Journal of Soils and Sediments

, Volume 10, Issue 2, pp 290–300

Soil contamination by organic and inorganic pollutants at the regional scale: the case of Piedmont, Italy

Authors

    • DI.VA.P.R.A., Chimica AgrariaUniversita` di Torino
    • A.R.P.A.-Piemonte
  • Mattia Biasioli
    • DI.VA.P.R.A., Chimica AgrariaUniversita` di Torino
  • Renzo Barberis
    • A.R.P.A.-Piemonte
  • Franco Ajmone-Marsan
    • DI.VA.P.R.A., Chimica AgrariaUniversita` di Torino
SOILS, SEC 3 * REMEDIATION AND MANAGEMENT OF CONTAMINATED OR DEGRADED LANDS * RESEARCH ARTICLE

DOI: 10.1007/s11368-009-0114-9

Cite this article as:
Fabietti, G., Biasioli, M., Barberis, R. et al. J Soils Sediments (2010) 10: 290. doi:10.1007/s11368-009-0114-9
  • 424 Views

Abstract

Background, aim, and scope

Diffuse soil contamination has often been neglected in scientific literature, as most studies focus on contaminants from point-sources (either of industrial or agricultural origin). However, soil pollution from diffuse sources is recognized as one of the major soil threats by the EU Soil Thematic Strategy. In fact, some pollutants are nowadays ubiquitarious in the soil system, and they have to be considered for the implementation of environmental legislation, the definition of clean-up values in remediation activities and, more generally, for a sustainable management of rural areas. In the literature large scale studies on diffuse contamination are few and scattered and often do not consider a wide range of contaminants, the effect of land use, the vertical variability and the potential natural contribution. Aim of this work was to provide an overview of the diffuse soil contamination on a regional scale for a large set of contaminants. Soil inorganic (Cd, Cr, Ni, Pb, Cu, Zn) and organic contaminants (polychlorinated dibenzo-p-dioxins(PCDD), polychlorinated biphenyls (PCBs), and polycyclic aromatic hydrocarbons (PAHs) as well as other soil general parameters were investigated on a 18 × 18 km grid covering a whole Italian region heavily industrialized and intensively cultivated. Soils were sampled at different depths both for natural–undisturbed soil and for agricultural–plow soil. Differences in the trends of investigated contaminants, as a consequence of land use, vertical variability, natural or geogenic origin, relationships among contaminants, and with main soil properties, were explored by means of enrichment factors, bi- and multi-variate statistics.

Materials and methods

The sampling scheme of this study is based on a systematic 18 × 18 km grid covering the whole region. Overall, 43 monitoring sites located at the center of each cell were sampled. At each site, five individual core samples within a 10 × 10 m area were taken at different depths for natural–undisturbed soil (topsoil 0–10 cm, subsoil 10–30 cm) and for agricultural–plow soil (topsoil Ap horizon, subsoil 20 cm below the Ap lower limit). Samples were processed for general soil properties as well as for organic (PAHs, PCBs, PCDDs/dibenzofurans (DFs)) and inorganic (Cd, Cr, Ni, Pb, Cu, Zn) contaminants analyses (aqua regia digestions). Contaminants were detected after extraction by means of ICP-MS and GC-MS. Statistical analysis was conducted using the software SPSS 13 (SPSS) and Minitab 15 (MINITAB). Data were geographically managed and processed with the Arcview 3.2 (ESRI, CA, USA) GIS software.

Results

Organic contaminants such as PCDDs were found to accumulate in natural areas while inorganics mostly concentrate in agricultural soils, confirming the presence of important phenomena of long-range diffuse contamination for the first and short-range for the latter. Soil use was also confirmed to be a major parameter in influencing the type, degree, and distribution of contaminants. Vertical variability was found to be high for organic contaminants and for Pb, suggesting their main anthropogenic origin, while other elements such as Cr and Ni appeared to be more related to the natural background. Data were compared with those from soils of a large industrial city present in the area, confirming the strong enrichment of the urban environment with respect to some contaminants such as metals, PCBs, and PAHs. Other contaminants such as PCDD/DFs showed lower differences, confirming their diffuse and almost ubiquitarious pollution.

