Waste management facilities in general, and incinerators in particular, have been traditionally affected by the NIMBY (Not In My Back Yard) syndrome (Kuhn and Ballard 1998). Although in comparison with other treatments for processing municipal solid waste (MSW) and hazardous waste (HW), incineration has multiple advantages (i.e., volume reduction, energy recovery, elimination of pathogen agents), public opposition to the siting and permitting of MSW and HW incinerators has been considerable (Chang et al. 2002, 2003; Domingo 2002a, 2002b; Hamer 2003). The main concern is related to the potential adverse consequences of the emission of pollutants for both the environment and public health (Schuhmacher et al. 1997a, 1997b; Domingo et al. 2002c).

Combustion processes have been catalogued as important contributors to environmental metal pollution. Among these processes, special attention has been paid to incineration (Hasselriis and Licata 1996; Thipse and Dreizin 2002). The number of hazardous waste incinerators (HWI) is notably lower than that of municipal solid waste incinerators (MSWI), and the volume of waste to be treated is also remarkably different. However, taking into account the potential number and variety of toxic residues contained in industrial wastes, these are expected, in principle, to be more dangerous than municipal wastes (van Veizen et al. 2002).

Environmental monitoring programs are common tools to control the state of pollution in a particular area (Rumbold et al. 1997; Morselli et al. 2002; Chang et al. 2003). These programs are used to assess the contribution of a potential contaminant source on the global environment. Because they can be easily collected and stored, soil and vegetation are two of the most standardized environmental monitors (Mumma et al. 1990; Klumpp et al. 1994, 2003; Moraes et al. 2002). Soil and vegetation have been widely used as cumulative matrices of long-term and short-term exposure, respectively, to environmental pollutants (Meneses et al. 1999; Domingo et al. 1999, 2001; Schuhmacher et al. 1996, 2000, 2002, 2003; Nadal et al. 2004).

In 1996, the construction of a HWI in Constantí (Tarragona, Spain) was initiated. Regular operations started in 1999. Because this facility was the first, and up until now the only HWI in Spain, the concern about its potential environmental impact and health risks has been, and still remains, considerable in public opinion. In response to that concern, a wide preoperational monitoring program was designed to evaluate the potential impact of emissions from the new HWI on the neighborhood, as well as to assess the health risks to the population living near the facility. Generally, evaluation of HWI stack emissions has been focused on heavy metals, semivolatile and volatile compounds, with special attention being paid to metals (carcinogenic elements such as cadmium and chromium are of particular concern), and dioxins and furans (Lisk 1988; Sedman and Esparza 1991; Dempsey and Oppelt 1993; Schuhmacher et al. 2002). A baseline survey was finished in 1998, just before the HWI began operating. Soil and herbage samples were collected in the vicinity of the facility under construction and were analyzed for metals (Llobet et al. 2000) and dioxins and furans (Schuhmacher et al. 1997a, 1998, 2000).

In 2001, 3 years after the baseline survey, the concentrations of metals and dioxins and furans were again determined in soil and herbage samples collected at the same points in which samples had been taken in the baseline survey (unpublished data). Recently, samples of both matrices were again collected and analyzed for metal levels. The results are here reported and compared with those found in the 2001 survey and the baseline surveys. On the other hand, the present study is also focused on establishing the health risks of metals for the population living in the neighborhood of the HWI.

Materials and Methods

Sampling

In April 2003, soil samples were collected at the same 40 sampling sites used in the baseline survey (1998) and the 2001 survey (Llobet et al. 2000). These points were selected taking into account the main wind directions and in a radius of 7 km from the HWI (Figure 1). Ten samples were collected in urban zones, whereas the remaining 30 were considered to belong to rural zones. Approximately five subsamples were collected as bulk samples at the same time within approximately 20 m2 in each sampling site. Soil samples were taken from the upper 3 cm of soil and stored in polyethylene bags for easy handling. In the laboratory, they were dried at room temperature and sieved through a 2-mm-mesh screen in order to obtain a homogeneous grain distribution.

Figure 1
figure 1

Sampling points of the area under study and wind rose (Departament de Medi Ambient 2002).

Full size image

Vegetation samples were also taken at the same 40 sampling points and at the same time as soil samples (April 2003). No rainfall was observed during the days before collection of samples. Herbage (Pipatherum paradoxum L) samples were obtained by cutting at about 5 cm from the ground. They were immediately stored in a double-aluminum fold and dried at room temperature.

