Chemosphere 65 (2006) 863–872

www.elsevier.com/locate/chemosphere

Assessing heavy metal sources in agricultural soils of an

European Mediterranean area by multivariate analysis
C. Micó *, L. Recatalá, M. Peris, J. Sánchez
Centro de Investigaciones sobre Desertificación – CIDE (Universitat de València, CSIC, Generalitat Valenciana),
Camı́ de la Marjal s/n, Apartado Oficial, 46470 Albal, Valencia, Spain

Received 31 October 2005; received in revised form 9 March 2006; accepted 10 March 2006
Available online 25 April 2006

Abstract

According to the European Thematic Strategy for Soil Protection, the characterization of the content and source of heavy metals in soils
are necessary to establish quality standards on a regional level that allow the detection of sampling sites affected by pollution. In relation to
this, the surface horizons of 54 agricultural soils under vegetable crops in the Alicante province (Spain), a representative area of the Euro-
pean Mediterranean region, were sampled to determine the content of Cd, Co, Cr, Cu, Fe, Mn, Ni, Pb and Zn. Analytical determinations
were performed by atomic absorption spectroscopy after microwave sample digestion in acid solution. Results indicated that heavy metal
levels were similar to those reported by authors working on agricultural soils from other parts of the Mediterranean region, with the excep-
tion of Cu and Pb in some samples. Multivariate analysis (principal component analysis and cluster analysis) was performed to identify a
common source for heavy metals. Moreover, soil properties were determined in order to characterize agricultural soils and to analyze rela-
tionships between heavy metal contents and soil properties. The content of Co, Cr, Fe, Mn, Ni and Zn were associated with parent rocks
and corresponded to the first principal component called the lithogenic component. A significant correlation was found between lithogenic
metals and some soil properties such as soil organic matter, clay content, and carbonates, indicating an important interaction among them.
On the other hand, elements such as Cd, Cu and Pb were related to anthropic activities and comprised the second (Cu and Pb) and third
principal components (Cd), designated the anthropogenic components. Generally, Cd, Cu and Pb showed a lower correlation with soil
properties due to the fact that they remain in available forms in these agricultural soils. Taking into account these results and other
achieved in other parts of the European Mediterranean region, it can be concluded that soil quality standards are highly needed to declare
soils affected by human induced pollution. This is particularly relevant for anthropogenic metals (Cd, Cu and Pb, and in some areas also
Zn). Further research in other agricultural areas of the region would improve the basis for proposing such soil quality standards.
Ó 2006 Elsevier Ltd. All rights reserved.

Keywords: Agriculture; Soil pollution; Multivariate statistics; Alicante; European Mediterranean region

1. Introduction are especially dangerous because of their persistence and

Soil pollution has become an important environmental
issue in developed countries (e.g., The Netherlands, United
Kingdom, Spain) due to changes in the land use pattern
(urban-industrial expansion, infrastructure development,
and tourism increase) over the last few decades. When con-
sidering the different kinds of contaminants, heavy metals
*
Corresponding author. Tel.: +34 96 1220540; fax: +34 96 1270967.
E-mail address: carolina.mico@uv.es (C. Micó).
toxicity (Adriano, 2001). Metals can be transferred from
soil to the other ecosystem components, such as under-
ground water or crops, and can affect human health
through the water supply and food web. In some areas with
heterogeneous lithology, heavy metal contents can be highly
variable, determined by the parent material and soil proper-
ties. For example, organic matter, clay and carbonates play
a relevant role in the availability of heavy metals in calcar-
eous soils (Sánchez-Camazano et al., 1998).
The natural input of several heavy metals to soils due to
pedogenic processes has been exceeded in some local areas