Discussion and conclusions

Even if natural soils in this study are mostly located far from major sources of contamination, notable differences appeared when compared to agricultural areas. In particular, the enrichment in the concentrations of organic contaminants such as PCDDs in natural areas and of inorganics in agricultural soils confirm the presence of important phenomena of long-range diffuse contamination for the first and short-range for the latter, which appeared also to be related to agricultural activities. Contaminants like PCDDs, PCBs, PAHs, and Pb presented high vertical variability, confirming their anthropogenic origin and strong affinity to soil organic matter, while others appear to be more related to the natural background. The comparison of data with those from soils within a large-industrial city present in the study area confirms the strong enrichment of the urban environment with respect to some contaminants such as metals, PCBs, and PAHs. Other contaminants such as PCDD/DFs showed lower differences confirming their diffuse and ubiquitarious pollution.

Recommendations and perspectives

Soil diffuse contamination revealed to be an important source for some contaminants that appear to distribute over a large scale. The type of land use strongly influences the distribution of pollutants in soils, by diluting them in depth (plowing in agricultural areas) or concentrating contaminants in the surface (natural areas or meadows). Data provided in this study constitute an important dataset of the soil environmental quality on a large scale that can be used for the development of guidelines for soil management, the definition of local clean-up values, and the implementation of risk assessment procedures.

Keywords

Diffuse pollutionHeavy metalsPAHsPCBsPCDD/DFsSoil

1 Background, aim, and scope

The diffuse pollution of soils is generally associated with atmospheric deposition of a variety of contaminants caused by a wide range of industrial activities, farming practices, and inadequate waste and wastewater recycling.

Diffuse contamination is a problem of high relevance, as it can bias the identification of sources of pollution in soils and complicate or nullify remediation strategies.

As a consequence of diffuse contamination, soils may lose fundamental functions with a decrease in their overall environmental quality (Mico et al. 2006). The Thematic Strategy for Soil Protection of the European Commission (2006) recognizes diffuse pollution as a threat to the soil quality.

Contaminants from diffuse sources are commonly persistent organic pollutants (POPs) and heavy metals (Tremolada et al. 2008). POPs are highly toxic molecules, recalcitrant to degradation, and some are carcinogenic or mutagenic. Among them are polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/DFs), polychlorinated biphenyls (PCBs), and polycyclic aromatic hydrocarbons (PAHs). While PAHs and PCDD/DFs are still being emitted mostly by industrial activities in large quantities into the environment, emissions of PCBs have been declining since their prohibition (1984 in Italy) (Katsoyiannis and Samara 2004). POPs can be transported in the atmosphere over short and long distances in gaseous or particulate forms and atmospheric dry and wet deposition constitutes the main input of these compounds to the soil (Cousins et al. 1999). Indirect deposition of compounds can occur via vegetation (Wania and Mclachlan 2001), where chemicals taken up by plants may enter the soil as litter falls to the ground and decays. POPs would then accumulate in horizons rich in organic matter (OM) where they may be retained for years (Masih and Taneja, 2006). It has been estimated that the soil in the UK contains about 90% of the total environmental burden of POPs in the country (Cousins and Jones 1998).

Natural contents of heavy metals in soils depend on the geological parent material, soil forming processes, and anthropogenic activities. In the past decades, the natural input of heavy metals to soils has been exceeded by the anthropogenic input (Shi et al. 2007) due to traffic, industry, waste disposal, and agricultural practices (Weber and Karczewska 2004). Among agricultural practices, the use of superphosphate fertilizers has been related to soil contamination by cadmium, while calcium nitrate can contain significant amounts of nickel. Certain fungicides contain of copper and zinc, which can increase availability of these elements in the upper soil horizons (Lopez-Mosquera et al. 2005).

After being deposited on the soil surface, soil use becomes an important factor that determines the vertical and/or horizontal distribution of pollutants. In agricultural soils, the dilution with uncontaminated subsoil by plowing decreases the concentration of contaminants. In natural soils, the lower disturbance, high levels of organic matter, and deposition of decaying plant litter generally enhance the accumulation of contaminants in the topsoil (Cousins and Jones 1998).

The knowledge of the extent of diffuse contamination is crucial for the correct implementation of environmental legislation and for the management of contaminated sites. In fact, some contaminants are nowadays ubiquitarious and the appraisal of their background content is necessary in the definition of clean-up values and mitigation strategies. A large number of studies have been conducted on diffuse soil contamination but data are scarce at the regional scale representing a typical territorial basis for the implementation of legislative action.

Aim of this study was to provide an appraisal of the diffuse contamination in soils at a regional scale for a range of organic and inorganic compounds, whose presence might be more related to long-range transport than to emissions from a point source. Associations among contaminants and soil properties, influence of the different land uses, vertical variability were also investigated.

