Assessment of metals in PM10 filters and Araucaria heterophylla needles in two areas of Quito, Ecuador.

The reliability of Araucaria heterophylla needles as a biomonitor was evaluated by analyzing the concentration of metals in PM10 filters and in Araucaria heterophylla needles. The sampling campaign was carried out at two sites in the city of Quito, Ecuador, in 2017-2019. Concentrations of Cr, Cu, K, Mn, Pb, Zn, Ca, Fe, Al and Mg were determined in PM10 filters and in Araucaria heterophylla needles using an Inductively Coupled Plasma - Optical Emission Spectroscopy (ICP-OES). The annual mean concentrations of PM10 ranged between 24.9 and 26.3 μg m-3, exceeding the limit established by the World Health Organization (20 μg m-3). Statistical analyses, performed for the PM10 filters, showed that dust resuspension and anthropogenic activities were important sources for PM10 emissions in the city. Metals related to natural emissions (Ca, Mg, K, Al and Fe) dominated in both types of samples, while the minor metals were those related to anthropogenic emissions (Zn, Cu, Cr and Pb). The former were positively associated with the needle samples, while the latter were associated with PM10 filters. This work not only improved scientific knowledge on the concentrations of PM10 and metals in the Andean city of Quito, but also greatly contributed to the progress of research on the use of Araucaria heterophylla needles as a biomonitor.

Introduction

Air pollution is a global problem that has adverse effects on ecosystems and living organisms. One of the main pollutants of concern is particulate matter (PM), due to its adverse effects on human health. The PM10 fraction (particles with aerodynamic diameter less or equal to 10 μm), and especially the PM2.5 fraction (particles with aerodynamic diameter less or equal to 2.5 μm), are the main cause of such effects. PM can penetrate deeply into human lungs and cause cardiovascular, respiratory and neurodegenerative disorders (Boda et al., 2020; Burnett et al., 2014; Costa et al., 2020; Middleton et al., 2008). Metals are considered one of the most challenging and toxic elements present in PM because they cannot degrade biologically or chemically and they tend to bioaccumulate (Chen et al., 2013; Li et al., 2015).

The investigation of chemical PM composition is essential to identify possible sources of pollutant emission and understand the toxicological effects. High-volume air samplers are a conventional monitoring technique, which allow to collect PM on a filter and perform the chemical analysis later (González et al., 2018; Morera-Gómez et al., 2018; Ramírez et al., 2018). On the other hand, the use of plants to identify changes in environmental quality is an ideal low-cost alternative (Ávila et al., 2019; Illi et al., 2017; Solgi et al., 2020; Turkyilmaz et al., 2018a). The ability of plants to assimilate trace metals from the surrounding atmosphere and bioaccumulate them in their tissues has made plants the most suitable biological air monitors (Sevik et al., 2019; Turkyilmaz et al., 2018b). Specifically, trees have received increasing attention, as they are long-lived species widely distributed in urban areas. Thus, they can provide comprehensive information on temporal and spatial scales (Darley, 2012; Sawidis et al., 2011). Furthermore, due to their size and persistence, trees have an outstanding adaptability and a greater accumulation of xenobiotics (Alatou and Sahli, 2019).

Monitoring and biomonitoring techniques can be used individually or they can be combined to analyze the concentration of metals in an area (Monaci et al., 2000; De Paula et al., 2015). Monaci et al. (2000) quantified various metals present in PM10 filters collected from 1995 to 1997 and in Quercus ilex leaves collected in 1997. The study (Monaci et al., 2000) was carried out along a busy road and in a park in Florence (Italy). Both materials (filters and plants) indicated that Ba and Zn were reliable tracers for vehicle traffic emissions, instead of Pb. On the other hand, De Paula et al. (2015) quantified various metals present in the hemiepiphyte herb Struthanthus flexicaulis and compared the concentrations with those found in PM2.5 filters. The study was carried out in urban, industrial and rural areas of Rio de Janeiro, Brazil. In general, the concentration of metals measured in the biomonitor was higher than those in the PM2.5 filters. Anthropogenic sources were identified using both types of samples, supporting the use of S. flexicaulis as a biomonitor.

Nevertheless, there are very few studies that address the relationship between metals accumulated in filters and plants, especially in conifer needles. Previous studies indicate that the use of conifers can be more effective in monitoring the concentration of metals than broad-leaved trees, mainly because the needles have additional resin channels that facilitate the capture of metals (Brown et al., 2017; Chen et al., 2017). Araucaria heterophylla is a conifer species of interest. However, the studies on its use as a biomonitor have been limited to the analysis of the air pollution tolerance index (APTI) (Anake et al., 2018, 2019). Whereas, its ability to accumulate metals is scarcely studied. In a previous work of the group (Alexandrino et al., 2020a), the Araucaria heterophylla was used for the first time to assess the presence of metals in areas with different vehicular traffic intensity. The samples were collected from a number of sites with high, moderate and low vehicular traffic intensity, showing that Araucaria heterophylla needles are suitable for monitoring metals associated with road traffic emissions. However, the comparison of the concentration of metals found in the needles of this conifer species with those found in PM filters is not yet available to date. Therefore, the aim of this work was to assess the reliability of Araucaria heterophylla needles as a biomonitor by analyzing the concentration of several metals in PM10 filters and in Araucaria heterophylla needles in two areas in the Andean city Quito, Ecuador.

