Monitoring of heavy metals in topsoils, atmospheric particles and plant leaves to identify possible contamination sources

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ “This is an Accepted Manuscript of an article published by Elsevier in: MICROCHEMICAL JOURNAL on 2007, available at: https://doi.org/10.1016/j.microc.2007.01.003” 1 Monitoring of heavy metals in topsoils, atmospheric particles and plant leaves to identify possible contamination sources 1S. Rossini Oliva and 2A.J. Fernández Espinosa* 1Department of Plant Biology and Ecology, University of Seville, Av. Reina Mercedes s/n, Apdo. 1095,E-41080 Seville, Spain, Tel.: +34954557056; Fax: +34954557059 E-mail: [email protected] 2 Department of Analytical Chemistry, University of Seville, C. Profesor García González s/n, E-41012, Seville, Spain, Tel.: +34954554368; Fax: +34954557168 E-mail: [email protected] *Corresponding author Abstract The research reports results on metal pollution of urban topsoils in relation to the metal content in leaves of two plant species and atmospheric particles. The content of pollutants (Ba, Cd, Cu, Fe, Mn, Ni, Pb, Ti and V) was determined by ICP-OES. Twenty-two samples of soil were collected during six months from two different sites of an urban area and from one rural zone. Regarding the pollution level, the studied soils were found to be low. Results for enrichment (EF) and concentration (CF) factors showed that soils were enriched in Pb, Ba, Cu and Ni. However, both species of plants showed a common behavior for all elements acting as excluders. ANOVA and different multivariate statistical analyses confirmed that the main pollution source of soil is traffic and fertilizers. Cd, Fe, Mn, Ti and V elements were attributed to natural sources. Also, It was suggested that N. oleander leaf is useful as a bio-monitors of soil pollution by Cu. Similarly, a direct relationship was found between the content of Cu in soils with the Cu level in PM10 atmospheric particles. The origin was attributed to dry and wet atmospheric deposition processes. Keywords: Soil; leaf; atmospheric pollutants; metals; PM10 2 1. Introduction Nowadays it is well known that cities suffer from considerable air pollution due to the release of different pollutants into the atmosphere whose levels can be toxic. Vehicular traffic is widely recognized to be a significant and increasing source of atmospheric and soil pollution in urban environments [1-3]. Traffic is particularly different to other sources of air pollution because vehicular emissions are readily inhaled by humans due to their close proximity before the atmospheric dilution can affect public health. Besides, important parameters such as the number and type of vehicles, slow moving traffic, etc., affects air pollution levels in urban areas. The relationship between emissions and pollution levels varies depending on cities. Infrastructure and town planning determine the emission pattern while meteorology and topography determine dispersion and transformation [4]. With regard to anthropogenic emissions in Seville, the main source is vehicular traffic, since the industrial influence is quite low [5,6]. Because of air pollution, urban soils can be contaminated by metals from different anthropogenic sources and from natural processes [7]. These pollutants can also bio-accumulate in plants. Atmospheric pollution is one of the major sources of heavy metal contamination in soils and roadside dusts in urban areas [8]. The contribution of metals from anthropogenic sources in soils is higher than the contribution from natural sources [9]. Soils in an urban environment may have a direct influence on human health by direct contact or suspended dust [10]. In addition, high levels of metals in urban soils have been recognized as important sources of metal intake in children, elevating metal levels in children’s blood. Also, soil ingestion constitutes an exposure route of contaminants to children as water or food ingestion [11]. It is important to assess the possible sources of pollutants in urban soils. The literature reported that urban soils often contain enriched levels of heavy metals relative to natural and anthropogenic background levels [10,12-17]. This suggests that the contribution of each one of them is often difficult