2 Materials and methods

2.1 Study area

The Piedmont Region (northwestern Italy) covers 25,400 km2 with a population of around 4,500,000. The regional landscape is characterized by a flat area (alluvial plain) in the middle, surrounded by a mountainous chain. Soils in the plain have developed on recent alluvial sediments derived from the erosion of the Alpine and Apennine chain. The land is locally heavily industrialized and intensively cultivated in the plains. The region represents a good example of the areas of central Europe where the development of heavy industry overlapped with modernization of agriculture in the last 60 years.

2.2 Soil sampling

The sampling scheme of this study is based on the systematic grid 18 × 18 km designed within the LUCAS project (European Community 2003) and covering the whole region (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs11368-009-0114-9/MediaObjects/11368_2009_114_Fig1_HTML.gif
Fig. 1

View of Piedmont region and of the sampling points. Sampling points are located on the corners of a systematic 18 × 18 km grid. Black squares are agricultural soils and white circles are natural areas

LUCAS is a survey launched by Eurostat in cooperation with DG-Agriculture of the European Union that aims at obtaining harmonized data at EU level on land use, land cover, and additional environmental features.

The monitoring has been running since 2001 and was designed to have reliable georeferenciated data of the environmental soil quality and to monitor both organic and inorganic pollutants in soils from diffuse contamination.

During spring and summer 2003, 43 monitoring sites located at the corners of each 18 × 18 km cell were sampled. At each site, five individual core samples within a 10 × 10 m square (four on the corners of the square and one in the middle) were taken from the topsoil and the subsoil layers, at different depths for natural–undisturbed soil (topsoil 0–10 cm, subsoil 10–30 cm) and for agricultural–plow soil (topsoil Ap horizon, subsoil 20 cm below the Ap lower limit). Core samples were mixed to obtain a composite sample for each soil layer. Of the 43 sites, 24 were classified as natural and 19 as agricultural. The geographical distribution among natural and agricultural soils can be observed in Fig. 1. The final 86 samples were analyzed for their general soil properties, total acid-extractable metal contents (Cd, Cr, Cu, Ni, Pb, and Zn), and for PCDD/DFs, PCBs, and PAHs concentration.

2.3 Chemical analyses

2.3.1 General soil properties

Samples were air-dried, gently crushed in ceramic mortars and sieved at <2 mm. A portion of each sample was further ground at <0.15 mm for aqua regia (HCl/HNO3, 3:1 solution) digestion (ISO 11466 1995). Particle size distribution was determined by the pipette method (Mipaf 1999). The pH was determined in a KCl solution, 1:2.5 soil/solution ratio (Mipaf 1999), organic carbon by dry combustion (ISO 10694 1995), and cation exchange capacity with BaCl2 at pH 8.1 (ISO 13536, 1995).

2.3.2 Metal analysis

The aqua regia extracts were analyzed for metals with an ICP-MS (Agilent, 7500CE). Duplicates were made for all samples and results accepted when the coefficient of variation was within 5%. A blank and the Certified Reference Material 141 R (Community Bureau of Reference, Geel, Belgium) were included in each batch of analyses for quality control. Results were satisfactory when within a range of ±10% from the certified value.

2.3.3 Organic contaminant analysis

Extraction of organic contaminants was carried out using an ASE 200 Accelerated Solvent Extractor (Dionex, Sunnyvale, CA, USA). Extractions were conducted using dichloromethane for PAHs, and toluene for PCBs and PCDDs/DFs. Extracts were purified in an automated Powerprep system (FMS, Waltham, MA, USA) and analyzed for PAHs after concentration to 1 ml (EPA 8270D 1996). The concentrations and profiles of PAH compounds were determined using a high-resolution gas chromatograph Agilent 6890, interfaced with a high-resolution mass spectrometer Agilent 5973 (Agilent, CA, USA). Identification and quantification of 16 PAH compounds (the so-called EPA list) was based on matching their retention time with a mixture of PAH standards. The fractionation and purification of PCBs and PCDDs/DFs after ASE extraction was performed by an automated Powerprep system (FMS, Waltham, MA, USA). The identification and quantification of PCBs and PCDDs/DFs (EPA 1613B 1994; EPA 1668A 1999) was carried out with the same instrumentation used for PAHs. The capillary column was 30 m long, with a 0.25 mm i.d. and 0.25 mm stationary phase film thickness (HT8, SGE Analytical Science, Australia). Thirty-two PCBs and 17 PCDDs/DFs isomers and congeners were determined by selected ion monitoring at the two most intensive ions of the molecular-ion cluster standards. Known amounts of 13C-labeled PCBs and PCDDs/DFs were added as internal standards. Recoveries ranged 75–92%, and concentrations were corrected for the recoveries accordingly. Results of dioxin congeners were expressed as I-TE (toxicity equivalents), calculated using toxicity equivalency factors reported by the World Health Organization (Van den Berg et al. 1998).