0045-6535/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chemosphere.2006.03.016
864 C. Micó et al. / Chemosphere 65 (2006) 863–872
by human input, even on a regional scale. In particular,
agricultural soils can be a long-term sink for heavy metals.
Intensification of agriculture, encouraged in Europe by the
Common Agricultural Policy (CAP) from the fifties, has
resulted in the incorporation of several types of pollutants
to soil, such as heavy metals, due to excess use of agrochem-
icals. These soils have also been influenced by other pollu-
tant activities such as the use of manures, sewage sludge
disposal or aerial fallout from industrial activities (Van
Camp et al., 2004). As a consequence, potentially toxic ele-
ments have accumulated in the soil profile. This can result in
a loss of soil functions concerning environmental quality
protection, maintenance of human health and productivity,
which are relevant aspects of soil quality (e.g., Doran and
Parkin, 1994). Pollutant activities can have implications
for the quality of agricultural soils, including phytotoxicity
at high concentrations and the transfer of heavy metals to
the human diet from crop uptake or soil ingestion by graz-
ing livestock (Nicholson et al., 2003). However, heavy metal
studies in agricultural soils are mainly referred to Northern
European countries (e.g., Dudka, 1992; Zhang et al., 1999;
Reimann et al., 2001) having a lack of information in
Southern Europe, particularly at the Mediterranean region
(Stanners and Bourdeau, 1995), where only few studies have
been carried out at regional level in some countries such as
Greece (e.g., Stalikas et al., 1997) and Italy (e.g., Facchinelli
et al., 2001; Abollino et al., 2002). In the Spanish Mediter-
ranean region, most of the studies on heavy metals in agri-
cultural soils have concerned local areas (e.g., Boluda et al.,
1988; Andreu and Gimeno-Garcı́a, 1996; Pomares et al.,
1998). This paper contributes to the knowledge of the con-
tent and potential source of heavy metals in agricultural
soils of the Alicante province, which is a representative area
of the Spanish Mediterranean region.
The knowledge of the heavy metal content in soils and
the origin of these levels are priority objectives in the Euro-
pean Union (EU). At the start of 2002, the European Com-
mission published a report ‘‘Towards a Thematic Strategy
for Soil Protection’’ (EC, 2002) that established the basis
and guidelines for maintaining or even improving soil
resources quality. Recently, different working groups pro-
duced reports addressing the state of soils, impacts and
pressures as well as recommendations for soil protection
policy making at the EU level. For example, the report pre-
sented by Van Camp et al. (2004) as part of the technical
group working on ‘‘Contamination and Land Manage-
ment’’ established the need to measure heavy metal concen-
trations in soils to assess the contamination risks and to
identify background values as a basis to establish soil qual-
ity standards. Analysis of heavy metal concentrations in
soils is, therefore, critical for policy making orientated at
reducing heavy metal inputs to soil and guaranteeing the
maintenance or even improvement of soil functions.
To establish reliable policies, it is necessary to have a
good knowledge of the content and variability of heavy
metals in soils, but also of the contribution of both parent
rock and human activities. This is a difficult task in ancient
populated areas such as the European Mediterranean
region, where totally unpolluted soils are almost impossible
to find. Multivariate analysis (principal component analysis
– PCA and cluster analysis – CA) has been widely used to
assist the interpretation of environmental data (e.g., Tuncer
et al., 1993; Einax and Soldt, 1999) and to distinguish
between natural and anthropogenic inputs (e.g., Garcı́a
et al., 1996; Facchinelli et al., 2001; Lucho-Constantino
et al., 2005).
The present study was carried out as part of a survey on
heavy metal contents in agricultural soils under vegetable
crops on a regional scale supported by the Regional
Government (Valencian Ministry of Agriculture, Fisheries
and Food). The aims of the study were (1) to determine
contents of Cd, Co, Cr, Cu, Fe, Mn, Ni, Pb and Zn in soils;
(2) to define their natural or anthropic source using multi-
variate analysis; and (3) to establish relationships between
heavy metals and some soil properties.
2. Material and methods
2.1. Site characteristics and soil sampling
The Alicante province is situated in southeast Spain
(Fig. 1), within the European Mediterranean region. The
province has been traditionally associated with agriculture,
but international tourism and industrialized activities have
been on the increase for several decades. This zone is well
known for its special microclimate, characterized by warm
temperatures all year round, an annual average temperature
of 18 °C and an annual average rainfall of 350 mm. The
province covers 5863 km2 and approximately 116 km2 are
occupied by vegetable crops, which form an important part
of the diet in the Mediterranean region, including leaf crops
(e.g., lettuce, Swiss chard), inflorescences (e.g., artichoke,
broccoli) or tubers (e.g., potato, rape).
Soil samples were collected at 54 random agricultural
plots under vegetable crops (see Fig. 1) using a random
number generator, on the basis of a land use map at
1:300 000 scale. Composite samples for each plot consisted
of 16 cores of soil collected from the surface horizon (25 cm
depth topsoil). Cores were mixed into one composite sam-
ple for each soil and analyzed in triplicate. Soils were coded
from Case 1 to 54 at each georeferenced location. All the
samples were transported to the laboratory where they
were air dried for several days at ambient temperature
and sieved to 2 mm for a general analysis of soil properties.
A representative sample of each soil was finely ground in
an agate mortar and stored in closed bottles to determine
heavy metal contents.
2.2. Chemical analysis
Selected soil properties relevant to control the mobility
and bioavailability of heavy metals were analyzed for the
general characterization of agricultural soils according to
the official laboratory methods of the Spanish Ministry of
 C. Micó et al. / Chemosphere 65 (2006) 863–872
Fig. 1. Location of the study area and the agricultural plots.
Agriculture, Fisheries and Food (MAPA, 1994). Soil pH
was measured in a 1:2.5 soil:water suspension and electrical
conductivity was determined in the saturation paste extract.
Organic matter concentration was determined by the Walk-
ley–Black method and the total carbonate percentage by
the Bernard calcimeter. Cation exchange capacity was
determined by saturation with sodium acetate solution,
replacement of the adsorbed sodium with ammonium and
determination of displaced sodium by flame atomic absorp-
tion spectrometry. Particle-size distribution was measured
by the pipette method to determine the sand, silt and clay
percentages.
Heavy metal contents were analyzed by microwave
extraction according to method 3051A of the United States
Environmental Protection Agency (USEPA, 1998), which
consists in a soil sample digestion with nitric acid (HNO3)
and hydrochloric acid (HCl) in a ratio of 3:1 (HNO3:HCl).
Method 3051A is extensively used elsewhere (e.g., Wei et al.,
1997; Chen and Ma, 1998; Link et al., 1998) and was
selected because it provided satisfactory results in relation
to recovery and precision data. Heavy metal contents in
the final solutions were determined by flame (air-acetylene)
atomic absorption spectroscopy (FAAS) with the exception
of some samples for Cd, which were determined by graphite
furnace atomic absorption spectroscopy (GFAAS).
2.3. Statistical analysis
Multivariate analysis was performed using SPSSÒ 12.0
for Windows. PCA was interpreted in accordance with
the hypothetical source of heavy metals (lithogenic, anthro-
pogenic or mixed). Varimax rotation was applied because
865
orthogonal rotation minimizes the number of variables
with a high loading on each component and facilitates
the interpretation of results. Correlation matrix (CM)
was used to identify the relationship between heavy metal
contents and soil properties. Pearson’s product moment
correlation coefficient was used for normal populations
whereas Spearman’s rank correlation coefficient was used
for non-normal populations. CA was applied to identify
different geochemical groups, clustering the samples with
a similar heavy metal content. CA was undertaken accord-
ing to the Ward-algorithmic method. Results are shown in
a dendrogram where steps in the hierarchical clustering
solution and values of the distances between clusters
(squared Euclidean distance) are represented.
3. Results and discussion
3.1. Descriptive statistics
Soil properties analyzed in the study area (Table 1) were
consistent with other agricultural soils characterized in the
Mediterranean region by some scientists (e.g., Millán et al.,
1983; Boluda et al., 1988; Andreu and Gimeno-Garcı́a,
1996; Stalikas et al., 1997; Pomares et al., 1998; Abollino
et al., 2002). The K–S test confirmed that soil properties
are normally distributed, with the exception of pH, electri-
cal conductivity and sand content.
The levels of organic matter, high percentages of clay
and the presence of carbonate seem to suggest an impor-
tant retention of heavy metals by these components.
According to these results, the bioavailability of trace ele-
ments would be expected to be low in the soils analyzed,
866 C. Micó et al. / Chemosphere 65 (2006) 863–872