to determine since urban soils are periodically disturbed by irrigation, construction activities and partial or total replacement [13]. The distribution of several heavy metals in some soils of Seville was 3 studied in 2002 [10], but concentrations of some toxic elements (e.g. Ba, Cd, V, etc.) remain poorly documented and need more investigation. On the other hand, it is well known that plants accumulate trace elements from the atmosphere and have been used in several bio-monitoring studies [18-22], offering low-cost information about environmental quality with the advantage of easy sampling. A close connection of the soil “O-horizon” with plant chemistry and atmospheric inputs was demonstrated in Northern Europe [23]. In our previous study [24], where more than twenty samples of plant leaves (Nerium oleander L. and Lantana camara L.) and airborne particles (PM10) were analyzed, it was reported that the only correlation found between metal concentrations in PM10 particles with either Nerium oleander or Lantana camara leaves were significant for Fe and Cu for oleander with respect to PM10. Regarding these results, it would be interesting to investigate if elements in atmospheric particles are also related with soil contents. At the same time, the assessment of metal enrichment and composition of urban soils is essential to establish if topsoils are polluted in the sampling sites. Therefore, the objectives of the present study were: firstly, the assessment of the grade of soil contamination in representative sites of the city of Seville. Secondly, the comparison study of trace element levels in PM10 particles, leaves of Nerium oleander L. and Lantana camara L. [24] in relation to those in the soil where the plants have grown. Thirdly, knowledge of the different behaviors of elements in both studied species of plants, investigating also the metals in which soils are enriched. The relationship between metal contents in plants and particles with soils might be useful to establish a pattern of behavior for bio-monitoring species, providing information of great interest for environmental authorities. 2. Experimental 2.1. Sampling sites Seville is the administrative center of Andalusia. It represents the most densely populated city of Southern Spain with around 700,000 inhabitants and a population density of 5,000 inhabitants per km2. It is situated 10 m above sea level on an extensive 4 plain crossed by the Guadalquivir River and has a Mediterranean climate with an average annual temperature of 19 °C and rainfall of 580 mm. Traffic of cars and trucks represents the most important pollution source because industries are rather scarce. Important crustal particle contributions in Seville and Southern Spain are well documented by the local scientific community and both regional and national governments. These particles come from the lands surrounding the city and from the North Africa deserts. 2.2. Sampling and analytical procedure Twenty-two samples of particles smaller than 10 µm (PM10) and samples of N. oleander and L. camara adult leaves were collected during autumn 2003 to spring 2004 from different sites [24] (Figure 1). According to local and regional authority information, one of the sampling sites, which represent the area of Seville most affected by the highest vehicular exhausts, is Torneo (TRN); another site with a low traffic density is Porvenir (POR). During the sampling period, the traffic densities in Torneo and Porvenir were 23,067 and 6,908 vehicles per day (VPD) respectively; the annual average daily traffic (AADT) in Seville was 16,838 vehicles per day (VPD). A third sampling site was chosen in Castilblanco as the control site (BLK) in a rural area at 40 km toward the Northern Seville. Twenty-two soil samples were collected from three sites (POR, TRN and BLK) for two different species of plants (N. oleander and L. camara). Soils where both plants grow were at different locations in each sampling site, but in close proximity. Soil samples were collected (5-10 cm depth) at the same time as plant and PM10 samples. All soil samples were dried at room temperature until a constant weight was achieved and sieved through a 2 mm mesh screen; one aliquot was ground in an agate ball mill (Retsch Model PM400) to obtain a more