2.4 Data handling

Statistical analysis was conducted using the software SPSS 13 (SPSS) and Minitab 15 (MINITAB). Data were geographically managed and processed with the Arcview 3.2 (ESRI, CA, USA) GIS software.

3 Results

3.1 General soil properties

Descriptive statistics of topsoil general properties by land use is reported in Table 1. Agricultural soils have a mean pH of 6.4, ranging 4.2–8.0, while natural soils have a mean pH of 4.9, ranging 3.4–7.5. The differences observed could be attributed to the lower organic matter content of agricultural soils (mean, 2.0%) than of natural ones (mean, 6.5%). This difference is a consequence of the biomass removal brought about by agricultural practices and of the litter accumulation typical of natural soils. Cation exchange capacity (CEC) in agricultural soils has a mean value of 25 cmol kg−1, ranging 17–37 cmol kg−1, while natural soils show a higher value, 37 cmol kg−1. The fairly high CEC contents observed can be attributed to the clay content of agricultural soils and to the OM enrichment of natural ones.
Table 1

Descriptive statistics of main physical–chemical properties of topsoils

 

Agricultural (n = 19)

Natural (n = 24)

pH KCl

OM

CEC

Sand

Silt

Clay

pH KCl

OM

CEC

Sand

Silt

Clay

%

cmol kg−1

%

%

cmol kg−1

%

Mean

6.4

2.0

25

43

42

15

4.9

6.5

37

60

33

8

Median

6.5

1.6

25

41

41

12

4.4

5.5

36

62

31

5

SD

1.2

1.0

6

18

12

11

1.4

4.5

15

22

17

8

Min

4.2

0.8

17

13

21

3

3.4

0.9

13

9

11

2

Max

8.0

4.8

37

76

63

44

7.5

17.5

76

87

80

35

3.2 Inorganic contaminants

In Table 2 the descriptive statistics of heavy metals content (aqua regia-extractable) of the soils are presented, compared with the Italian legislative limits for contamination for green and residential areas (no specific limits are currently adopted for agricultural soils). Both agricultural and natural soils show mean and median contents of heavy metals always below legislative limits. However, with the exception of Pb (agricultural soils) and Cd (natural soils), limits are always passed at least once in some of the samples, showing the presence of hot-spots of contamination, likely due to point-source contamination phenomena.
Table 2

Descriptive statistics of heavy metals content

 

Layers

Agricultural (n = 19)

Natural (n = 24)

Cd

Cr

Ni

Pb

Cu

Zn

Cd

Cr

Ni

Pb

Cu

Zn

mg kg−1

Mean

Topsoil

0.56

96

72

21

50

72

0.47

74

46

42

25

69

Median

0.30

79

66

20

22

68

0.29

56

34

34

16

58

SD

0.51

50

40

7

83

26

0.42

49

36

27

29

37

Min

0.05

45

17

10

9

34

0.29

18

6

11

1

28

Max

1.50

257

166

40

342

155

1.60

191

127

103

139

187

Mean

Subsoil

0.63

104

71

20

30

69

0.40

75

47

31

21

60

Median

0.29

93

61

18

25

67

0.29

59

36

29

14

50

SD

0.77

52

35

8

30

28

0.37

47

39

18

27

33

Min

0.04

49

16

7

6

28

0.05

18

5

8

1

21

Max

3.10

265

136

39

143

167

1.80

171

135

90

122

187

Italian limits

 

2

150

120

100

120

150

2

150

120

100

120

150

Results are presented divided by land use (agricultural and natural) and depth (topsoil and subsoil). Legislative limits (D.M. 152/06 for green and residential areas) are also shown

On the average, agricultural soils always show higher concentrations of metals than natural areas, except for Pb. This could reflect the influence of agricultural practices on the load of contaminants reaching the soil, which adds to the contribution of atmospheric deposition. The higher concentration of Pb reported for natural soils can be ascribed to their high organic matter content and to the origin of this element, mostly related to traffic, that contributed to a more diffuse deposition. Lead is in fact among the less mobile metals in soils and forms stable chelates with organic matter.

Average metal concentrations tend to decrease with depth, reflecting the accumulation in surface layers typical of diffuse contamination. This pattern is more evident for natural soils, as a consequence of plowing in agricultural areas.