Materials and methods

Study area and sampling site

Quito, the capital of Ecuador, covers a surface of about 372.4 km2 and is at an altitude of approximately 2850 m above sea level (m.a.s.l.) (Figure 1).The city has a population density of 7347.1 inhabitants km−2, and is the hub of the Metropolitan District of Quito (DMQ) that extends over an area of 4235 km2 with approximately 2.2 million inhabitants. Quito has a mean precipitation value of 1204 mm, concentrated mainly in the rainy period (September–May). The accelerated development and population growth of the city have brought deterioration of the air quality. Hence, an air quality monitoring network was installed in 2003 in the city. The main source of air pollution in Quito is road-traffic (Herrera et al., 2016). Moreover, the city also has industries located mainly in surrounding areas that also contribute to air pollution (Zalakeviciute et al., 2020a).

Figure 1

Map showing the Metropolitan District of Quito (blue area), city of Quito (red area) and the location of the sampling points (black symbols).

The investigation was carried out using high-volume samplers installed in two sites of the atmospheric monitoring network of Quito: Belisario (S1, 78o29′45.1″W, 0o11′04.7″S) and Los Chillos (S2, 78o27′19.5″W, 0o17′48.8″S) (Figure 1). Belisario is an urban area with a high vehicular traffic according to the Google Maps Traffic application and a real traffic sampling using the QGis software (Alexandrino et al., 2020b; Zalakeviciute et al., 2020b). On the other hand, Los Chillos is considered a suburban area influenced by industry activities (Cevallos et al., 2017; Zalakeviciute et al., 2020a).

Sampling and chemical analysis

The procedure followed to collect and determine the concentration of metals in PM10 filters and needles was explained in detail previously (Alexandrino et al., 2020a; Zalakeviciute et al., 2020a). Briefly, the PM10 samples were collected twice a month between January 2017 and December 2018 over a 24-h period. High-volume samplers (Tish Environmental, INC.; EPA reference method), with quartz fiber filters (Whatman, 20.3 cm × 25.4 cm), were used to collect the PM10. The equipment were installed approximately 10 m above ground level, operating with a flow rate of approximately 1.1 m3 min−1 under actual conditions of temperature and pressure (U.S. EPA, 1999). After the sampling period, the filters were stored in zip-lock bags and kept at 4 °C until analysis. The filters were conditioned (constant temperature of 20 ± 3 °C and relative humidity of 50 ± 5%) and weighted before and after the collection of PM10. The concentration of PM10 was determined using the mass of PM10 (difference in weight of the conditioned filter before and after the collection of particles) and the volume of air passing through the filter in standard conditions of temperature and pressure (U.S. EPA, 1999). A total of 88 PM10 filters were collected for the analysis of the concentration of PM10 and metals. Additionally, meteorological parameters (precipitation, relative humidity, pressure, temperature and wind speed) were recorded at 1-h intervals, from which a mean daily reading was calculated. The instruments used are reported in Table 1.

Table 1

Equipment for the measurement of meteorological parameters.

Meteorological parameter	Equipment
Precipitation	Rain Fall sensor model 382 (Met One Instruments)
Relative humidity	Hygro-ThermoTransmitter-compact model 1.1005.54.161 (Thies Clima)
Pressure	PTB101B Barometric Pressure Sensor (Vaisala)
Temperature	Hygro-ThermoTransmitter-compact model 1.1005.54.161 (Thies Clima)
Wind speed	010C Wind Speed Sensor (Met One Instruments)

On the other hand, a single Araucaria heterophylla tree was identified close to each monitoring station (see Figure 1), at a maximum distance of 260 m. Approximately 8 g of Araucaria heterophylla needles were collected and chemically analyzed once a month during August, September and November 2019 at both sites. The needles were sampled in the directions of the four cardinal points, at a height of about 2.5 m from the ground using plastic gloves to avoid any contamination. The collected samples were stored in clean zip-lock bags, wrapped in aluminum foil, transported to the laboratory and stored at 4 °C until analysis. Part of the samples was used to determine the water content and another faction was used for chemical analysis. Samples were only collected when no precipitation was recorded at least 5 days before sampling, in order to avoid the loss of information caused by the washing effect.

The methods by Zalakeviciute et al. (2020a) and Alexandrino et al. (2020a) were used to obtain the chemical composition of PM10 filters and needles, respectively. Briefly, 16 cm3 of loaded filter or 0.5 g of fresh needle sample was digested (microwave digestion system MARS 6 - CEM Corporation) with 10 ml of concentrated HNO3 or with 7 ml of concentrated HNO3, 2 ml of H2O2 and 1 ml of ultrapure water, for the filters and needles, respectively. The digestion was carried out at 200 °C for 45 min. The extracts were filtered and diluted with ultrapure water to a final volume of 50 ml for the filters, and 25 ml for the needles. The diluted solutions were analyzed for concentrations of Cr, Cu, K, Mn, Pb, Zn, Ca, Fe, Al and Mg using Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP-OES, Thermo Scientific iCAP 7000 Series). The extraction efficiencies were determined using the certified reference materials NIST1648a - Urban particulate matter and NIST SRM 1575a – Trace elements in Pine Needles, as explained previously (Alexandrino et al., 2020a; Zalakeviciute et al., 2020a). Moreover, blank filters were subjected to the same chemical procedure as the sampled filters in order to determine the limit of detection (LOD) and quantification (LOQ) for each element (see Zalakeviciute et al. (2020a) for details).