Table 1
Descriptive statistics of soil properties and heavy metal contents
Variable Symbol Unit Descriptive statistics
Mean SD Min. Max. Skewness K–S test
pH pH – 8.3 0.2 7.8 9.1 0.977 0.037
Electrical conductivity EC dS m 1 1.4 1.0 0.1 4.9 1.233 0.010
Organic matter OM % 2.1 0.5 0.8 3.3 0.048 0.200
Total carbonate TC % 46.5 6.9 23.6 67.5 0.049 0.051
1
Cation exchange capacity CEC cmol(+) kg 16.3 4.2 8.1 25.1 0.005 0.200
Sand content Sand % 21 16 2 65 0.953 0.003
Silt content Silt % 45 9 14 63 0.905 0.060
Clay content Clay % 34 12 8 56 0.087 0.200
Cadmium Cd mg kg 1 0.34 0.20 0.10 0.88 0.964 0.000
Cobalt Co mg kg 1 7.1 1.7 2.9 10.1 0.260 0.200
Chromium Cr mg kg 1 26.5 5.9 14.8 42.5 0.533 0.200
Copper Cu mg kg 1 22.5 8.9 12.6 76.8 4.427 0.000
Iron Fe mg kg 1 13 608 3107 6588 19 807 0.241 0.200
Manganese Mn mg kg 1 295 61 133 406 0.865 0.079
Nickel Ni mg kg 1 20.9 5.1 11.4 32.0 0.089 0.200
Lead Pb mg kg 1 22.8 16.1 6.8 101.3 3.247 0.000
Zinc Zn mg kg 1 52.8 14.9 20.3 88.8 0.158 0.200
SD – standard deviation; Min. – minimum value; Max. – maximum value; K–S test – Kolmogorov–Smirnov test.

although anthropogenic sources (Alloway, 1990; Adriano,
2001) and some soil properties such as salinity (Doner,
1978) could facilitate the mobility of some trace elements
(e.g., Cd, Cu). In fact, most of the soils are saline (87%
of the samples) with an electrical conductivity in the satu-
ration paste extract above 2 dS m 1. This issue has been
tackled in other studies on soils (e.g., Ortı́z et al., 1985)
and on waters (e.g., Andreu et al., 2002) of the Alicante
province and was attributed to the use of poor quality
water for irrigation from underground water resources,
especially in the south of the province (Segura river basin).
The mean values of the heavy metal contents (Table 1)
arranged in the following decreasing order: Fe > Mn >
Zn > Cr > Pb > Cu > Ni > Co > Cd. Therefore, micronu-
trients such as Fe, Mn and Zn presented higher levels in
agricultural soils, whereas Co and Cd presented the lowest
values. The application of the K–S test confirmed that the
heavy metal contents in these soils are normally distributed
with the exception of Cd, Cu and Pb. In the case of Co, Cr,
Fe, Mn, Ni and Zn, the raw data sets conformed to the
background population, whereas Cu and Pb presented
anomalous values that were identified as outliers. The pop-
ulations of Cu and Pb fitted a lognormal distribution for
the raw data set and a normal distribution when outliers
were not considered as part of the population data (back-
ground population). This may be due to the presence of
disturbed soils from a point pollution source as explained
below. On the other hand, Cd showed a non-normal distri-
bution for its population data due to the determination of
some samples by GFAAS as a consequence of its values
being below the detection limit by FAAS. This produces
a bias of the population in the low value zone and deter-
mines a non-normal population.
The heavy metal levels in all samples do not exceed ref-
erence values established by the 86/278/EEC Directive for
agricultural soils with a pH > 7. However, there are some
samples with a metal level (particularly, Cu and Pb) higher
than other works on Spanish agricultural soils in similar
conditions (Table 2). In fact, the maximum values of Cu
(76.8 mg kg 1) and Pb (101 mg kg 1) were above the values