Chromium and nickel show high contents in both land uses, with numerous samples above the legislative limit. In the literature, these elements are reported to range in world soils 10–40 mg kg–1 (Ni) and 50–100 mg kg−1 (Cr; Kabata-Pendias and Pendias 1992), but soils developing over ultramafic rocks can show much higher contents. In Piedmont, several outcrops of these rocks exist, that are responsible for the high values observed in the present study (Biasioli et al. 2006; Facchinelli et al. 2001; Lombini et al. 1998). This seems to be confirmed by the minimal differences observed between Cr and Ni content in topsoil and subsoil and between the two categories of land use considered. Copper concentrations double in agricultural soils (mean 50 mg kg−1) with respect to natural ones (mean 25 mg kg−1). This is probably to be attributed to vineyards highly present in some parts of the region, having received loads of pesticides based on CuSO4 for more than a century (Facchinelli et al. 2001). Cadmium and zinc do not show significant differences in their concentrations among the land uses and sampling depths.

3.3 Organic contaminants

All organics investigated were found to be below the legal limits for green and residential areas (Table 3). In contrast with what observed for heavy metals, organic contaminants content tends to be lower in agricultural soils compared to natural ones. This can probably be attributed to the plowing activities of agricultural areas and to the higher organic matter content of natural soils (Sweetman et al. 2005). In natural areas, moreover, a significant contribution can derive from the input of decaying plant material in afforested areas, as living vegetation has been shown to be an effective scavenger of organics from the atmosphere (Wania and Mclachlan 2001). A clear trend with depth for organic contaminants can be observed from data in Table 3, as mean and maximum values decrease from topsoil to subsoil. This corroborates the hypothesis of atmospheric deposition as major source of these contaminants and reflects their low vertical mobility.
Table 3

Means ± standard deviations of PCDDs/DFs (ng kg−1 I-TE), PCBs and PAHs (ng g−1), divided by land use and sampling depth

  

Agricultural (n = 19)

Natural (n = 24)

Italian limits D.Lgs. 152/06

Topsoil

Subsoil

Topsoil

Subsoil

 

Σ PCDDs/DFs

 

1.4 ± 0.7

1.0 ± 0.8

3.0 ± 2.6

1.5 ± 1.6

10

Σ PCBs

 

2.2 ± 3.5

1.7 ± 4.3

3.5 ± 3.3

1.6 ± 1.8

60

 

Rings

     

Naphthalene

2

5 ± 1.4

<1.0

9 ± 6.8

7 ± 5.9

 

Acenaphthylene

3

<1.0

<1.0

<1.0

<1.0

 

Acenaphthene

3

6 ± 1.9

<1.0

<1.0

<1.0

 

Fluorene

3

<1.0

<1.0

<1.0

<1.0

 

Phenanthrene

3

<1.0

<1.0

<1.0

<1.0

 

Anthracene

3

<1.0

<1.0

<1.0

<1.0

 

Fluoranthene

4

22 ± 47.1

19 ± 29.3

12 ± 17

9 ± 12.7

 

Pyrene

4

9 ± 8

10 ± 10.7

9 ± 11.4

6 ± 4.2

5,000

Benz[a]anthracenea

4

12 ± 7.5

10 ± 9.3

9 ± 7.5

7 ± 3.9

500

Chrysenea

4

12 ± 8.2

9 ± 6.9

15 ± 15.6

8 ± 5.9

5,000

Benzo[b]fluoranthenea

5

11 ± 15.8

12 ± 19

23 ± 22.6

13 ± 17.2

500

Benzo[k]fluoranthenea

5

6 ± 2.8

6 ± 3.6

8 ± 10.2

6 ± 3.9

500

Benzo[a]pyrenea

5

14 ± 10.5

11 ± 12.4

21 ± 17.3

11 ± 9.5

100

Indeno[1,2,3-cd]pyrenea

5

7 ± 5.3

8 ± 5.8

13 ± 12.9

6 ± 4.5

100

Benzo[g,h,i]perylene

5

7 ± 4.4

7 ± 5.2

10 ± 8.8

6 ± 3.4

100

Dibenz[a,h]anthracenea

6

<1.0

<1.0

<1.0

<1.0

100

Σ 16 PAHs

 

138 ± 76.4

129 ± 86.5

155 ± 119

110 ± 58.1

 

Σ PAHs > 4 rings

 

107 ± 75.9

96 ± 84.1

121 ± 114.8

75 ± 56.2

10,000

Σ PAHs carca

 

68 ± 45.1

60 ± 54.6

90 ± 76.5

56 ± 44.6

 

Sums of 16 PAHS, of PAHs above four rings, and of carcinogenic PAHs are also presented

aEPA Group B2—probable human carcinogens

The concentrations of PCDDs/DFs in surface agricultural topsoils range 0.6–3.9 ng kg−1 I-TE with a mean of a 1.4 ng kg−1 I-TE. Surface natural soils present higher values, ranging 0.7–8.7 ng kg−1 I-TE and a mean value (3.0 ng kg−1 I-TE) more than doubled compared with agricultural soils. Comparing the two sampling depths, a clear decreasing trend for PCDDs/DFs can be observed, especially in the case of natural soils that decrease from 3.0 (topsoil) to 1.5 ng kg−1 I-TE (subsoil).