Additionally, the water content in the needles was determined to express the results in a dry weight basis (μg g−1 DW). This allows comparing the concentration of metals found in this work with those reported in the literature. The process consisted of drying, in triplicate, 2.5 g of fresh material at a constant temperature of 75 °C until verifying that a constant weight was reached. The mean water content in the samples was 57.50 ± 3.37% (mean value ± standard deviation).

Statistical analyses

Univariate, bivariate and multivariate statistical analyses were performed on the datasets (Taheri Shahraiyni and Sodoudi, 2016) in order to identify the emission sources and to assess the relationship between the metals found in the PM10 filters and needles. An Exploratory Data Analysis (EDA) was performed on the datasets for the PM10 filters and the Araucaria heterophylla needles. The EDA was performed for univariate and bivariate cases. Univariate analysis allowed expressing the results as the mean value ± standard deviation for each concentration of metal. The bivariate analysis performed was namely a Pearson's correlation for every pair of variables with their respective significance test. A one-way analysis of variance (ANOVA) was performed together with a pairwise comparison of means by a post-hoc Tukey's test whenever significant differences were indicated by the ANOVA between sampling years (PM10 filters) or monitoring stations (PM10 filters and needles). The ANOVA assumptions were verified by a Shapiro-Wilk normality test, as well as by inspection of residual vs. fitted values plot. Moreover, a Principal Component Analysis (PCA) was performed on the metal data of the PM10 filters to identify the potential sources of pollution that affect the study areas and evaluate their contribution (Braga et al., 2004). PCA is the most widely used multivariate statistical analysis technique in atmospheric sciences (Calesso et al., 2012). This technique seeks to reduce the size of the data set but preserves the initial information by grouping the variables that have similar characteristics into the same component (Wannaz et al., 2013), which may indicate a common source (Abril et al., 2014a; Wannaz et al., 2012).

Additionally, a PCA was performed using the metal data of the PM10 filters and needles as classification criteria in order to assess the relationship between both samples. PCA was used to highlight how strongly each characteristic (vector) influence the principal components, in order to identify the importance and relation of the measured variables (concentration of metals) according to the data of the materials (PM10 filters and needles). All statistics were performed using R language.

Results and discussion

PM10 filters

Figure 2 shows the annual variation of the concentration of PM10 for both study sites. It is observed that the limit value for the daily concentration of PM10 established by The World Health Organization (50 μg m−3, WHO, 2005), the European Union (50 μg m−3, European Commission, 2019) and the Ecuador Standards Air Quality (100 μg m−3, TULS, 2015) is not exceeded at the urban site S1. However, the WHO and the European Union standards were exceeded on two days (01/01/2017 and 22/09/2017) at the suburban site S2.

Figure 2

Annual variation in concentration of PM10 found in the filters at the two study sites: S1 - Belisario and S2 - Los Chillos.

A separate statistical analysis was performed for each year in order to provide more complete information on compliance with the guidelines-based annual concentration limit values, as well as to provide information on the variation in the concentration of PM10 between years and monitoring stations. Table 2 shows the mean values and standard deviation (SD) of the concentrations of PM10 (μg m−3) found in the filters for each sampling year in both sites. The p-values for the ANOVA test between years and monitoring stations are also shown. It is observed that the mean annual concentration of PM10 is in the range of 24.9–26.3 μg m−3. Thus, the WHO mean annual limit value (20 μg m−3) is exceeded in both years (WHO, 2005). However, these concentrations are below the mean annual limit established by the Ecuadorian Standards for Air Quality (50 μg m−3) (TULS, 2015) and the European Union (40 μg m−3) (European Commission, 2019). The annual mean concentration was higher in 2018 than in 2017 at both sites, with a higher concentration at site S2 than at site S1. However, no significant difference is observed between years or monitoring stations.

Table 2

Concentration of PM10 (mean ± SD) found in the filters at sites S1 and S2 in 2017 and 2018. Unit: (μg m−3). p-values for the ANOVA test between years and monitoring stations.

Site	2017	2018	ANOVA between years
S1 (n = 44)	24.9 ± 8.7	25.1 ± 8.9	0.9
S2 (n = 44)	26.1 ± 11.5	26.3 ± 9.5	0.8
ANOVA between stations	0.7	1.0

Table 3 shows the mean value and standard deviation of the concentration of metals (μg m−3) found in the filters analyzed in 2017 and 2018, in the two study sites. The p-values for the ANOVA test between years and monitoring stations are also shown. It is observed that the major metals found in both stations are Ca (6.6–8.7 μg m−3), Al (0.4–2.6 μg m−3), Mg (1.6–2.0 μg m−3), K (0.4–0.8 μg m−3) and Fe (0.4–0.6 μg m−3), which is in line with previous works (Celis et al., 2004; Loyola et al., 2012; Monaci et al., 2000; Vargas et al., 2012). These metals are major components of crustal materials and soil, thus they are typically related to natural origin, such as dust resuspension (Arditsoglou and Samara, 2005; Loyola et al., 2009, 2012). On the other hand, mean concentration of the minor metals Cr, Cu, Pb, Zn and Mn were found in the range of 0.009–0.2 μg m−3. These metals are typically related to anthropogenic emissions (Loyola et al., 2012; Pacyna and Pacyna, 2001), although Cu and Zn are also micronutrients which are essential for life but are toxics at high concentrations (Gandois and Probst, 2012).