homogenous grain distribution at a size <50 µm. The soil characteristics surrounding the three sampling sites were different. Therefore, sampling sites can be classified in the following manner: BLK is characterized only by natural and unaltered soil. TRN and POR substrates had been 5 heavily disturbed and were characterized by natural soil and asphalt. Soils were differentiated according to the plant where they grew, oleander soil and lantana soil. Filter and plant samples were treated according to a previous work [24]. Soil samples were mineralized with 8 ml of a mixture of HNO3 and HCl (1:3) in a microwave digestion system. This method is widely used in environmental geochemistry studies and recommended by the National Government regulation. After digestion and cooling, sample solutions were transferred to 15 mL sterile tubes. The concentrations of Ba, Cd, Cu, Fe, Pb, Mn, Ni, Ti and V were determined simultaneously by inductively coupled plasma atomic emission spectrometry (ICPOES), using a Fisons-ARL 3410 sequential multi-element instrument. The accuracy of the determinations was previously checked by analysis of a NIST reference material (SRM 2711). The recovery range was from 71.9 to 97.6 %, which proved the validity of the methodology used. A gold solution was used as the internal standard to spike samples, blanks and reference materials in order to check the correct extraction of the elements. Previously, a study concluded that gold was not present in urban air [25] or plants [18] and its presence within the Teflon vessel had no influence on analytical determinations of the other metals. All ultra-pure reagents and standards were supplied by Merck. Distilled water was of Milli-Q grade obtained from a Waters-Millipore apparatus, model Plus. 2.3. Data Analysis Two parameters were calculated in order to characterize the origin and transfer of elements: 1. The Concentration factor (CF), expresses the ratio of metal concentration (M) between plants and soils: CF = (Mplant/Msoil). The ratio was calculated for both species of plants revealing the behavior about the pollutants studied. 2. The Enrichment factor of soil (EFsoil), is the relative abundance, with regards to Iron, of one element (M) in a soil compared to its relative abundance (M/Fe) in the local control site: EFsoil = (M/Fe)soil/ (M/Fe)control. The enrichment factor was used to establish which elements of soils had been relatively enriched, allowing the evaluation of the anthropogenic impact. 6 Statistical analyses were carried out using the CSS: STATISTICA (StatSoft) software package. Basic statistic was used to calculate the mean values and standard deviations while ANOVA was used to detect significant differences between sampling sites (p0.05). The Scheffé's multiple mean comparison test was also used in the ANOVA. The relationship between metal concentration in atmospheric particles and plants and soils was tested by the Pearson correlation coefficients (r). Cluster analysis (CA) and Principal Component Analysis (PCA) were used as classification tools. The components of the PCA were rotated using a Varimax rotation. 3. Results and discussion 3.1. Element concentrations in soil samples Descriptive statistics of metal concentrations of soils in the different sites are reported in Table 1. The dispersion of data is small for elements in oleander soils and lantana soils except for Cu, Ba, Ni and, mainly Pb in lantana soils; this indicates that soil compositions are homogeneous for the other elements, being Pb, Cu, Ba and Ni elements of variable composition possibly originated from anthropogenic sources [6]. Large element variability rather than a high concentration provides an indication of an unusual element source [6,17]. In Table 1 the average content of elements is compared to those reported in Seville in 2002 [10] and in urban soils of Madrid in 1998 [13]. There are some similarities between the present data and those of Seville in 2002 [10], whilst the comparison also suggests that the soils studied have higher contents in Ni and V than in Madrid. This can correspond to some degree of pollution by these metals in Seville probably due to frequent emissions during the past from nearby big industries actually disappearing [25]. With regards to traffic sources, it is well