PCBs in surface agricultural soils present a wide range, with a minimum value of 0.1, a maximum of 16.0 ng g−1, and a mean value of 2.2 ng g−1. Surface natural soils present a range similar to natural ones, varying from 0.2 to 14 ng g−1, with an average of 3.5 ng g−1. Subsoils always present lower values than topsoils, with a sharp decrease in natural areas, where PCBs decrease from 3.5 to 1.7 ng g−1. Agricultural soils decrease down to a mean value of 1.6 ng g−1 in subsoils.

In surface agricultural soils, PAHs have a mean value of 138 ng g−1 (calculated as the sum of the 16 compounds analyzed), ranging 80–304 ng g−1. Compared to agricultural areas, surface natural soils show a higher mean value (160 ng g−1), and a very wide range of concentrations, from 80 to 601 ng g−1. Agricultural subsoils show a mean PAH content of 129 ng g−1 and, in contrast with what observed for PCDDs/DFs and PCBs, do not differ markedly from mean values in surface. Mean PAHs concentration in natural soils decreases with depth, moving from 160 ng g−1 in the topsoil to 110 ng g−1 in the subsoil. Data presented by single compounds show, on the average, similar trends than those described for the sum of PAHs, with mean concentrations lower for agricultural than natural soils and a decrease in concentration with depth more evident for natural soils. Mean concentrations of the individual compounds are in general low and below the relative legislative limits, revealing a moderate degree of pollution by PAHs. Compounds with a number of rings <4 (naphthalene, acenaphthylene, acenaphthene, fluorene, phenanthrene, and anthracene), show concentrations often below the detection limit (<1 ng g−1). This is due to the fact that compounds with two to three rings are more subject to degradation and volatilization, due to their lower molecular weight, higher vapor pressure, higher fugacity ratio, and relatively higher solubility than the four-ring molecules (Nadal et al. 2004). As a consequence, these compounds are less abundant in soils when contamination is moderate and sources are distant. The PAHs found to be more abundant in agricultural and natural soils are benzo[b]fluoranthene, chrysene, fluoranthene, and benzo[a]pyrene. The latter has been recognized as the most carcinogenic among the PAHs (Srogi 2007), and represents an important tracer of PAHs diffuse contamination.

3.4 Vertical variability of contaminants

In general contaminants from diffuse sources tend to accumulate at the surface of soils, and their concentrations decrease with depth. For heavy metals, whose content in soils can also be influenced by parent material, concentrations can follow different patterns.

To test differences among contaminants at the two sampling depths considered and for the two land uses, a paired T-test was conducted. The test was run under the assumptions of equal variances, as confirmed by a Levene test for homoscedasticity. The results of the test for agricultural and natural soils are reported in Table 4.
Table 4

Results of the paired T-test on contaminants distribution at the two sampling depths (topsoil, subsoil)

 

Cd

Cr

Ni

Pb

Cu

Zn

ΣPCDDs

ΣPCBs

ΣPAHs

Agricultural (n = 19)

Stat t

−0.80

−1.64

0.03

1.53

1.39

0.93

3.32

2.03

0.74

P(T<=t)

0.43

0.12

0.98

0.14

0.18

0.37

0.00

0.06

0.47

t critical

2.10

2.10

2.10

2.10

2.10

2.10

2.10

2.10

2.10

Natural (n = 24)

Stat t

0.99

−0.22

−0.49

3.01

1.71

2.68

3.11

3.67

3.19

P(T<=t)

0.33

0.83

0.63

0.01

0.10

0.01

0.00

0.00

0.00

t critical

2.07

2.07

2.07

2.07

2.07

2.07

2.07

2.07

2.07

In bold are data where H0 could not be rejected (99% CI)

For Cr and Ni, regardless of the land use, no significant differences were found between topsoil and subsoil, confirming the main lithogenic origin of these metals in the soils investigated. Also Cd and Cu did not appear to be significantly different in both land uses, testifying the absence of major sources of contamination for these two elements. This is partly in contrast with what was observed for Cu in agricultural soils that, on average, presented an enrichment in the topsoil that can be attributed to some hot-spots of contamination, as well described by the high variability (range 9–342 mg kg−1 in the agricultural topsoils) of this element.