Table 3

Concentration of metals (mean ± SD) found in the extract of the PM10 filters at sites S1 and S2 in 2017 and 2018. Unit: (μg m−3). p-values for the ANOVA test between years and monitoring stations.

	Site	Cr		Cu		K		Mn		Pb		Zn		Ca		Fe		Al		Mg
		2017	2018	2017	2018	2017	2018	2017	2018	2017	2018	2017	2018	2017	2018	2017	2018	2017	2018	2017	2018
	S1	0.10 ± 0.3	0.009 ± 0.003	0.1 ± 0.09	0.1 ± 0.07	0.7 ± 0.8	0.8 ± 0.3	0.04 ± 0.02	0.06 ± 0.02	0.02 ± 0.02	0.04 ± 0.04	0.1 ± 0.1	0.07 ± 0.04	7.6 ± 1.9	8.7 ± 2.8	0.4 ± 0.2	0.5 ± 0.2	0.4 ± 0.2	2.6 ± 2.0	1.8 ± 0.5	2.1 ± 1.0
	S2	0.009 ± 0.009	0.009 ± 0.006	0.2 ± 0.2	0.09 ± 0.05	0.4 ± 0.2	0.8 ± 0.3	0.03 ± 0.01	0.06 ± 0.02	0.02 ± 0.02	0.03 ± 0.03	0.1 ± 0.08	0.1 ± 0.1	6.6 ± 2.0	8.0 ± 2.1	0.4 ± 0.2	0.6 ± 0.2	0.5 ± 0.5	2.6 ± 2.0	1.6 ± 0.5	2.0 ± 0.8
ANOVA between years	S1	0.1		1.0		0.9		0.008 ∗∗		0.07		0.04 ∗		0.2		0.08		1.6e-05 ∗∗∗		0.2
	S2	0.8		0.02 ∗		8.7e-06 ∗∗∗		9.3e-06 ∗∗∗		0.6		1.0		0.02 ∗		0.008 ∗∗		4.7e-05 ∗∗∗		0.04 ∗
ANOVA between stations		0.2	0.9	0.06 ∗∗	0.6	0.09	0.5	0.04 ∗	0.9	0.6	0.5	0.5	0.2	0.1	0.4	0.2	0.9	0.4	1.0	0.08	0.7

∗∗∗ 0 < p < 0.001; ∗∗ 0.001 < p < 0.01; ∗ 0.01 < p < 0.05.

Statistically significant difference between years, in both sites, was only observed for the concentrations of Mn and Al, for which the mean concentrations were higher in 2018 than in 2017. Moreover, there is also a significant difference in the concentration of Zn at site S1 between years, with 2017 being the year with the highest concentration. A significant difference is also observed in the concentration of Cu, K, Ca, Fe and Mg at site S2 between years, with the highest concentration of Cu in 2017, while the mean concentrations of K, Ca, Fe and Mg were higher in 2018 than in 2017. In summary, it is noticed that the concentrations of the metals mainly associated with soil dust (K, Ca, Mg, Fe and Al) were higher in 2018, while the concentrations of the metals mainly associated with anthropogenic emissions (Zn and Cu) were higher in 2017. These variations may have been due to different factors, such as irregularity of the behavior of emission sources, like vehicular sources and power plants, and to meteorological parameters, as mentioned in previous works (Braga et al., 2005; Manalis et al., 2005; Ramírez et al., 2018; Sanchez-Ccoyllo and Andrade, 2002).

In order to evaluate the effect of meteorological parameters on the metal composition of PM, Table 4 shows the mean and standard deviation of precipitation, relative humidity, pressure, temperature and wind speed during the samplings in 2017 and 2018. It is observed that wind speed was lower in 2018 than in 2017, which indicates that the increase in the concentration of metals associated with soil dust in 2018 may be due to the absence of wind that causes these metals to accumulate in the ambient air. On the other hand, the higher relative humidity in 2017 compared to 2018, can causes worsened combustion efficiency (Zalakeviciute et al., 2018), and thus a greater emission of metals associated with anthropogenic activities occurs in 2017.

Table 4

Meteorological parameters values (mean ± SD) during the samplings in 2017 and 2018 at sites S1 and S2.

Site	Precipitation (mm)		Relative humidity (%)		Pressure (mmHg)		Temperature (oC)		Wind speed (m s−1)
	2017	2018	2017	2018	2017	2018	2017	2018	2017	2018
S1	3.4 ± 7.3	2.6 ± 5.7	67.6 ± 13.3	65 ± 14.5	726 ± 0.7	726 ± 1.0	14.4 ± 1.1	14.1 ± 1.1	169 ± 38.9	165 ± 45.6
S2	1 ± 3	2.7 ± 5.8	69.2 ± 11.4	67.3 ± 11	759 ± 0.7	760 ± 0.9	16.2 ± 1	16 ± 0.9	190 ± 30.8	180 ± 59.3