Analytical Procedure

Approximately 0.5 g of dried soil and herbage samples were treated with 5 ml of HNO3 (65% Suprapur, E. Merck, Darmstadt, Germany) in hermetic Teflon bombs and maintained at room temperature for 8 h. Subsequently, they were heated at 80°C for 8 h. After cooling, solutions were filtered and made up to 25 ml with deionized water (Meneses et al. 1999; Schuhmacher et al. 2003).

The concentrations of arsenic (As), beryllium (Be), cadmium (Cd), chromium (Cr), mercury (Hg), manganese (Mn), lead (Pb), tin (Sn), thallium (Tl), and vanadium (V) in soils were determined by inductively coupled plasma spectrometry (ICP-MS, Perkin Elmer Elan 6000), while atomic absorption spectrometry with graphite furnace atomization (Varian spectrophotometer, Spectra A-30) was used to determine nickel (Ni) concentrations. The levels of As, Be, Cd, Hg, Mn, Pb, Sn, and Tl in herbage samples were assessed by ICP-MS, while the concentrations of Cr, Ni, and V were determined by atomic absorption spectrometry with graphite furnace atomization.

The limits of detection (LOD) for soils and herbage were the following: 0.1 μg/g for As; 0.025 μg/g for Cd, Mn, Ni, Pb, and Tl; 0.05 μg/g for Hg and Sn; and 0.25 μg/g for Be. In herbage, detection limits of Cr and V were 0.04 and 0.1 μg/g, respectively, whereas in soils, the LOD for these elements was 0.25 μg/g. The accuracy of the methodology was checked by determining the levels of duplicate samples, as well as those of blanks (control samples). Moreover, a reference material (Lobster hepatopancreas, NRC Canada, TORT 2), which was run after every 10 samples, was also used to check for drift in the sensitivity of the instrument. Recovery rates ranged from 79% to 123% for soils and from 82% to 124% for vegetation.

Statistics

In the case of values under the respective detection limits, concentrations were assumed to be one-half of the limit of detection (ND = 1/2 LOD). Statistical significance of the data was computed by one-way analysis of variance followed by Student’s t-test or by the Kruskal-Wallis test. A probability of 0.05 or lower was considered significant. All the analyses were done using the statistical package SPSS 11.0. A multivariate analysis was also done. Data were evaluated through principal component analysis (PCA), which is a multivariate technique for reducing matrices of data to their lowest dimensionality. It allows identification of the number of significant factors (principal components) and use of this information for classification.

Results and Discussion

The concentrations of metals in soils and herbage collected in 2001 and 2003 in the vicinity of the new HWI, together with the results of the baseline (1998) survey, are shown in Table 1. The percentages of variation during the periods 1998–2003 and 2001–2003 are also given. During the period 1998–2003, As, Be, Cr, Ni, and V levels showed significant increases in soils. In contrast, the levels of Cd, Hg, and Sn significantly diminished. With respect to herbage, Cr, Mn, and V concentrations were significantly increased, while As levels significantly diminished (p < 0.001). Although the 2003 concentrations of Cd, Pb, and Sn in vegetation were lower than those of the baseline (1998) study, the differences did not reach the level of statistical significance. Vanadium was the only element showing a similar tendency in both matrices: increases of 54.4% and 100% in soils and herbage, respectively. In general terms, important fluctuations were observed in the levels of metals in soils and vegetation between the 1998 and 2003 surveys. It must be taken into account that in addition to combustion processes, environmental levels of metals can be notably influenced by a remarkable number of different emission sources. The HWI is located in an industrial zone, where several chemical and petrochemical companies are settled in. Moreover, the area is surrounded by a highway and several important roads. Consequently, the potential impact of other sources must be specially considered. On the other hand, no statistically significant correlation (Pearson) between soil and herbage concentrations was found for any of the analyzed elements. This means that root uptake of heavy metals by plants can be considered negligible. This was an expected finding, because only the A soil horizon (upper layer) was sampled. Likewise, the lack of a homogeneous change between soil and vegetation levels also means that other potential emission sources, different from incineration, could be even more important than they were supposed.