known that Pb and Cu come from leaded fuel and brake linings in motor vehicles, respectively [6]. According to Quebec Ministry of Environment (QME), the Pb and Cu contents in soils of both sampling sites can be considering as Level B, which is the acceptable limit for residential, recreational and institutional sites. Levels of Pb and Cu observed in the background site (BLK) are 7 lower than those usually considered as control values (Cu = 40 mg kg-1, Pb = 50 mg kg-1) and they are lower than values reported in Seville in 2002 for the Alamillo Park, a site considered as background by this study [10]. However, Ti and again V contents in BLK were higher than background values reported in Madrid in 1998 [13]. Oleander soils Results of the Scheffé Test show that element concentrations in oleander soil differ significantly (p <0.05) among the sampling sites, except for Mn. The Pb concentration was significantly low in BLK with respect to the other two sampling sites and differences in Pb concentrations between them were not found. However, both Pb mean values in POR y TRN are much higher than those reported for soils sampled around an industrial area of Huelva (Spain) [17] and for an urban soil of Siena (Italy) [19]. Lead is linked to the deposition of atmospheric particles on soils generated primarily from traffic and from other anthropogenic activities [13]. Differences between soils suggest that Pb content on the soil surface is the result of accumulation process from atmospheric sources during the past years, mainly traffic. Traffic in Seville was concentrated surrounding the historical center before the new roads and avenues made due to the Exposition Universal of 1992 (Expo92). This high traffic density in the past affecting TRN and POR is probably the reason for high Pb levels in soils. Regarding Cu and Ni concentrations, they were significantly lower in BLK than in POR and TRN, but differences between both were not observed. Both elements were considered as anthropogenic, which explains the lower values found in BLK. Apart from traffic, contamination of soils by Cu is probably derived from use of fertilizers, sprays, bactericides, fungicides and agricultural or municipal wastes as well as from industrial emissions [26,27]. Ni emissions normally come from combustion of fossil fuels. Furthermore, atmospheric pollution by Ni is high in Seville due to industrial sources [24]. Similarly, Ba and Ni concentrations were also found to be lower in BLK than in the other sites, while concentrations in TRN were significantly higher than in POR. Barium is another element associated with traffic [19,24]. However, Cd, Fe, Ti and V concentrations were significantly higher in BLK than in the urban sites, whilst no 8 differences were observed between TRN and POR. In addition, the Mn concentration in BLK was significantly higher than in TRN, but lower than in POR. This constitutes a contrary pattern than the previous elements commented on above, i.e. Cu, Ni, Ba and Pb. Lantana soils In the case of the lantana soils, results of the Scheffé Test showed that, as in oleander soils, the Cd, Fe, V and Ti concentrations were significantly higher in BLK than in the urban sites. Besides, values of Ti concentrations in BLK were higher than in the urban soils of other work [20]. Fe and V can be considered elements from soils [28] and Ti is a lithophilic element [24]. Therefore, results of the Scheffé Test on oleander soils and lantana soils confirm a contrary behavior between the two groups of elements, indicating that Fe, Ti, V and Cd were associated to crustal elements and the elements Pb, Cu, Ni and Ba were associated to vehicular traffic. Besides these two groups, conclusions on Manganese suggest a different behavior. 