Regarding the differences between sites, statistically significant differences are found in 2017 for Cu (p = 0.06) and Mn (p = 0.04), with the mean concentrations of these metals being higher at site S2 and at site S1, respectively. Cu is associated with sources like brake linings (Hjortenkrans et al., 2006), and metallurgical industry (López et al., 2011). Mn, in addition to being associated with natural soil source (Bilos et al., 2001; Thorpe and Harrison, 2008), is also associated with components of steel and alloys used in the automotive industry (Fujiwara et al., 2011). Table 5 compares the concentration of the metals found in Quito with those found in other Latin American cities (Buenos Aires – Argentina, São Paulo – Brazil, Cunha – Brazil, Chillan – Chile, Santiago – Chile, Bogota – Colombia). The concentration of Cr found in Quito is 10 times higher than that reported in Bogota, but 10 times lower that reported in Santiago. Mn, Pb and Zn follow the same pattern: the concentrations of these metals are around 10 times lower in Quito than in Santiago. The concentrations of Cu and Al in Quito are similar to those found in Chillan and Santiago, and also Buenos Aires and Bogota in the case of Al, but much higher than the concentrations found in the other cities. Lower concentration of Fe is found in Quito compared to Buenos Aires, Chillan and Bogota, but higher than the ones reported in São Paulo. The concentration of Mg in Quito is similar to Santiago and Bogota, and higher than Chillan. Ca is the element which appears to be enriched in Quito compared to other world cities, and is similar to the Ca concentration in Santiago. It could be due to the composition of the carbonaceous soil-source material of this city.

Table 5

Comparison of the concentrations of metals found in PM10 filters in Quito and in other Latin American cities. PW: present work. ∗nd not detected.

City (Country)	Concentration (ng m−3)										Reference
	Cr	Cu	K	Mn	Pb	Zn	Ca	Fe	Al	Mg
Quito (Ecuador)	8.5–95	88–200	410–820	32–56	19–45	71–120	6550–8690	360–550	380–2600	1560–2120	PW
Buenos Aires (Argentina)	-	17.2–30.0	-	16.0–24.9	14.2–52.0	46.8–109	1029–1419	672–1215	763–1133	-	Reich et al. (2009)
São Paulo (Brazil)	-	29.3	-	18.90	16.6	118	-	97–0	54.3	-	Marie Bourotte et al. (2011)
Cunha (Brazil)	-	4	-	7.70	13.2	221	-	nd∗	39.6	-	Marie Bourotte et al. (2011)
Chillan (Chile)	-	180	1910	40	20	300	1610	2110	1410	290	Celis et al. (2004)
Santiago (Chile)	240–920	80–320	-	60–110	80–960	450–1080	6090–11230	-	2620–3820	1300–3940	Richter et al. (2007)
Bogota (Colombia)	7	13–28	236–262	5–20	128	37–194	1082–1534	663–1068	898–1107	1050–1380	Vargas et al. (2012)

Finally, to identify possible sources of pollution in both sites, Pearson and PCA tests were performed using the PM10 filter data. Table 6 shows the Pearson correlation coefficients for the metals found in the PM10 filters in both sites. At site S1, high correlations were found for the following metal groups: (i) Ca, Mg, Fe, Al and Mn; (ii) Cr, Cu, K and Zn; and (iii) Pb, K and Mn. The first group is clearly derived from natural sources, because these metals are major components of crustal material and soil and are indicative of the significant contribution of dust resuspension to particulate matter formation (Arditsoglou and Samara, 2005; Loyola et al., 2009, 2012). Particle resuspension can be due to the action of both wind and road-traffic activity, and can come from different sources such as wear and tear of roads and tires, construction and maintenance of buildings and roads or agricultural activities (Marcazzan et al., 2001). On the other hand, a clear anthropogenic origin is observed for the other two groups.

Table 6

Pearson analysis for metals found in PM10 filters at sites S1 (bold) and S2 (normal).

	Mg	Fe	Al	Cr	Zn	Cu	Mn	K	Pb
Ca	0.9∗∗∗0.9∗∗∗	0.5∗∗0.6∗∗	0.6∗∗0.5∗	0.060.09	-0.10.02	0.2-0.02	0.6∗∗∗0.7∗∗∗	0.20.4∗	0.10.1
Mg	1	0.5∗0.8∗∗∗	0.7∗∗∗0.8∗∗∗	0.10.02	-0.10.2	0.3-0.03	0.8∗∗∗0.9∗∗∗	0.30.8∗∗∗	0.30.01
Fe		1	0.6∗∗∗0.7∗∗∗	-0.04-0.1	-0.30.3	0.07-0.2	0.30.8∗∗∗	-0.010.8∗∗∗	0.020.1
Al			1	-0.2-0.1	-0.20.3	0.3-0.1	0.7∗∗∗0.8∗∗∗	0.10.8∗∗∗	0.4∗0.09
Cr				1	0.20.3	0.5∗0.08	0.30.02	0.6∗∗0.03	0.20.5∗
Zn					1	0.6∗∗0.2	0.30.3	0.4∗0.5∗	0.30.2
Cu						1	0.8∗∗∗-0.1	0.6∗∗∗-0.03	0.4∗0.05
Mn							1	0.4∗0.8∗∗∗	0.4∗0.3
K								1	0.6∗∗0.3

