on the lichen photobiont. Organic carbon content, which measures physiological damage on the lichen mycobiont, was analyzed through wet combustion using the Walkley-Black method with external heat (^{APHA 2005}), on the Laboratory of Soil Analyses at Universidade Federal do Rio Grande do Sul.

CHEMICAL ANALYSES

We tested total content of the following pollutants in the lichen thalli: sulphur (S), copper (Cu), zinc (Zn), iron (Fe), manganese (Mn), chromium (Cr), nickel (Ni), lead (Pb), vanadium (V) and aluminum (Al). Analyses were made through inductively coupled plasma optical emission spectrometry (ICP-OES) (^{Boss and Fredeen 2004}) with previous acid digestion. Samples were dried before being macerated. Chemical analyses were made in the Laboratory of Soil Analyses at Universidade Federal do Rio Grande do Sul and were validated with the use of spikes and the certified reference material SRM 1515, for Apple Leaves and Domestic Sludge SRM 2781.

OTHER POLLUTION VARIABLES

Other pollution variables were available for some of the monitored sites, such as particulate matter (PM10), sulphur dioxide (SO2), ozone (O3), carbon monoxide (CO) and nitrogen dioxide (NO2). These data were obtained through automatic samplers in Charqueadas and Esteio and semi-automatic air samplers in Santo Antônio da Patrulha. Samplers belonged to the Environmental Department of the State of Rio Grande do Sul (FEPAM - Fundação Estadual de Proteção Ambiental Henrique Luiz Roessler), but in some cases data were not available for all the monitoring period and did not include all parameters (PM10, SO2, O3, CO and NO2). Sample data were compared with the standards of air quality in Brazil (^{CONAMA 1990}) and with the guidelines of World Health Organization (^{WHO 1999}, 2006).

We also assessed modeled atmospheric concentrations of NOX and PM2.5, from the Chemistry Coupled Aerosol-Tracer Transport Modeling System (CCATT-BRAMS, ^{Freitas et al. 2009}, ^{Longo et al. 2013}), for each study site. Data were provided by the Group of Modeling of the Atmosphere and its Interfaces from Universidade Federal de Pelotas, state of Rio Grande do Sul (http://ccatt.ufpel.edu.br). Because the model had been recently implemented on the state, data were only available for 2014 on.

STATISTICAL ANALYSES

Aiming to test whether monitored sites (i.e. seven cities) differed according to air quality, we performed a PCA (Principal Component Analysis) using a matrix of the sites described by metals, sulphur and modeled pollutants. The result scores were plotted in a biplot and variables scores were scaled by the square root of eigenvalues. This analysis was made with the “rda” function of CRAN software R using package Vegan (^{R Core Team 2015}).

LICHEN PHYSIOLOGICAL RESPONSES

Overall, monitored sites had only slight differences regarding physiological damages on the thalli of Parmotrema tinctorum. The organic carbon contents (OC) after exposure did not show any clear differences among cities, and neither did the chlorophyll contents (Table I). Regarding the percentage of algal live cells, Triunfo (TR) was the site with the lowest values at the end of the exposure period in 2013 and 2014 (29.8% and 10.4%, respectively), while Montenegro (MO) showed the highest values after seven months of exposure in 2013 (46.1%). The highest values for Maquiné (MA) and Charqueadas (CH) were in 2014 (both around 46%). This parameter also showed a decrease in almost all sites after exposure, with the exception of MA and MO in 2014 (Table I). The percentage of dead algal cells in 2013 had a similar, but opposite, pattern to the live cells. However, the same did not happen in 2014, when the highest values in seven months were found in Esteio (ES) and the lowest in TR, the latter showing also an expressive higher value of plasmolysed algal cells in 2014 (59.8%) (Table I).

AIR POLLUTANTS MONITORING

Most air pollutants analyzed with the lichen biomonitor, the air samplers and the modeled data had different concentrations in industrial and rural cities (Table II, Figure 3 and 4).

Figure 3 Contents of sulphur and metals absorbed by the lichen Parmotrema tinctorum during seven months of exposure in 2013 and 2014. The red column represents the basal values, prior to exposure in the cities. T0 = contents prior to exposure; T1 = after two months; T2 = after five months; T3 = after seven months. 

Figure 3(continuation) Contents of sulphur and metals absorbed by the lichen Parmotrema tinctorum during seven months of exposure in 2013 and 2014. The red column represents the basal values, prior to exposure in the cities. T0 = contents prior to exposure; T1 = after two months; T2 = after five months; T3 = after seven months.. 

Figure 4 Diagram of ordination of a PCA (Principal Component Analysis) based on a matrix of sites (cities) described by the pollution variables. Sites are the black letters and the variables, the grey ones. Abbreviations: CA = Caraá, CH = Charqueadas, TR = Triunfo, MO = Montenegro, MA = Maquiné, SA = Santo Antônio da Patrulha, ES = Esteio, Cr = Cromium, Al = Aluminum, Pb = Lead, V = Vanadium, Fe = Iron, Ni = Nickel, S = Sulphur, Mn = Manganese, Cu = Copper, PM2.5 = PM_2.5, NO_x = NO_x. 