Among all the contaminants investigated, only PCDDs/DFs showed significant differences with depth in agricultural soils, whereas all organic contaminants as well as Pb and Zn where found to be statistically different in natural soils. This would confirm the effect of agronomic practices in agricultural soils and their low organic carbon content. Results for natural soils, instead, point to atmospheric deposition as a major source of contamination for Pb, Zn, and organics, and reflect the strong binding capacity of natural surface soils towards these pollutants.

3.5 Top enrichment factors

Surface accumulation of contaminants is sometimes assessed using the so-called top enrichment factor (TEF) (Boruvka et al. 2005; Facchinelli et al. 2001). Under the assumption of an enrichment of contaminants from diffuse sources in surface soil layers, the intensity of contamination is evaluated calculating TEF, by dividing surface mean content of contaminants with data from the deep layer. A TEF >> 1 indicates surface enrichment of contaminants, reflecting a likely atmospheric deposition. The TEFs calculated for metals, PCDDs/DFs, PCBs, and PAHs for agricultural and natural soils are presented in Fig. 2.
https://static-content.springer.com/image/art%3A10.1007%2Fs11368-009-0114-9/MediaObjects/11368_2009_114_Fig2_HTML.gif
Fig. 2

Top enrichment factor—TEF for agricultural (n = 19) (AGR) and natural (n = 24) (NAT) soils for inorganic and organic contaminants. The horizontal line indicates no enrichment

For all contaminants TEFs are always higher in natural than in agricultural soils, except for Cu, confirming the dilution of contaminants brought about by agricultural practices. Nickel and chromium in agricultural soils show a TEF of about 1, confirming their main lithological origin. Agricultural soils show some elevated TEFs for Cu (>1.5) and for PCDDs/DFs and PCB (>1.3). Natural soils show TEF > 1 for Cu, Zn, and Pb and very high TEFs for organic compounds, with values >1.5.

3.6 Association among contaminants

Pearson correlations and cluster analysis were applied to the raw data to investigate the relationships among contaminants. These techniques have been successfully applied to investigate associations among contaminants and to evaluate their potential sources (Zhang 2006; Boruvka et al. 2005) and can give important information on the study of soil diffuse contamination.

3.6.1 Pearson correlations

The correlation matrix among organic matter, Cr, Ni, Pb, Zn, PAHs, PCBs, and PCDDs/DFs for agricultural and natural topsoils is shown in Table 5. Other general soil properties as well as Cd and Cu are not presented as they did not show any significant correlation. A strong correlation between Cr and Ni for both agricultural and natural soils is evident, further substantiating their lithological origin. In natural soils organic matter is strongly correlated with PCDDs/DFs, PCBs, and PAHs, while this correlation disappears for agricultural soils, in line with the lower organic matter content of these soils and the disturbance of agricultural practices. Lead shows good correlations with Zn, PCDDs/DFs, PCBs, and PAHs for both agricultural and natural soils, suggesting a common origin for these contaminants regardless of the land use.
Table 5

Pearson correlation matrix among organic matter (OM), Cr, Pb, Zn, and organic compounds in agricultural and natural topsoils

 

Agricultural (n = 19)

Natural (n = 24)

OM

Cr

Pb

Zn

PCBs

PCDDs

OM

Cr

Pb

Zn

PCBs

PCDDs

Cr

−0.032

     

−0.453

     
 

0.897

     

0.026

     

Ni

−0.100

0.887

    

−0.417

0.877

    
 

0.684

0.000

    

0.043

0.000

    

Pb

0.223

0.221

    

0.748

−0.369

    
 

0.358

0.364

    

0.000

0.076

    

Zn

0.237

0.251

0.867

   

0.065

−0.037

0.535

   
 

0.330

0.299

0.000

   

0.762

0.865

0.007

   

PCBs

0.179

0.450

0.716

0.725

  

0.643

−0.353

0.398

−0.259

  
 

0.465

0.053

0.001

0.000

  

0.001

0.091

0.054

0.221

  

PCDDs

0.058

0.607

0.534

0.549

0.877

 

0.754

−0.320

0.556

0.045

0.576

 
 

0.815

0.006

0.019

0.015

0.000

 

0.000

0.127

0.005

0.833

0.003

 

PAHs

−0.030

0.286

0.636

0.539

0.665

0.536

0.689

−0.286

0.538

−0.030

0.769

0.513

 

0.903

0.236

0.003

0.017

0.002

0.018

0.000

0.175

0.007

0.891

0.000

0.010

Cells show Pearson correlation coefficient and the correspondent P value. In bold are reported significant correlations (r > 0.500)

3.6.2 Cluster analysis

Cluster analysis was applied on the raw data on the variables organic matter, Cd, Cr, Cu, Ni, Pb, Zn PAHs, PCBs, and PCDDs/DFs. Clusters were obtained using single linkage method and correlations for distance measure.