Similar results were obtained from the PCA (Table 7). Two components with eigenvalues higher than 1, and accounting for a total of 68.5% of the data variability, were identified for site S1. The first component is attributed to the soil dust (Mn, Ca and Mg). Component 2 presents high loads for Cr, Cu, K and Zn. Due to the proximity of this sampling point to areas with high vehicular flow, these metals, with an exception of K, could be associated to vehicular traffic emissions, due to their high content in the brake and tire composition (Adachi and Tainosho, 2004; Thorpe and Harrison, 2008). On the other hand, K is naturally present in the soil, but it is also found in the smoke from the combustion of vegetative biomass (Celis et al., 2004). Thus, the occurrence of K in this factor may also reveal a small contribution from biomass burning (Amato et al., 2009). Moreover, this second component is also loaded by Pb, Fe and Al. Pb may be emitted from several sources, such as motor oil combustion, resuspension of enriched road dust and brake wear (Lough et al., 2005). On the other hand, the high correlation of Fe and Al with the crustal elements Ca (0.5 for Fe and 0.6 for Al) and Mg (0.5 for Fe and 0.7 for Al) (Table 6), and their presence in the component loaded with the anthropogenic elements (Cr, Cu, Zn and Pb, Table 7), can suggest that Fe and Al may also have additional source different than the dust resuspension, specifically, vehicle traffic. In fact, these metals have been associated with components of steel and alloys used in the automotive industry (Fujiwara et al., 2011). Moreover, Fe has been found in high concentration in car brake linings, car brake dust and in the exhaust of diesel and gasoline engines, while Al has been identified in relatively high concentration in motor exhaust (Loyola et al., 2012).

Table 7

Factor loading for metals in PM10 filters at sites S1 (bold) and S2 (normal).

Metal	Dim. 1	Dim. 2
Cr	-0.20.02	0.30.6
Cu	-0.4-0.03	0.30.3
K	-0.30.4	0.40.08
Mn	-0.40.4	0.02-0.02
Pb	-0.30.1	0.20.6
Zn	-0.10.2	0.50.4
Ca	-0.40.3	-0.3-0.09
Fe	-0.20.4	-0.4-0.1
Al	-0.40.4	-0.3-0.1
Mg	-0.40.4	-0.3-0.1
Eigenvalue	4.25.0	2.61.7
Variance (%)	42.549.1	26.017.2
Cumulative variance (%)	42.549.1	68.566.4

At site S2, high correlations were found for the following metal groups: (i) Ca, Mg, Fe, Al, Mn and K; (ii) K and Zn; and (iii) Pb and Cr (Table 6). As in the case of the site S1, the first group is crustal elements, identifying a soil dust source, while the other two groups can be associated with anthropogenic emissions. Similar results were observed with the PCA (Table 7). Two dimensions accounting for a total of 66.4% of the data variability were identified for site S2. Dimension 1, which accounts for the 49.1% of the total variance, presents high loads for K, Mn, Ca, Fe, Al and Mg. Dimension 2 presents high loads for Cr, Cu, Pb and Zn, which are associated with industrial and road traffic emissions. Emissions of Cr, Cu, Pb and Zn, with respect to industrial activities, could be related to chemical industries (Wannaz et al., 2012), cement plants (Ogunkunle and Fatoba 2014; Wannaz et al., 2012), metallurgical industries (Quiterio et al., 2004) and manufacturing processes of power transformers (Bermudez et al., 2009). Unlike at the urban site S1, Fe and Al seems to come mainly from resuspension of dust, due to the tight correlation with crustal elements (Table 6) and its presence in the same dimension (Table 7).

Araucaria heterophylla needles

Table 8 shows the mean and standard deviation of the concentration of metals found in the Araucaria heterophylla needles. It is observed that, in both areas, the metals with the highest mean concentration are Ca (27627 μg g−1 DW (S1); 18236 μg g−1 DW (S2)), Mg (3999 μg g−1 DW (S1); 4093 μg g−1 DW (S2)), K (3212 μg g−1 DW (S1); 3032 μg g−1 DW (S2)), Al (996 μg g−1 DW (S1); 642 μg g-1 DW (S2)), Fe (734 μg g−1 DW (S1); 278 μg g−1 DW (S2)) and Mn (382 μg g−1 DW (S1); 136 μg g−1 DW (S2)). The highest concentrations were also observed for Ca, Mg, K, Al and Fe in the PM10 filters (see Table 3). Similar sequence for metal accumulation was also obtained in works using biomonitors such as mosses (Lazo et al., 2019; Stafilov et al., 2018; Steinnes et al., 2016), pine needles (Brown et al., 2017) and lichen (Bergamaschi et al., 2007). Ca, Mg and K are macronutrients considered essential for plant metabolism and therefore are part of the tissue composition (Fernández and Carballeira, 2000). On the other hand, Al, Fe and Mn are considered geological elements but are also related to components used in the automotive industry (Fujiwara et al., 2011). Zn, Cu, Cr and Pb are other minor metals found in the Araucaria heterophylla needles, with concentrations ranging from 0.6 to 48.3 μg g−1 DW. These results indicate that Araucaria heterophylla needles are suitable for biomonitoring because they tend to accumulate metals and the needles indicate similar differences with PM10 in quantitative relationships in the two study areas (Tables 3 and 8).