Table 2
1
Heavy metal contents in Spanish agricultural soils (in mg kg dry weight)
Metal In this Millán et al. Boluda et al. Andreu and Pomares et al.
work (1983)a (1988)b Gimeno-Garcı́a (1996)c (1998)d
Cd 0.3 – – 0.6 –
Co 7 – 16 7 –
Cr 27 – 49 – 33
Cu 23 39 – 26 31
Fe 13 608 20 787 – – 25 930
Mn 295 382 – – 261
Ni 21 – 37 16 12
Pb 23 – 42 47 22
Zn 53 57 62 81 59
a
Ebro River Valley.
b
‘‘La Plana de Requena-Utiel’’ area in the Valencia province.
c
Albufera Natural Park in the Valencia province.
d
Castellón province.
 C. Micó et al. / Chemosphere 65 (2006) 863–872 867

found in these studies and higher than background values
enforced in European countries (Kabata-Pendias and Pen-
dias, 2001). Therefore, levels of heavy metal analyzed cor-
responded to background populations, whereas Cu and Pb
presented moderately polluted samples. The most impor-
tant activity, which produces the highest concentrations
of Cu in agricultural soils, is the use of fungicidal sprays
that reach the soil surface directly or indirectly as leaf litter
(Tiller and Merry, 1982). Copper sulphate is also used in
the study area as an aquatic algaecide in irrigation ditches
and can reach soils or vegetable crops (Rico et al., 2003).
On the other hand, some authors (e.g., Alloway, 1990; Til-
ler, 1992; Turer et al., 2001) suggested that atmospheric
deposition of Pb could occur in soils near roads with con-
siderable traffic, from solid particles and toxic fumes in the
atmosphere. However, other authors (e.g., Aucejo et al.,
1997; Facchinelli et al., 2001) suggested that wastewater
was the main source for these elements in agricultural soils,
especially in densely populated areas with an inefficient
cleaning process. In the Alicante province, both of these
sources could contribute to increasing Pb levels. Finally,
Cd contents analyzed not exceed recommended levels
although a slight enrichment of the top layer had occurred
in some parts of the study area as identified in some sam-
ples from the banks of the Segura river.
In summary, the soils studied can be considered as
unpolluted as their heavy metal contents are within typical
ranges recorded in other similar studies, with the exception
of Cu and Pb. However, because the total heavy metal con-
tent is not a good predictor of the availability and mobility
of heavy metals in soils (Kabata-Pendias and Pendias,
2001), further research is required to assess the pool of
potentially available species (e.g., extractable fraction) to
determine the probability of these transferring from the soil
to other ecosystem components (e.g., crops, underground
water).
3.2. Multivariate analysis approaches
3.2.1. Principal component analysis (PCA)
The results of PCA for heavy metal contents are pre-
sented in Table 3. According to these results, the eigen-
values of the third extracted components are greater than
one both before and after the matrix rotation. As a conse-
quence, heavy metals could be grouped into a three-com-
ponent model that accounts for 82% of all the data
variation. Spatial representation of the three rotated com-
ponents is shown in Fig. 2.
PC1 can be considered to be a lithogenic component, as
the variability of the elements seems to be controlled by
parent rocks. The parent rock factor explains 56% of the
total variance and is the most important component. This
result suggests that the distribution of Co, Cr, Fe, Mn,
Ni and, partially, Zn has a lithogenic control, included in
the first principal component. In fact, a non-point source
contamination for these elements was detected in the study
area. A similar chemical behaviour is presented by Fe and
Mn, partly in oxide form and partly present as hydroxides,
which appear in precipitated forms in sedimentary rocks
under Mediterranean conditions. The parent material

Table 3
Total variance explained and component matrixes (three principal components selected) for heavy metal contents
Component Initial eigenvalues Extraction sums of squared loadings Rotation sums of squared loadings
Total % of variance Cumulative (%) Total % of variance Cumulative (%) Total % of variance Cumulative (%)
Total variance explained
1 5.035 55.946 55.946 5.035 55.946 55.946 4.904 54.492 54.492
2 1.289 14.323 70.269 1.289 14.323 70.269 1.310 14.559 69.050
3 1.078 11.980 82.249 1.078 11.980 82.249 1.188 13.199 82.249
4 0.762 8.470 90.719
5 0.317 3.523 94.242
6 0.249 2.772 97.014
7 0.125 1.389 98.403
8 0.080 0.892 99.295
9 0.063 0.705 100.000
Element Component matrix Rotated component matrix
PC1 PC2 PC3 PC1 PC2 PC3
Component matrixes
Ni 0.932 0.249 0.003 0.962 0.076 0.046
Fe 0.947 0.070 0.064 0.942 0.044 0.133
Co 0.846 0.367 0.188 0.904 0.093 0.244
Mn 0.928 0.097 0.044 0.895 0.238 0.122
Cr 0.874 0.009 0.145 0.853 0.059 0.233
Zn 0.836 0.317 0.020 0.762 0.378 0.276
Pb 0.024 0.798 0.298 0.162 0.825 0.142
Cu 0.408 0.373 0.520 0.347 0.641 0.209
Cd 0.251 0.448 0.810 0.140 0.002 0.949
Extraction method: Principal Component Analysis; Rotation method: Varimax with Kaiser normalization.
868 C. Micó et al. / Chemosphere 65 (2006) 863–872