Regarding the metals absorbed by the lichen P. tinctorum, differences were found especially in the concentration of copper, zinc, iron, chromium and nickel, with higher values in CH and MO in general. There was no clear difference among sites for the other heavy metals and sulphur contents. However, when comparing these values with the ones before exposure, in the control sample, we observed that almost all pollutants were more concentrated after the exposure period (Figure 3).

High contents of pollutants were even found in the rural areas (CA, MA and SA): suplhur, zinc, chromium, nickel, lead, vanadium and aluminum had similar or higher concentrations than those found in some industrial sites (Figure 3). It is important to note that chromium had a considerable peak in CA and SA in 2013 (more than 160 mg/kg), which did not repeat in 2014. Moreover, sulphur peaked at the beginning of 2014, reaching all sampled cities. At the industrial sites (MO, TR, CH and ES) all pollutants increased after exposure when compared to the control (Figure 3).

Data from the air samplers were not available for all monitored cities during the seven months when the lichens were exposed (2013 and 2014). However, based on the available data we found differences among some cities (Table II). The modeled contents of NOx and PM2.5 (Figure 4) also varied among cities, being higher in ES and CH, both industrial cities.

CH had higher values of PM10 than SA, similar to the modeled PM2.5, where even higher values were found for ES (Figure 4). Particulate matter in CH also exceeded the recommended values for health security in both periods evaluated (^{WHO 2006}) (Table II). Measured contents of SO2, showed differences among CH, ES and SA, with Charqueadas (CH) showing higher concentrations than Esteio (ES) and Santo Antônio da Patrulha (SA). NOx also appeared in higher values in ES and CH (Figure 4). We observed greater values of ozone in CH in both years when compared to Esteio (ES), while carbon monoxide did not show any difference.

MAIN POLLUTION COMPOSITION IN EACH CITY

Cities were grouped based on pollution characteristics through a principal component analysis (Figure 4). Axis 1 explained 43.4% of the observed variance and axis 2, 28.8%. Caraá (CA) was more related to higher contents of chromium, while Charqueadas (CH), is closely related to many different metals. A greater relation of Mn, Cu and modeled NOX and PM2.5 was found in Esteio (ES). Zn was more important in Triunfo (TR) and Montenegro (MO).

PHYSIOLOGICAL RESPONSES

The effect of exposure to pollutants on Parmotrema tinctorum was visible in most of the thalli. The overall pattern was an increase in the physiology parameters, with the exception of the percentages of live algal cells. Distinctively, the sites of Montenegro and Maquiné had the highest percentages of live algal cells after the exposure period, while Triunfo had the lowest values. It is also noteworthy that Triunfo had higher values of dead and plasmolysed algal cells. While some works have noted that the percentage of dead algal cells is related to air pollution (^{Calvelo and Liberatore 2004}, ^{Käffer et al. 2012}), here we show that live algal cells can also be a useful parameter.

Many authors have demonstrated that lichens are highly sensitive to contents of particulate matter, PAHs (Polycyclic Aromatic Carbons), NO2 and SO2, showing physiological damages when exposed to these pollutants (^{Van Herk et al. 2003}, ^{Riddell et al. 2012}, ^{Käffer et al. 2012}). We did not evaluate organic compounds, however their presence is likely reflecting in loss of lichen vitality in the industrial cities, namely PAHs, dioxins, furans and others, which tend to bioaccumulate as a result of atmospheric deposition (^{Augusto et al. 2013}). Although we did not measure organic pollution, we hypothesize that important damage on lichen thalli in Triunfo could be due to the effect of these pollutants from the petrochemical complex nearby the study area, since most metal concentrations were not higher than in other industrial areas.

AIR POLLUTION MONITORING

Almost all the pollutants absorbed by P. tinctorum increased after exposure, a pattern also found in other studies (^{Käffer et al. 2012}). Lichens are known to accumulate and retain many heavy metals in quantities that exceed their physiological requirements because they are able to keep metals extracellularly as oxalate crystals or turn them into lichen acids (^{Backor and Loppi 2009}).

Zinc contents were higher in sites with industrial activities such as Esteio and Montenegro. Esteio has a high demographic density (2918 inhabitants/km², one of the highest in the state of Rio Grande do Sul), which results in an intense daily traffic. According to ^{Minganti et al. (2003)} the main potential sources of zinc are indeed industrial emissions and traffic flow. Sulphur peaks were registered at the beginning of 2014 in all sampled sites. These peaks are unlikely to be related to local pollution events, but instead to some global emissions as maybe a volcanic eruption (^{Thomas and Prata 2011}) or another important source.