Table 1 Metal concentrations (μg/g ± SD) in soil and herbage samples collected near the HWIa
Full size table

Tables 2 and 3 summarize the concentrations of metals in soils and herbage, respectively, according to the specific zones of sampling, urban or rural. Data are given for the 1998 and 2003 surveys. In the present study, the levels of all elements were higher in rural soils than in those collected in urban zones. The differences were statistically significant for Be and Tl in 2003, whereas in the baseline survey the level of significance had been reached only for Hg and Sn. The study of the temporal variation of the data shows that the rural soil samples collected in 2003 presented higher metal concentration than those found in 1998. The differences were statistically significant for all elements, with the exception of Mn, Pb, and Tl. The homogeneous increase in the concentrations observed in rural samples was not noted in those collected in urban zones, in which Ni and V levels showed significant increases, while Hg concentrations decreased (Table 2).

Table 2 Metal concentrations (μg/g ± SD) in urban and rural soil samples collected in 1998 and 2003
Full size table
Table 3 Metal concentrations (μg/g ± SD) in urban and rural herbage samples collected in 1998 and 2003
Full size table

In vegetation, the differences according to the area of collection were not so evident. Thus, only Sn levels found in rural samples were significantly higher than those observed in urban samples. Manganese concentrations were significantly increased between 1998 and 2003 in both urban and rural samples. Other elements also presented significant changes during this period in rural samples. The levels of Cr increased, while those of V decreased. In turn, only Sn concentrations showed a significant decrease in urban samples.

To assess the influence of the HWI on the area under its direct influence, sampling points collected at the main wind directions (N, NW, E, and S) were distributed into 2 groups: “a” (those located up to 1500 m from the stack), and “b” (those located at greater distances). This classification was based on data such as the height of the HWI stack (55 m) and the wind regime of the zone (predominant winds: N and NW). A comparison of the 1998 and 2003 metal levels in soil and herbage samples is depicted in Figures 2 and 3, respectively. For soils, significant changes were not found for any wind direction in the “a” zone. With respect to the “b” area, significant changes were noted for Be, Hg, Ni, Tl, and V in the NW direction, as well as for Cd, Hg, Ni, and Pb in the S direction. Significant differences at the N, S, and E directions were also observed for various elements in herbage samples (Figure 3).

Figure 2
figure 2

Concentrations of metals in soils (μg/g) according to the directions of the wind (N, NW, E, and S) and the distance to the HWI (zones a and b). An asterisk means significant differences between 1998 and 2003.

Full size image
Figure 3
figure 3

Concentrations of metals in herbage (μg/g) according to the directions of the wind (N, NW, E, and S) and the distance to the HWI (zones a and b). An asterisk means significant differences between 1998 and 2003.

Full size image

The significant differences in metal concentrations appeared mainly at distances more than 1.5 km from the plant. It might be an indicator that the HWI is not the main contributor of the presence of metals into the environment of the area under evaluation. Data fluctuations would be due to the presence of other pollution sources (traffic, industrial activities, etc.), and/or to the environmental variations of these elements in the zone. In general terms, the current metal concentrations in soil and vegetation were very similar to those obtained in previous studies in the vicinity of a MSWI also operating in Tarragona (Llobet et al. 1999, 2002), and in an important petrochemical complex located in Tarragona County (Nadal et al. 2004). They were also similar to those found in recent years by our research group in samples collected in zones under the influence of another MSWI (Schuhmacher et al. 1997a; Meneses et al. 1999) and a cement plant (Schuhmacher et al. 2003) of Catalonia.

In order to study temporal changes of the heavy metals concentration in soil and vegetation, multivariate analysis (PCA) was carried out. The input data corresponded to 80 samples collected in 1998 and 2003. In soils, PCA provided a three-dimensional model, which accounted for 81.3% of the variance. The first principal component (PC), which explained 34.7% of the variance, was positively correlated with As, Be, Mn, Cr, and Ni. The second PC (25.3% of the variance) was correlated with Sn, Tl, and V. Finally, the third PC (21.3% of the variance) showed a correlation with Cd, Hg, and Pb. With regard to herbage samples, the three-dimensional model accounted for 66.8% of the variance. The first component correlated positively with Cr, As, Pb, and V and explained 37.6% of the variance. The second PC was correlated with Ni and the third PC with Mn and Hg (15.1% and 14.1% of the variance, respectively). Scatter plots for soil and vegetation samples are depicted in Figure 4a and 4b and in Figure 5a and 5b, respectively. In general, most soil and herbage samples form a cluster. However, no characterization was observed according to the year of collection. Some samples showed high concentrations of some components, although only a few of them were collected near the incinerator. It can be determined that these samples could be contaminated by other emission sources of metal different from the HWI assessed here. Consequently, PCA allowed establishing again that the influence of the incinerator over the surrounding environment could not be considered more important than other potential sources.