agronomic practices, whereas vehicle and industrial fumes

Fig. 2. Principal component analysis loading plots for the three rotated
components.
present in alluvial and alluvial–colluvial areas, principally
of calcareous nature, determines Co, Cr and Ni contents.
Zn can have a lithogenic source as it forms a number of sol-
uble salts (e.g., chlorides, sulphates and nitrates) or insolu-
ble salts (e.g., silicates, carbonates, phosphates, oxides and
sulphides) according to the prevailing pedogenic processes
(Adriano, 2001). However, Zn and its compounds are also
used in different manufactured goods (e.g., paints, cosmet-
ics, automobile tyres, batteries, and electrical apparatus)
and in agricultural fertilizers. This element has the follow-
ing loading factor in the second and third principal compo-
nents, indicating a mixed source both from lithogenic and
from anthropogenic inputs. On the other hand, relatively
lower loading factors of Cd, Cu and Pb in the lithogenic
component (PC1) suggested other effects and also other
influences (e.g., agricultural practices or industrial conta-
mination) on the status of these elements in the soils.
PC2 includes Cu and Pb and can be defined as an
anthropogenic component due to the presence of high lev-
els in some soils. As explained above, high Cu values can
come from Cu-based agrochemicals related to specific
and wastewater used to irrigate some agricultural areas
can also be the source for the high Pb values found in some
soils. In fact, some plots that contain high Pb contents are
located near roads that have a high traffic density and are,
probably, affected by vehicle fumes. However, other plots
are far from roads but close to some important villages
and have probably been influenced by wastewater.
PC3 only includes Cd content. As we mentioned previ-
ously, enrichment of the top layer occurs in some areas
although the recommended level is not exceeded. Phos-
phate fertilizers are an important source of heavy metals
entering agricultural soils, especially Cd, Cu and Zn. Con-
cern about the accumulation of Cd in the environment
have resulted in some fertilizer manufacturers, encouraged
by the European Union Commission, changing the source
of raw materials to reduce inputs (Nicholson et al., 2003).
Other sources of Cd may include other inorganic fertilizers
(e.g., nitrogen or potash), atmospheric deposition or
anthropic wastes such as sewage sludge, wastewater or
waste materials. Therefore, PC3 can also be an anthropo-
genic component due to the isolation of Cd to lithogenic
elements in the graph (see Fig. 2).
Facchinelli et al. (2001) also drew conclusions about two
groups of heavy metals. For these authors, Cu and Zn were
associated with specific agronomic practices and Pb was
derived from car exhausts, and all three metals were related
with human activities. On the other hand, Cr, Co and Ni
were associated with and controlled by parent rocks (Cd,
Fe, and Mn were not analyzed by these authors). In our
study, it seems reasonable to conclude that Cd, Cu and
Pb constitute an anthropogenic component related to spe-
cific human activities, whereas the remaining elements (Co,
Cr, Fe, Mn and Ni) appear to be associated with parent
rocks. In the case of Zn, this element displays a combined
relationship with both groups and seems to have both nat-
ural and anthropic origin.
3.2.2. Correlation matrix (CM)
Organic matter (OM) showed a positive linear relation-
ship with all the elements analyzed despite the fact