In Fig. 3 are shown dendrograms obtained from cluster analysis for agricultural and natural topsoils. In natural soils some clear clusters are identified, as organic matter is strongly associated with PCDDs/DFs and Pb and subsequently with PCBs and PAHs, confirming what previously discussed. Cr and Ni are strongly associated as well, being regrouped in a separate cluster regardless of the land use. Instead, differences due to the land use are observed in the association of the organic matter with the contaminants, which is very weak in agricultural soils. Some inorganic contaminants such as Cd and Cu do not seem to be associated with the other variables, showing no or a moderate degree of contamination. Results from the cluster analysis confirm the hypothesis made on the origin of contaminants, the influence of land use on their trend and on the association they have with the other contaminants and with the soil properties.
https://static-content.springer.com/image/art%3A10.1007%2Fs11368-009-0114-9/MediaObjects/11368_2009_114_Fig3_HTML.gif
Fig. 3

Cluster analysis applied to organic matter, heavy metals, and organic contaminants in natural and agricultural topsoils

3.7 Urban vs. rural

To give an appraisal of the degree of contamination in soils in Piedmont and on the effect of different land uses on the presence of contaminants, data from the present study were compared with those of a previous study on the urban soil system conducted within the city of Torino, the largest of the region (Biasioli and Ajmone-Marsan, 2007). Data from the two studies can be easily compared as they were carried out with the same sampling and analytical procedures. In Figs. 4 and 5 mean concentrations (±SD) of organic and inorganic contaminants obtained from the present study for natural and agricultural soils are compared with data from urban soils.
https://static-content.springer.com/image/art%3A10.1007%2Fs11368-009-0114-9/MediaObjects/11368_2009_114_Fig4_HTML.gif
Fig. 4

Mean concentrations (histograms) and standard deviations (error bars) of organic contaminants in urban (n=30), agricultural (n = 19), and natural (n = 24) topsoils. Dashed lines are Italian legislative limits

https://static-content.springer.com/image/art%3A10.1007%2Fs11368-009-0114-9/MediaObjects/11368_2009_114_Fig5_HTML.gif
Fig. 5

Mean concentrations (histograms) and standard deviations (error bars) of inorganic contaminants contents in urban (n=30), agricultural (n = 19), and natural (n = 24) topsoil. Dashed lines are Italian legislative limits. Cd concentrations are multiplied by 100

From the data presented in Fig. 4, clear differences appear in PAHs and PCBs average contents, these contaminants resulting strongly enriched in urban soils with respect to their surrounding natural and agricultural area. The urban environment plays therefore an important role in the emission of these pollutants and in concentrating them within its borders. Lower differences can be observed for PCDDs/DFs, indicating the diffuse drift of these contaminants over a large area and their peculiarity of being transported over long distances.

The city is also enriched in all inorganic contaminants analyzed (see Fig. 5), with average concentrations often above legislative limits, reflecting a strong emission of these contaminants into the urban atmosphere.

4 Conclusions

Even if natural soils in this study are mostly located far from the major sources of contamination (industries, traffic, agriculture) notable differences appeared when compared to agricultural areas. In particular, the enrichment in the concentrations of organic contaminants such as PCDDs in natural areas and of inorganics in agricultural soils confirms the presence of important phenomena of long-range diffuse contamination for the first and short-range for the latter, which appeared also to be related to agricultural activities. Contaminants like PCDDs, PCBs, PAHs, and Pb presented high vertical variability, confirming their anthropogenic origin and strong affinity to soil organic matter, while others appear to be more related to the natural background.

The comparison of data with those from soils within a large industrial city present in the study area confirms the strong enrichment of the urban environment with respect to some contaminants such as metals, PCBs, and PAHs. Other contaminants such as PCDD/DFs showed lower differences confirming their diffuse and ubiquitarious pollution.

Acknowledgements

The help of ARPA Piemonte, Polo Bonifiche of Alessandria, with the analytical work is gratefully acknowledged.

Copyright information

© Springer-Verlag 2009