Figure 4
figure 4

(a) Principal component analysis. Plot PC1/PC2 for metals in soil samples. (b) Principal component analysis, Plot PC1/PC3 for metals in soil samples.

Full size image
Figure 5
figure 5

(a) Principal component analysis. Plot PC1/PC2 for herbage samples. (b) Principal component analysis. Plot PC1/PC3 for herbage samples.

Full size image

Risk assessment consisted of comparing the current metal concentrations in soils with Preliminary Remediation Goals (PRG), considered as safe values for people living in residential areas (US EPA 2004). In relation to noncarcinogenic risks (Figure 6), all the elements presented a value inside the safe interval. Among them, V and Mn were the metals with the highest percentage of soil screening level. With respect to the carcinogenic elements (Figure 7), Cd and Cr were included in the safe interval, while As levels exceeded clearly the regulatory limits. The reason could be the relative low levels marked by the US EPA as PRG for As, in comparison with the environmental concentrations of this element usually found. Nevertheless efforts should be focused on minimizing the release of As to the environment.

Figure 6
figure 6

Noncarcinogenic risk: Comparison between metal concentrations in soils collected in the vicinty of the HWI and Preliminary Remediation Goals.

Full size image
Figure 7
figure 7

Carcinogenic risk: Comparison between levels of carcinogenic elements in soils collected in the vicinity of the HWI and Preliminary Remediation Goals.

Full size image

Table 4 presents the predicted oral exposure for adults and children living in residential areas near the HWI. Criteria for the calculations were previously reported (Nadal et al. 2004; US EPA 2004). Hazard Quotient for all metals was less than 1, which is considered the upper bound. Table 5 summarizes the predicted results for inhalation exposure of people living in the same area. In this case, a mean concentration of particulate matter in air of 71 μg/m3 was assumed (Departament de Medi Ambient 2002).

Table 4 Predicted oral exposure of metals for adults and children and hazard quotients
Full size table
Table 5 Predicted exposure (mg kg−1 day−1) corresponding to metal inhalation by adults and children
Full size table

With respect to carcinogenic risks, only those elements for which a slope factor has been established were assessed: As ingestion, and As, Cd, and Cr inhalation. The risk of cancer was recalculated from the oral and the inhalation predicted exposures. Results are shown in Table 6. It can be noted that As ingestion would exceed the safe value of 10−6, especially when assessing cancer risk in children. Some considerations should be noted: since the PRG used here corresponds to residential areas, this is the most conservative scenario. However, most samples were collected in rural zones. In addition, no age adjustment was carried out for calculations in children, overestimating the real value. Apart from As, only Cr inhalation could play a positive role in the potential increase in the number of cancers. The reason is that PRG for total chromium is established on the assumption that the ratio between Cr+6 and Cr+3 concentrations in soil is 1:6 (US EPA 2004). In the current study, no speciation was carried out. Therefore, this conservative calculation was taken into account.

Table 6 Risk of cancer due to ingestion and inhalation of carcinogenic elements
Full size table

Compared with metal levels found in recent years in a number of industrial areas (Morselli et al. 1993; Adamo et al. 2002; Sterckeman et al. 2002; Tuzen 2003; Yusuf et al. 2003), the current metal concentrations in the area under potential influence of the HWI are in the low part of these levels. Recently, Morselli et al. (2002) examined the impact of an Italian MSWI on the surrounding environment. Levels of heavy metals in soils and vegetation similar to those found near the HWI assessed here were observed. On the other hand, the concentrations reported by a number of investigators in urban soils were higher than those found in the present study (Kaminski and Landsberger 2000; Mielke et al. 2001; Granero and Domingo 2002). The current values are even lower than those reported as background by Frink (1996).

In conclusion, the levels of metals found in soil and vegetation samples collected near the HWI do not show a continuous and homogeneous increase during the period 1998–2003 (approximately 5 years of regular operations). The fluctuations in the concentrations would suggest that the influence of the facility is minimal in relation to other metal pollution sources. Further studies should be focused on discovering the contribution of these sources to the global pollution of the area.