3.2. Comparison with plant and particle contents Inter-element relationships provide interesting information on heavy metal sources and pathways. Correlation analysis (Table 2) shows that Pb, Ba, Cu and Ni were negatively correlated with Cd, Fe, V and Ti suggesting that these elements come from different sources. These correlations corroborate the results of Section 3.1. Additionally, Ba, Cu, Ni and Pb were inter-correlated in oleander and lantana soils, suggesting that they have a common origin associated with traffic. Correlation coefficients were higher in oleander soils than in lantana soils (Ba-Cu r=0.88/0.75, Ba- Ni r=0.87/0.76, Ba-Pb r=0.87/0.81, Cu-Pb r=0.88/0.87, Cu-Ni r=0.81/0.44, Pb-Ni r=0.86/0.51). Similarly, Fe, Ti, V and Cd were also inter-correlated between them with correlation coefficients similar in both soils (Fe-V r=0.98/0.95, Fe-Ti r=0.93/0.72, Fe- Cd r=0.99/0.97, Cd-V r=0.97/0.95, Cd-Ti r=0.92/0.52, Ti-V r=0.97/0.87). Therefore, the highest coefficients were found in oleander soils for anthropogenic elements while similar coefficients were obtained in both types of soils for earth crustal elements. It 15 [9] Nriagu, J.O., Pacyna, J.M. Quantitative assessment of worldwide contamination of air, water and soils by trace metals. Nature 1988; 333: 134-139. [10] Madrid, L., Díaz-Barrientos, E., Madrid, F. Distribution of heavy metal contents of urban soils in parks of Seville. Chemosphere 2002; 49: 1301-1308. [11] Ljung, M., Selinus, O., Otabbong, E. Metals in soils of children’s urban environments in the small northern European city of Uppsala. Sci Total Environ 2005; in press. [12] Sánchez Camazano, M., Sánchez- Martín, M.J., Lorenzo, L.F. Lead and cadmium in soils and vegetables from urban gardens of Salamanca (Spain). Sci Total Environ 1994; 146-147: 163-168. [13] De Miguel, E., Jimenez de Grado, M., Llamas, J.F., Martín Dorado, A., Mazadiego, L.F.. The overlooked contribution of compost application to the trace element load in the urban soil of Madrid (Spain). Sci Tot Environ 1998; 215: 113– 122. [14] Manta, D., Angelone, M., Bellanca, A., Neri, R., Sprovieri, M. Heavy metals in urban soils: a case study from the city of Palermo (Sicily), Italy. Sci Total Environ 2002; 300: 229-243. [15] Hernandez, L., Probst, A., Probst J.L., Ulrich, E. Heavy metal distribution in some France forest soils: evidence for atmospheric pollution. Sci Total Environ 2003: 312: 195-219. 16 [16] Wen-hua, L., Jing-zhu, Z., Zhi-yun, O., Leif, S., Guo-hua, L. Impacts of sewage irrigation on heavy metal distribution and contamination in Beijing, China. Environ Int 2005; 31: 805-812. [17] Mingorance, M.D., Valdés, B., Rossini Oliva, S. Distribución de metales en suelos y plantas que crecen en un área sujeta a emisiones industriales. In: Abstracts of 6th Iberian and 3rd Iberoamerican Congress of Environmental Contamination and Toxicology Cádiz. Encuadernaciones Martínez, Puerto Real, Spain, 2005, p. 41. [18] Markert, B. Instrumental element and multi-element analysis of plant samples. Sussex, England: John Wiley & Sons, 1996, 296 pp. [19] Monaci, F., Bargagli, R. Barium and other trace metals as indicators of vehicle emissions. Water, Air Soil Pollut 1996; 100: 89-98. [20] Bargagli, R., Monaci, F., Agnorelli, C. Oak leaves as accumulators of airborne elements in an area with geochemical and geothermal anomalies. Environ Pollut 2003; 124: 321-329. [21] Rossini Oliva, S., Valdés, B. Ligustrum lucidum Ait. f. leaves as bioinicators of the air quality in a Mediterranean city. Environ Monit Assess 2004; 96: 221-232. [22] Rossini Oliva, S., Rautio, P. Could ornamental plants serve as passive biomonitors in urban areas? J Atmos Chem 2004; 49: 137-148. [23] Reinmann, C., Kashulina, G., De Caritat, P., Niskavaara, H. Multi-element, multi médium regional geochemistry in the European Artic: element concentration, variation and correlation. Appl Geochem 2001a; 16: 759-780. 17 [24] Fernández Espinosa, A. J., Rossini Oliva, S. The composition and relationships between trace element levels in inhalable atmospheric particles (PM10) and in leaves of Nerium oleander L. and Lantana camara L. Chemosphere 2006; 62: 1665-1672. [25] Fernández, A. J., Ternero, M., Barragán, F. J., Jiménez, J. C. A chemical speciation of trace metals for fine urban particles. Atmos Environ 2002; 36: 773-780. [26] Kabata-Pendias A., Pendias, H. Trace elements in soils and plants. CRC Press, An Arbor, Michigan, 1992, 365 pp. [27] Adriano, D.C. Trace elements in the terrestrial environment. Springer-Verlag, New York, Berlin, Heidelberg, Tokyo, 2001, 880 pp. [28] Rossini Oliva, S., Mingorance, M.D. Assessment of airborne heavy metal pollution by aboveground plant parts. Chemosphere 2006; in press. [29] Reinmann, C., Koller, F., Frengstad, B., Kashulina, G., Niskavaara, H., Englmaier, P. Comparison of the element composition in several plant species and their substrate from 1500000-Km2 area in Northern Europe. Sci Total Environ 2001b; 278: 87-112. [30] Rossini Oliva, S., Mingorance, M.D. Study of the impact of industrial emission on the vegetation grown around Huelva (South of Spain) city. J Atmos Chem 2004; 49: 291-302. [31] Rossini Oliva, S., Valdés, B., Mingorance, M.D. Nerium oleander L. as a means to minimise the effects of pollution. Bocconea 2006; 19, in press. 