Table 8

Concentration of metals (mean ± SD) found in the Araucaria heterophylla needles collected near to the monitoring stations. Unit: (μg g−1 DW). p-values for the ANOVA test between monitoring stations.

Site	Cr	Cu	K	Mn	Pb	Zn	Ca	Fe	Al	Mg
S1 (n = 3)	3.6 ± 1	12.2 ± 3.5	3212 ± 1459	382 ± 208	3.0 ± 2.1	48.3 ± 17.1	27627 ± 2501	734 ± 211	996 ± 793	3999 ± 409
S2 (n = 3)	0.6 ± 0.3	3.3 ± 1.4	3032 ± 1312	136 ± 23.9	1.2 ± 0.2	17.5 ± 2.8	18236 ± 1825	278 ± 91.9	642 ± 505	4093 ± 214
ANOVA between stations	0.008∗∗	0.02 ∗	0.9	0.1	0.2	0.03 ∗	0.006 ∗	0.03 ∗	0.6	0.7

∗∗ 0.001 < p < 0.01; ∗ 0.01 < p < 0.05.

The mean concentrations of Cr, Cu, Zn, Fe and Ca show statistically significant differences between monitoring stations, with S1 being the site with the highest concentration of these metals (Table 8). As mentioned above, Cr, Cu and Zn are largely associated with traffic emissions (Al-Alawi and Mandiwana, 2007; Solgi et al., 2020). Thus, the higher concentration of these metals at site S1 could be due to higher traffic flux than at site S2. On the other hand, the higher concentrations of Fe and Ca at site S1, compared to site S2, could be related to the natural composition of the plant and to the soil resuspension (Abril et al., 2014b).

The concentration of Cr, Cu, Mn, Pb, Zn and Fe found in the present work was compared with data from studies carried out in other cities of the world (Table 9). Data for K, Ca, Al and Mg were not included because there were no values for these elements in the other studies. Because there is no information in literature on the detection of metals in the air using Araucaria species, the concentrations present in Table 9 are based on other plant species. Therefore, the concentrations are not directly compared because the capacity to capture metals depends on the plant species (Bergamaschi et al., 2007; Čeburnis and Steinnes, 2000; Minganti et al., 2003). However, the comparison serves to give an approximate idea of the concentrations of metals in different countries. In general, the concentrations are in the same order of magnitude, with the exception of Pb and Zn, for which the concentrations were up to 100 and 10 times higher, respectively, in the Pinus halepensis L. needles used in Jordan – Amman as a biomonitor.

Table 9

Comparison of the concentrations of Cr, Cu, Mn, Pb, Zn and Fe found in the Araucaria heterophylla needle samples and in other cities of the world. Different plant species are included. Unit: (μg g−1 DW). PW: present work.

City/Country and plant species	Cr	Cu	Mn	Pb	Zn	Fe	Reference
Quito - EcuadorAuracaria heterophylla	0.5–3.6	3.3–12.2	136–382	1.2–3.1	17.5–48.3	278–734	PW
Cordoba - ArgentinaTillandsia capillaris		10–13	294–319	0.81–1.02	18.7–25.5	442–581	Wannaz et al. (2012)
LithuaniaPicea abies needles	0.2	-	678	0.7	35	-	Čeburnis and Steinnes (2000)
LithuaniaJuniperus communis	0.1	-	725	0.6	22	-	Čeburnis and Steinnes (2000)
LituhaniaMoss	1.3	-	273	9.9	47	-	Čeburnis and Steinnes (2000)
Amman - JordanPinus halepensis L. needles	-	16.5–37	-	40.0–196	25–262	-	Al-Alawi and Mandiwana (2007)
TurkeyPinus brutia needles	0.8–12.1	3.1–10.1	-	2.8–12.1	-	93.1–596	Dogan et al. (2010)
São Paulo - BrazilTillandsia usneoides	1.0–3.7	4.3–56	135–234	-	41–189	612–1324	Figueiredo et al. (2007)

Comparison between the concentration of metals in PM10 filters and Araucaria heterophylla needles

Table 10 shows the mean concentration, in mg kg−1, of the metals found in PM10 filters and Araucaria heterophylla needles. It is observed that, in both areas, the concentrations of crustal materials and soil (K, Ca, Mg, Mn and Fe) were higher in the plant material than in the PM10 filters, while the concentrations of the anthropogenic metals Cr, Cu, Pb and Zn were higher in the PM10 filters. Additionally, a PCA analysis was carried out using the materials (PM10 filters and Araucaria heterophylla needles) as classification criteria, to have a clear view of the metals that are more associated with each material (Table 11, Figure 3). A positive association is observed between the needles and K, Ca, Mg, Mn, Fe and, to a lesser extent, Al, while Cr, Cu, Pb and Zn are associated with PM10 filters (see Figure 3). This result is opposite to that observed in the work of De Paula et al. (2015) where, in general, the concentration of the metals measured in the Struthanthus flexicaulis leaves were higher than those found in the PM filters. This was attributed to the fact that the recollection of PM filters was over a 24-h period, while leaves present a long-term effect. However, in the present work, taking into account frequent high precipitation events in Quito throughout the year, the lower concentration of anthropogenic metals in the Araucaria heterophylla needles could be mainly due to a wash effect of the rain that contribute to eliminate metals retained in the wax layer (Çelik et al., 2005; Al-Alawi and Mandiwana, 2007; Solgi et al., 2020; De Nicola et al., 2008). Roberts (1972) indicated that the retention of pollutants in plant leaves depends largely upon the physical, rather than the chemical, state of the pollutant and the physical form of the vegetation, and the wash effect is different depending of the plant species (Little, 1973; Moraghan, 1991; Rossini and Valdés, 2004).