Table 4
Correlation matrix between heavy metal contents and soil properties in the studied soils
pH EC OM TC CEC Sand Silt Clay
Cd 0.054a ns 0.030a ns 0.437**,a 0.229a ns 0.272*,a 0.246a ns 0.288*,a 0.138a ns
Co 0.388**,a 0.186a ns 0.508** 0.423** 0.617** 0.812**,a 0.393** 0.760**
Cr 0.162a ns 0.104a ns 0.469** 0.130 ns 0.549** 0.695**,a 0.229 ns 0.752**
Cu 0.008a ns 0.349**,a 0.434**,a 0.243a ns 0.388**,a 0.349**,a 0.064a ns 0.441**,a
Fe 0.337*,a 0.043a ns 0.535** 0.460** 0.619** 0.842**,a 0.369** 0.805**
Mn 0.246a ns 0.025a ns 0.559** 0.399** 0.652** 0.720**,a 0.397** 0.688**
Ni 0.324*,a 0.097a ns 0.514** 0.302* 0.746** 0.912**,a 0.430* 0.850**
Pb 0.047a ns 0.067a ns 0.270*,a 0.069a ns 0.050a ns 0.013a ns 0.199a ns 0.033a ns
Zn 0.099a ns 0.212a ns 0.614** 0.226 ns 0.585** 0.639**,a 0.295* 0.578**
ns – not significant; EC – electrical conductivity; OM – organic matter; TC – total carbonate, CEC – cation exchange capacity.
*
P < 0.05.
**
P < 0.01.
a
Spearman’s coefficient.
 C. Micó et al. / Chemosphere 65 (2006) 863–872
that organic matter content was low in the study area
(Table 4). The clay content exhibited a significant relation-
ship with all the elements, with the exception of Cd and Pb.
The third soil property with a large number of correlations
with heavy metals was total carbonate (TC), which showed
a correlation with Co, Fe, Mn and Ni. These results sug-
gest that adsorption and retention of heavy metals in these
agricultural soils are influenced by several factors, such as
organic matter, clay minerals and carbonates. On the other
hand, soil properties with a low relationship with heavy
metal content were pH and electrical conductivity. For
the former, significant negative correlations were only
observed with Co, Fe and Ni. This could be due to the nar-
row interval of pH values in the study area confirmed by
the low standard deviation of this parameter. Electrical
conductivity (EC) only showed a linear and negative rela-
tionship with Cu content. Doner (1978) found an increased
mobility of Cu(II) through the soil profile as a consequence
of chlorocomplex formation. In addition, the substitution
of sodium in the exchange positions can produce desorp-
tion of some elements such as Cd or Cu and a higher
mobility of these elements. This result may suggest a
higher bioavailability of Cu in some soils with salinity
problems.
Results obtained by PCA were confirmed by analyzing
the CM. Lithogenic metals such as Co, Fe, Mn or Ni
showed a higher correlation with soil properties. This con-
clusion is in agreement with other studies (Ma et al., 1997;
Chen et al., 1999) and suggests that the parent material and
subsequent pedogenic processes are major factors in the
amounts and distribution of lithogenic metals in these soils.
On the other hand, anthropogenic metals (Cd and Pb)
showed a weaker relationship with soil properties as a con-
sequence of their external sources. Only Pb showed a cor-
relation with organic matter, showing that this soil
component is an important sink for Pb in the soils studied.
These results have also been reported by other authors in
calcareous soils (e.g., Sánchez-Camazano et al., 1998). In
the case of Cu, this element was correlated with a large
number of soil properties, but high levels in some samples
also indicated an anthropogenic source. In addition, heavy
metals from an anthropogenic source could be more mobile
than lithogenic metals according to Burt et al. (2003). This
could constitute a threat to human health through the food
web and should be investigated in the agricultural soils of
the study area. This would require, for example, analysis
of the extractable contents in soils or heavy metal contents
in vegetable crops, especially where high levels were found
in soils.
3.2.3. Cluster analysis (CA)
The dendrogram obtained (Fig. 3) was used to define
four geochemical groups (Group 1–4) that illustrate the
complexity of the study area. Group 1 only includes a soil
sample (Case 38) affected by high Cu content (76.8 mg kg 1)
due to the application of fungicidal sprays or algaecides.
This agricultural plot is located in Villena (northwest of
869
the province) where an intensification of agricultural prac-
tices has occurred over the last few decades. Group 2
includes three soil samples (Case 3, 35 and 51) having high
contents of Pb. Case 3 is also located in Villena, whereas
Cases 35 and 51 are located in Elche (southeast of the prov-
ince). These three soils are located close to local roads with
dense traffic. In both cases (Group 1 and 2), further studies
are required to evaluate a possible risk for human health or
the environment due to high soil levels of Cu and Pb.
Soils in Group 3 are located in the Segura Valley (south
of the province), which includes most of the cases (72%).
This group includes soils with heavy metal contents within
normal ranges, although a few plots can be identified with
an enrichment of some elements (e.g., Cd, Cr or Zn). For
this group, it would desirable to monitor these levels in
order to avoid a continued increase of heavy metal contents
as a consequence of human activities. Finally, Group 4
includes plots with a low content for most of the elements
and accounts for 20% of the total. For these cases (Group 3
and 4), normal levels of heavy metals were identified and
include most of the soils analyzed (92%). Therefore, an
acceptable quality of agricultural soils was detected in the
study area according to the heavy metal contents analyzed
and background values may be fixed on the basis of levels
recorded in plots included within Group 3 and 4.
The work presented here is consistent with those per-
formed by other scientists (e.g., Stalikas et al., 1997; Fac-
chinelli et al., 2001; Abollino et al., 2002) in other parts
of the European Mediterranean region. Within this region,
it seems clear that Co, Cr, Fe, Mn and Ni have a lithogenic
origin whereas Cu and Pb have an anthropogenic origin in
the agricultural soils. For other metals such as Cd or Zn,
the origin varies between different areas of the region
depending on specific human activities that are locally rel-
evant. For instance, the Cd content identified in the agri-
cultural soils of Alicante province seems to come from
irrigation water polluted by industries. The Zn contents
identified in the agricultural soils of the Piedmont region
in Italy come from a long-term anthropic activity con-
nected with grape-growing (Facchinelli et al., 2001).
The results achieved in this paper contribute to the
knowledge of the content and the origin of heavy metals
in agricultural soils within the European Mediterranean
region, which is relevant for soil quality policy making at
regional level. For instance, the concentrations of Co, Cr,
Fe, Mn, Ni and Zn found in Alicante correspond to back-
ground values, similarly to other parts of the region. By
applying several approaches such as descriptive statistics
(Esser, 1996), probability plots (Davies, 1983; Tobı́as
et al., 1997) or regression models (Tack et al., 1997) to
background values found in the region it can be derived
critical (or trigger) values to allow distinguish between pol-
luted and non-polluted soils. This is highly needed for Cd,
Zn, Cu and Pb, particularly for the two latter, which have
an anthropogenic origin in the agricultural soils of the
region. However, further information on heavy metal con-
tents in soils from other agricultural areas of the European
870 C. Micó et al. / Chemosphere 65 (2006) 863–872

Fig. 3. Dendrogram obtained by cluster analysis for heavy metal contents in soil samples.