The observed peak of chromium in the monitored sites of Caraá and Santo Antônio da Patrulha recorded in 2013 could be either related to leather waste burning in any of the small shoe factories in these cities or could have been released (accidentally or not) by some metalworking industry, since it only occurred in one month during the monitored period. In the urban environment of Porto Alegre, the capital and the largest and most populated city of Rio Grande do Sul, ^{Käffer et al. (2012)} found 19.0 mg/kg as the mean highest concentration of chromium in P. tinctorum, while in our study that amount was more than three times higher (Caraá = 73.0 mg/kg). Charqueadas also had higher contents of chromium (averaging 17.4 mg/kg), but constant throughout the monitoring period. In Europe, a review of chromium averages in mosses between 1990 and 2000 showed that no country had higher concentrations than 8.5 mg/kg during this period and the trend expected for the next years was a reduction in chromium emissions (^{Harmens et al. 2007}).

Despite only having data for PM10, SO2, NO2, O3 and CO at some sites, this information was valuable to understand the differences among them. It is important to note that all these pollutants were below the levels required by the National Patterns of Air Quality (^{CONAMA 1990}). However, if we consider the guidelines of the World Health Organization for particulate matter (PM10), Charqueadas exceeded the suggested limit for health security, which is an annual mean of 20 μg/m³ (^{WHO 2006}) and Santo Antônio da Patrulha is also very close to reaching the limit. Charqueadas has a greater demographic density and more intense industrial activity, while Santo Antônio da Patrulha has not many industries, but lately has had a large increment in motor vehicles (82% increase in seven years) (^{Silva et al. 2015}).

Particulate matter may include a wide range of chemical substances, such as elemental and organic carbon compounds; oxides of silicon, aluminum and iron; trace-metals; sulfates; nitrates and ammonia (^{Bergamaschi et al. 2007}). The great danger of high values of PM10 is due to its association to increased rates of bronchitis, reduced lung functions (^{WHO 2000}), and also an increase of cancer risk, since many compounds of the particulate matter are also mutagenic (^{Coronas et al. 2008}, ²⁰⁰⁹).

In a recent study, ^{Coronas et al. (2008)} analyzed the mutagenic activity of airborne particulate material (PM10) in Triunfo through the Salmonella/microsome assay, and found positive genotoxic activity, which they related to the presence of PAHs (Polycyclic Aromatic Hydrocarbons). The concentration of PM10 found by this study in 2008 was even higher than was registered here (up to 100 μg/m³). In Esteio, ^{Coronas et al. (2009)} found mutagenic activity in all samples of particulate matter (PM10), using the same method. Similar results were found by ^{Pereira et al. (2013)} for Montenegro and Santo Antônio da Patrulha. In the latter city, ^{Silva et al. (2015)} recorded peaks of 64 μg/m³ of PM2.5, much higher than the limit defined by WHO (25 μg/m³ for this parameter), as well as high values of PAHs.

Almost all other greenhouse gases differed among sites, with the exception of Carbon Monoxide. This gas is highly related to vehicular emissions (^{Mayer 1999}) and has been controlled by the Brazilian government directly at car industries, through several actions. Since 2003, when this program started, two thirds of emissions have been reduced in the country (^{IBAMA 2014}). As we have observed in this study, it is indeed true that emissions of CO are controlled, since their concentrations did not change between sites and are much lower than the WHO recommendations (^{WHO 1999}).

As for the other gases, high contents were found in Charqueadas (SO2 and O3) and Esteio (NO2), which is comparable to the results found through modeled data of NOx. The higher concentration of SO2 in Charqueadas might be due to the coal-fueled power plant, a well-known source of SO2, heavy metals and other elements (^{Garty 2000}). The higher concentration of ozone could be a result of pollutants coming from elsewhere, since this greenhouse gas arises from the oxidation of nitrogen oxides, carbon monoxides and/or volatile organic compounds (^{Jenkin and Clemitshaw 2000}). Nitrogen dioxide (NO2), which had the highest values in Esteio, derives mainly from the burning of fossil fuels, and so it is closely related to motor vehicle traffic (^{Mayer 1999}). Esteio has indeed one of the highest demographic densities in our study (2918 inhabitants/km²), almost eighteen times the density of Charqueadas (163 inhabitants/km²) (^{IBGE 2014}).

MAIN POLLUTION COMPOSITION IN EACH CITY

We created an air pollution profile for the seven monitored cities using data for metals, sulphur, modeled PM2.5 and NOx concentrations. It showed the most polluted sites to be Charqueadas and Esteio, with high contents of many pollutants as already discussed. The former has a thermoelectric power plant; the latter, important petrochemical industries; in both cities there are large steel manufacturing industries. These, together with other kinds of industries and also traffic emissions, are probably the main air pollution sources in these cities, which had also high concentrations of gases (SO2, NO2 and O3) and PM10.

The cities of Triunfo and Montenegro, mainly industrial, also accounted for the observed air pollution. Triunfo holds a large Petrochemical Complex, while Montenegro is situated at the main direction of winds coming from this complex. However, organic compounds that were not evaluated would probably better separate these sites from the rural ones (SA, MA and CA). Caraá (CA) seemed to be highly influenced by the great peak of chromium in 2013, while Santo Antônio da Patrulha (SA) and Maquiné (MA) were characterized by lower concentrations of the analyzed pollutants.