Table 10

Concentration of metals (mean value ± SD) found in PM10 filters and in Araucaria heterophylla needles. Unit: mg kg−1.

Site and materials	Cr	Cu	K	Mn	Pb	Zn	Ca	Fe	Al	Mg
S1
PM10 filter	13.2 ± 48.6	27.6 ± 22.6	208 ± 155	13.8 ± 5	7.8 ± 8.3	26.6 ± 25.3	2301 ± 693	139 ± 50.1	440 ± 509	554 ± 218
Araucaria heterophylla	1.6 ± 0.5	5.5 ± 1.3	1434 ± 564	175 ± 102	1.4 ± 1	21.9 ± 8	12571 ± 1688	332 ± 92.1	456 ± 374	1812 ± 175
S2
PM10 filter	2.6 ± 2	40.1 ± 44.3	182 ± 98.4	13.1 ± 5.7	7.7 ± 8.6	29.2 ± 26.1	2117 ± 617	130 ± 67.5	463 ± 517	516 ± 204
Araucaria heterophylla	0.3 ± 0.2	1.6 ± 0.6	1442 ± 603	65.5 ± 13	0.6 ± 0.07	8.8 ± 0.5	8723 ± 694	132 ± 41.4	305 ± 236	1961 ± 128

Table 11

PCA for the metals found in PM10 filters and in Araucaria heterophylla needles.

Metal	Dim. 1	Dim. 2	Dim. 3
Cr	-0.02	0.3	0.7
Cu	-0.2	0.6	0.3
K	0.9	0.03	0.3
Mn	0.8	-0.1	0.2
Pb	-0.1	0.6	-0.08
Zn	-0.02	0.7	0.2
Ca	1.0	-0.1	0.1
Fe	0. 7	0.3	-0.4
Al	0.3	0.5	-0.7
Mg	0.9	0.05	0.02
Eigenvalue	3.8	1.7	1.3
Variance (%)	38.3	16.9	12.6
Cumulative variance (%)	38.3	55.2	67.8

Figure 3

PCA analysis to assess the associated of metals with each material (PM10 filter and Araucaria heterophylla needles).

We are aware of the limitation of the present study related to the analysis of the effect of rain on the removal of metals from the needles. However, we are confident that the dataset available here can give preliminary results of this effect. Therefore, Figure 4 shows, as an example, the concentration of Cr, Cu, Pb and Zn in the Araucaria heterophylla needles and the precipitation level in sampling months (August, September and November 2019) at site S1. It is observed that the concentration of the metals decrease in the months with highest precipitation levels (September and November). In this way, the Araucaria heterophylla needles could be a good biomonitor, but it must be taken into account that the samples must be collected in a period when it does not rain to obtain more reliable results of the metals accumulated in the plant.

Figure 4

Concentration of Cr, Cu, Pb and Zn found in the Araucaria heterophylla needles and precipitation in August, September and November (site S1).

Conclusions

The concentrations of different metals were investigated in PM10 filters and Araucaria heterophylla needles at two sites in the city of Quito, Ecuador. There is a clear significant contribution of dust resuspension and anthropogenic activities to particulate matter presence in both sites. Ca, Mg, K, Al and Fe were the major metals found in PM10 filters and Araucaria heterophylla needles. At the urban site S1, Fe and Al were found to have an additional source different to the resuspension of dust, specifically vehicular contributions. Minor metals found in both materials were Zn, Cu, Cr and Pb, which are associated with anthropogenic activities. Araucaria needles showed higher concentration of Cr, Cu, Zn, Fe and Ca at site S1 compared to the suburban site S2. This could be due to higher traffic flux in that area. The concentrations of crustal materials and soil (K, Ca, Mg, Mn and Fe) were higher in the Araucaria heterophylla needles than in the PM10 filters, while the concentrations of the anthropogenic metals Cr, Cu, Pb and Zn were higher in the PM10 filter. In fact, a PCA analysis showed a positive association between the needles and K, Ca, Mg, Mn, Fe and, to a lesser extent, Al, while Cr Cu, Pb and Zn were associated with PM10 filters. The lower concentration of anthropogenic metals in the Araucaria heterophylla needles than in the PM10 filters was attributed to a wash effect of the rain that contributed to eliminate metals retained in the wax layer. Based on the results obtained in this work, Araucaria heterophylla needles are suitable for biomonitoring because they tend to accumulate metals and the needles indicate similar differences with PM10 in quantitative relationships. However, the samples should be taken in a period when it does not rain, to obtain more reliable results on the ability to accumulate metals.

Declarations

Author contribution statement

K. Alexandrino: Conceived and designed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

T. Mancheno: Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

R. Zalakeviciute and M. González-Rodríguez: Contributed reagents, materials, analysis tools or data.

Funding statement

This work was supported by Universidad de Las Américas (Grant numbers: AMB.KAF.19.04 and AMB.RZ.19.01), Quito, Ecuador.

Data availability statement

Data will be made available on request.

Declaration of interests statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.