Mediterranean region would be needed to fix agricultural spective of the recent Thematic Strategy for Soil Protection
soil quality standards at regional level according to the per- of the European Union.
 C. Micó et al. / Chemosphere 65 (2006) 863–872
4. Conclusions
The results obtained in this work increase our knowl-
edge of the heavy metal contents and their possible source
in the agricultural soils of the province of Alicante, a rep-
resentative area of the European Mediterranean region.
The PCA performed on nine heavy metals identified three
principal components controlling their variability in agri-
cultural soils under vegetable crops. Co, Cr, Fe, Mn, Ni
and, partially, Zn content were associated in the same com-
ponent (PC1), determined principally by soil parent rocks.
In fact, a point source of contamination for these elements
was not identified in the study area. The Cu and Pb (PC2)
content were related to the anthropogenic component due
to the presence of high levels of these elements in a large
number of samples. Finally, the Cd content (PC3) was con-
trolled by a long-term anthropic activity as a consequence
of inorganic fertilizers, atmospheric deposition or indus-
trial wastes. Therefore, this element constituted an anthro-
pogenic component. These results could be confirmed by
the fact that lithogenic metals were better correlated with
soil properties indicating an important interaction with
some components such as organic matter, carbonates or
clay. On the other hand, Cd and Pb were less correlated
with soil properties, possibly due to it remaining in avail-
able forms in these agricultural soils. In the case of Cu, this
element was significantly correlated with soil properties,
but the high levels analyzed in some samples and its inclu-
sion into the second component with Pb also indicated an
anthropogenic source. These results are consistent with
those achieved by other scientists in other parts of the
European Mediterranean region, highlighting the need of
establishing soil quality standards for heavy metals in order
to declare soils affected by human induced pollution. This
is particularly relevant for anthropogenic metals (Cd, Cu
and Pb, and in some areas also Zn). Further knowledge
from other areas would extend and improve the basis for
proposing agricultural soil quality standards at regional
level according to the European Strategy for Soil
Protection.
Acknowledgements
The authors wish to thank the Valencian Ministry of
Agriculture, Fish and Food (Generalitat Valenciana) for
financial support (contract GV-CAPA00-21). The con-
structive and thorough reviews of the referees and the Edi-
tor of the journal are warmly acknowledged.
References
Abollino, O., Aceto, M., Malandrino, M., Mentasti, E., Sarzanini, C.,
Petrella, F., 2002. Heavy metals in agricultural soils from Piedmont,
Italy. Distribution, speciation and chemometric data treatment.
Chemosphere 49, 545–557.
Adriano, D.C., 2001. Trace Elements in Terrestrial Environments.
Biogeochemistry, Bioavailability and Risks of Metals. Springer-
Verlag, New York.
871
Alloway, B.J., 1990. Heavy Metals in Soils. Blackie & Son, Ltd., London.
Andreu, V., Gimeno-Garcı́a, E., 1996. Total content and extractable
fraction of cadmium, cobalt, copper, nickel, lead, and zinc in
calcareous orchard soils. Commun. Soil Sci. Plant 27, 2633–2648.
Andreu, J.M., Garcı́a-Sánchez, E., Cerón, J.C., Pulido-Bosch, A., Jordán-
Vidal, M.M., Navarro, J., Mataix-Solera, J., 2002. Algunas Conside-
raciones sobre la Calidad de las Aguas del Acuı́fero del Tarafa
(Alicante, España). In: Bocanegra, E., Martı́nez, D., Massone, H.
(Eds.), Groundwater and Human Development. Mar de la Plata,
Argentina, pp. 129–133.
Aucejo, A., Ferrer, J., Gabaldón, C., Marzal, P., Seco, A., 1997. Diagnosis
of boron, fluorine, lead, nickel and zinc toxicity in citrus plantations in
Villareal, Spain. Water Air Soil Poll. 94, 349–360.
Boluda, R., Andreu, V., Pons, V., Sánchez, J., 1988. Contenido de metales
pesados (Cd, Co, Cr, Cu, Ni, Pb y Zn) en suelos de la comarca La
Plana de Requena-Utiel (Valencia). Anal. Edafol. Agrob. 47, 1485–
1502.
Burt, R., Wilson, M.A., Mays, M.D., Lee, C.W., 2003. Major and trace
elements of selected pedons in the USA. J. Environ. Qual. 32, 2109–
2121.
Chen, M., Ma, L.Q., 1998. Comparison of four USEPA digestion methods
for trace metal analysis using certified and Florida soils. J. Environ.
Qual. 27, 1294–1300.
Chen, M., Ma, L.Q., Harris, W.G., 1999. Baseline concentrations of 15
trace elements in Florida surface soils. J. Environ. Qual. 28, 1173–
1181.
Davies, B.E., 1983. A graphical estimation of the normal lead content of
some British soils. Geoderma 29, 67–75.
Doner, H.E., 1978. Chloride as a factor in mobilities of Ni(II), Cu(II), and
Cd(II) in soil. Soil Sci. Soc. Am. J. 42, 882–885.
Doran, J.W., Parkin, T.B., 1994. Defining and assessing soil quality. In:
Doran, J.W., Coleman, D.C., Bezdicek, D.F., Stewart, B.A. (Eds.),
Defining Soil Quality for a Sustainable Environment, SSSA Special
Publication no. 35. Soil Science Society of America, Madison, WI,
pp. 3–21.
Dudka, S., 1992. Factor analysis of total element concentrations in surface
soils of Poland. Sci. Total Environ. 121, 39–52.
EC, 2002. Communication of 16 April 2002 from the Commission to the
Council, the European Parliament, the Economic and Social Com-
mittee and the Committee of the Regions: Towards a Thematic
Strategy for Soil Protection. European Commission (EC), Brussels.
Einax, J.W., Soldt, U., 1999. Geostatistical and multivariate statistical