18 [32] Mertens, J., Luyssaert, S., Verheyen, K. Use and abuse of trace metal concentrations in plant tissue for biomonitoring and phytoextraction. Environ Pollut 2005; 138: 1–4. [33] Rossini Oliva, S., Raitio, P. Review of cleaning techniques and their effects on the chemical composition of foliar samples. Boreal Environ Res 2003; 8: 263-272. [34] Chamberlain AC. Fallout of lead and uptake by crops. Atmos Environ 1983; 17: 693-706. [35] Baker AJM., Accumulators and excluders: strategies in the response of plants to heavy metals. J Plant Nutr 1981; 3: 643-654. 19 Table 1. Mean element concentrations (mg kg-1 ± SD) in soil samples collected in different sites. N: soil of Nerium oleander; L: soil of Lantana camara. Sampling site Ba Cd Cu Fe (%) Mn Ni Pb V Ti N, POR 266±28* 6.50±0.44* 72.1±9.8* 2.25±0.16* 554±23 23.7±1.2* 174±16* 44.5±5.3* 1162±315* N,TRN 339±60* 6.33±0.82* 71±17* 2.24±0.29* 612±73 23.7±2.3* 142±29* 43.9±5.6* 1690±234* N,BLK 94.3±8.1 10.30±0.92 30.3±5.4 3.67±0.40 612±89 13.2±3.4 22.4±4.2 102±14 6232±1639 L, POR 294±36* 7.8±1.1* 80±21 2.71±0.54* 583±74* 29.3±7.1* 123±20* 57.9±7.6* 1687±449* L, TRN 189±74* 6.58±0.89* 77±49 2.28±0.33* 470±50* 18.9±3.7* 92±57* 48±12* 1776±562* L, BLK 86.0±3.7 10.60±0.80 25.7±1.8 3.87±0.16 649±56 10.60±0.50 20±12 110.0±5.6 8353±1490 Madrid et al. (2002) n.a. n.a. 68±64 2.01±0.32 471±103 22±6 137±160 n.a n.a. De Miguel et al. (1998) 369±64 n.a. 72±37 2.31±0.35 437±118 14±4 161±82 30±3 2135±319 *Indicate significant (Scheffé test, p<0.05) statistical differences with respect to the BLK. n.a., not available. 20 Table 2. Correlation matrix of the total element concentrations in N. oleander and L. camara soils N. oleander Ba Cd Cu Fe Mn Ni Pb V Ti Ba 1.00 Cd -0.71 1.00 Cu 0.88 -0.64 1.00 Fe -0.70 0.99 -0.66 1.00 Mn 0.15 0.37 0.00 0.41 1.00 Ni 0.87 -0.76 0.81 -0.75 0.17 1.00 Pb 0.81 -0.78 0.88 -0.81 -0.21 0.86 1.00 V -0.78 0.97 -0.74 0.98 0.30 -0.85 -0.87 1.00 Ti -0.75 0.92 -0.75 0.93 0.22 -0.90 -0.88 0.97 1.00 L. camara Ba Cd Cu Fe Mn Ni Pb V Ti Ba 1.00 Cd -0.03 1.00 Cu 0.75 -0.47 1.00 Fe -0.26 0.97 -0.45 1.00 Mn 0.10 0.76 -0.35 0.84 1.00 Ni 0.76 -0.13 0.44 -0.17 0.15 1.00 Pb 0.87 -0.62 0.87 -0.67 -0.38 0.51 1.00 V -0.50 0.95 -0.62 0.95 0.73 -0.42 -0.76 1.00 Ti -0.62 0.52 -0.46 0.72 0.58 -0.47 -0.63 0.87 1.00 Correlations significant at p < 0.05 are in bold 21 Table 3. Factor loadings for the metal contents in oleander and lantana soil samples. Marked loadings are >0.7. Variable PC1 PC2 Oleander soil Ba 0.92 0,21 Cd - 0.86 0,44 Cu 0.89 0,12 Fe - 0.86 0,47 Mn - 0.0 1 0,96 Ni 0.96 0,16 Pb 0.92 - 0,15 V -0.93 0,34 Ti - 0.94 0,25 Eigenvalue 6.8 1.4 % Var 76.1 15.7 Lantana soil Ba - 0,19 -0,97 Cd 0,93 0,21 Cu - 0,38 -0,76 Fe 0,97 0,19 Mn 0,90 - 0,14 Ni 0,13 - 0,87 Pb - 0,49 -0,78 V 0,87 0,45 Ti 0,71 0,58 Eigenvalue 5.8 2.1 % Var 64.0 21.1 % Var, percentage of explained variance. 22 Table 4. Mean CF value for all elements in oleander soil and lantana soil. Sampling site Ba Cu Fe Mn Ni Pb V Ti N, POR 0.086 0.067 0.005 0.076 0.007 0.004 0.016 0.008 N,TRN 0.097 0.099 0.005 0.030 0.004 0.006 0.014 0.005 N,BLK 0.243 0.086 0.002 0.039 0.000 0.012 0.005 0.002 L, POR 0.036 0.149 0.008 0.030 0.004 0.010 0.022 0.009 L, TRN 0.031 0.078 0.003 0.050 0.008 0.005 0.009 0.003 L, BLK 0.043 0.303 0.003 0.052 0.004 0.038 0.008 0.003 23 FIGURE CAPTIONS Figure 1. Location of sampling sites in the studied area of Seville Figure 2. Hierarchical clustering results (dendrogram) of heavy metal concentrations in oleander (a) and lantana (b) soil samples. Figure 3. Hierarchical clustering results (dendrogram) of heavy metal concentrations in topsoil, PM10 and oleander leaf samples. OS, oleander soil, P, PM10, O, oleander leaves. Figure 4. Mean EF value for all elements. a), N. oleander soil; b), L. camara soil. The horizontal line indicates the EF threshold =2. 24 BLK (40 Figure 1