methods for the assessment of polluted soils: merits and limitations.
Chemometr. Intell. Lab. 46, 79–91.
Esser, K.B., 1996. Reference concentrations for heavy metals in mineral
soils, oat and orchard grass (Dactylis glomerata) from three agricul-
tural regions in Norway. Water Air Soil Poll. 89, 375–397.
Facchinelli, A., Sacchi, E., Mallen, L., 2001. Multivariate statistical and
GIS-based approach to identify heavy metal sources in soils. Environ.
Pollut. 114, 313–324.
Garcı́a, R., Maiz, I., Millán, E., 1996. Heavy metal contamination analysis
of roadsoils and grasses from Gipuzkoa (Spain). Environ. Technol. 17,
763–770.
Kabata-Pendias, A., Pendias, H., 2001. Trace Elements in Soils and
Plants. CRC Press, Boca Raton, FL.
Link, D.D., Walter, P.J., Kingston, H.M., 1998. Development and
validation of the new EPA microwave-assisted leach Methods 3051A.
Environ. Sci. Technol. 32, 3628–3632.
Lucho-Constantino, C.A., Álvarez-Suárez, M., Beltrán-Hernández, R.I.,
Prieto-Garcı́a, F., Poggi-Varaldo, H.M., 2005. A multivariate analysis
of the accumulation and fractionation of major and trace elements in
agricultural soils in Hidalgo State, Mexico irrigated with raw waste-
water. Environ. Int. 31, 313–323.
Ma, L.Q., Tan, F., Harris, W.G., 1997. Concentrations and distributions
of eleven metals in Florida soils. J. Environ. Qual. 26, 769–775.
MAPA, 1994. Métodos Oficiales de Análisis. Tomo III. Servicio de
Publicaciones del Ministerio de Agricultura, Pesca y Alimentación,
Madrid.
872 C. Micó et al. / Chemosphere 65 (2006) 863–872
Millán, E., Abadia, A., Montañés, L., 1983. Niveles de Fe, Mn, Cu, y Zn
en suelos cultivados del valle del Ebro. An. Aula Dei. 16, 305–317.
Nicholson, F.A., Smith, S.R., Alloway, B.J., Carlton-Smith, C., Cham-
bers, B.J., 2003. An inventory of heavy metals inputs to agricultural
soils in England and Wales. Sci. Total Environ. 311, 205–219.
Ortı́z, R., Hernández, S., Alcaraz, F., 1985. Caracterización edafogenética
de los suelos halomorfos de un sector meridional de la Albufera de
Elche (Alicante). Anal. Edaf. Agrob. 44, 93–117.
Pomares, F., Estela, M., Tarazona, F., Sala, M.O., Canet, R., 1998.
Estado de contaminación por metales pesados en suelos de cı́tricos con
cultivo ecológico. In: SEAE (Ed.), Actas del III Congreso de la
Sociedad Española de Agricultura Ecológica: Una Alternativa para el
Mundo Rural del Tercer Milenio. Sociedad Española de Agricultura
Ecológica (SEAE), Valencia, pp. 247–252.
Reimann, C., Koller, F., Frengstad, B., Kashulina, G., Niskavaara, H.,
Englmaier, P., 2001. Comparison of the element composition in several
plant species and their substrate from a 1,500,000 km2 area in
Northern Europe. Sci. Total Environ. 278, 87–112.
Rico, A., Olcina, J., Paños, V., 2003. Rasgos geográficos de las tierras del
Alto Vinalopó. In: Bello, A., López-Pérez, J.A., Garcı́a Álvarez, A.
(Eds.), Biofumigación en Agricultura Extensiva de Regadı́o. Pro-
ducción Integrada de Hortı́colas. Mundi-Prensa S.A. y Fundación
Rural Caja Alicante, Alicante, pp. 39–88.
Sánchez-Camazano, M., Sánchez-Martin, M.J., Lorenzo, L.F., 1998.
Significance of soil properties for content and distribution of cadmium
and lead in natural calcareous soils. Sci. Total Environ. 218, 217–226.
Stalikas, C.D., Mantalovas, A.Ch., Pilidis, G.A., 1997. Multielement
concentrations in vegetable species grown in two typical agricultural
areas of Greece. Sci. Total Environ. 206, 17–24.
Stanners, D., Bourdeau, P., 1995. Europe’s Environment. The Dobris
Assessment. European Environment Agency (EEA), Copenhagen.
Tack, F.M.G., Verloo, M.G., Vanmechelen, L., Van Ranst, E., 1997.
Baseline concentration levels of trace elements as a function of clay
and organic carbon contents in soils in Flanders (Belgium). Sci. Total
Environ. 201, 113–123.
Tiller, K.G., 1992. Urban soil contamination in Australia. Austr. J. Soil
Res. 30, 937–957.
Tiller, K.G., Merry, R.H., 1982. Copper pollution of agricultural soils. In:
Loneragan, J.F., Robson, A.D., Graham, R.D. (Eds.), Copper in Soils
and Plants. Academic Press, London, pp. 119–140.
Tobı́as, F.J., Bech, J., Sánchez, P., 1997. Establishment of the background
levels of some trace elements in soils of NE Spain with probability
plots. Sci. Total Environ. 206, 255–265.
Tuncer, G.T., Tuncel, S.G., Tuncel, G., Balkas, T.I., 1993. Metal
pollution in the Golden Horn, Turkey – contribution of natural and
anthropogenic components since 1913. Water Sci. Technol. 28, 59–
64.
Turer, D., Maynard, J.B., Sansalone, J.J., 2001. Heavy metal contami-
nation in soils of urban highways: comparison between runoff and
soil concentrations at Cincinnati, Ohio. Water Air Soil Poll. 132, 293–
314.
USEPA, 1998. Method 3051A. Microwave assisted acid digestion of
sediments, sludges, soils and oils. Revision 1, US Environmental
Protection Agency (USEPA), Washington, DC.
Van Camp, L., Bujarrabal, B., Gentile, A.-R., Jones, R.J.A., Montana-
rella, L., Olazabal, C., Selvaradjou, S.-K., 2004. Reports of the
Technical Working Groups Established Under the Thematic Strategy
for Soil Protection, vol. IV. Contamination and Land Management.
EUR 21319 EN/4, European Communities, Luxembourg.
Wei, Y.-L., Shyu, H.-M., Joehuang, K.-L., 1997. Comparison of
microwave vs. hot-plate digestion for nine real-world river sediments.
J. Environ. Qual. 26, 764–768.
Zhang, Ch., Selinus, O., Kjellström, G., 1999. Discrimination between
natural background and anthropogenic pollution in environmental
geochemistry – exemplified in an area of south-eastern Sweden. Sci.
Total Environ. 243/244, 129–140.