respectively for Co, V and Mn (Xu et al., 2008). E r
i is called the
monomial potential ecological risk factor, and it corresponds to the
potential ecological risk of a given contaminant.
The local background (see Section 6.1) is selected as the reference
baselines in this study. The calculated RI values can be categorized into
four classes of potential ecological risks: low risk (< 150), moderate
(150–300), considerable (300–600) and very high (> 600).
4.3. Sediment quality guidelines method (SQGs)
Various sediment quality guidelines are used to protect aquatic
biota from the harmful and toxic effects related with sediment contaminants (McCready et al., 2006). These guidelines evaluate the degree to which the sediment-associated chemical status might adversely
affect aquatic organisms and are designed for the interpretation of sediment quality. They can be used to rank and prioritize both sites and
chemicals of potential concern (Dias De Alba et al., 2011).
4.3.1. Effects range-low and effects range-medium values
The “Effects range low” (ERL) and “Effects range medium” (ERM)
are sediment quality guidelines developed by Long and Morgan (1990)
to relate metal concentrations with their effects. These concentrations
have been separated into three categories: scarcely observed or predicted effect (below ERL), occasionally observed effect (ERL – ERM)
and frequently observed effect (above ERM) (Long et al., 1995a,
1995b). The mean ERM quotient (m-ERM-Q) method is applied, to
determine the possible biological effect of a group of heavy metals, by
calculating a mean quotient (Long and Morgan, 1990; Carr et al., 1996;
Long et al., 2000) using the following equation:
∑
−
− =
m ERM Q
[C /(ERM )]/n
i
i
(6)
where Ci is the concentration of metal i in the sediment, ERM i the
guideline values for the metal and n is the number of metals. The
toxicity probabilities of m-ERM quotient of < 0.1, 0.11–0.5, 0.51–1.5,
and > 1.50 are 9%, 21%, 49%, and 76%, respectively (Long et al.,
2000).
4.3.2. Threshold effect level (TEL) and probable effect level (PEL) approach
The TEL and PEL levels (also referred as TEC and PEC) are other
sediment quality guidelines widely used to assess the ecotoxicology of
sediments. This approach is also based on the relation between measured concentrations of metals and observed biological effects, such as
mortality, growth or reproduction of living organisms. The Threshold
effect level (TEL) refers to the concentration below which adverse effects are expected to occur only rarely, whereas Probable effect level
(PEL) indicates the concentration above which adverse effects are expected to occur frequently (Long and MacDonald, 1998 and Mac
Donald et al., 2000).
The mean PEL quotient (m-PEL-Q) was calculated using:
∑
−
− =
m PEL Q
[C /(PEL )]/n
i
i
(7)
where m-PELQ values of < 0.1, 0.11–1.5, 1.51–2.3 and > 2.3 coincide
with 10%, 25%, 50% and 76% likehood of toxicity, respectively (Long
et al., 1995a, 1995b). Consequently, four relative levels of priority
(highly toxic, medium toxic, slightly toxic and non toxic) have been
proposed to interpret m-ERM-Q and m-PEL-Q values.
4.3.3. Potential acute toxicity
For assessing the extent to which the aquatic organisms may be
influenced by sediment pollution, a potential acute toxicity can be estimated (Swarnalatha et al., 2013; Radhouan et al., 2015 and Soliman
et al., 2015). Potential acute toxicity of contaminants in sediment
samples is the sum of the toxic units (ΣTU) defined as the ratio of the
determined concentration of element i (C i ) to PEL value of element i
(P i ), (Pedersen et al., 1998).
=
TU
C/P
i
i i
(8)
In this study, toxic units (TU) were calculated to normalize the
toxicities caused by various trace metals, which allows the comparison
of their relative effects.
5. Statistical analysis
Pearson correlation analysis was implemented to determine the
relationship between the twelve heavy metals investigated in the surface sediments of Algerian coast. Multivariate analysis (Principal
component analysis (PCA) is an effective tool for providing suggestive
information regarding trace metal sources and pathways (Hu et al.,
2013), and it has been applied on our data set of fifty one stations and
twelve metals. Both correlation and PCA were performed using the
statistical software package STATISTICA version 6.1 for Windows.
6. Results and discussion
6.1. Choice of reference backgrounds
This is not the scope of this paper to discuss the advantages of using
general or local geochemical backgrounds in order to evaluate contamination levels, but the fact is that such choice may drastically affect
the results in term of ecological risk. If the average crustal or shale
values are easy to use, it is also known that they do not represent
correctly the carbonate watersheds (Viers et al., 2009), which are
particularly developed on the Algerian coast. Table 2 reports these
crustal and shale values, and it is clear that the differences observed for
V, Cr, Ni or Co for example will affect the comparison.
We thus used unpublished data to define local reference background
values, in order to get the most realistic evaluation of the contamination. These values reported on Table 2 were derived from three sediment cores collected in the Bay of Algiers, between 40 and 100 m water
depth. These samples were analyzed for heavy metals at the CEREGE
with the same protocol; and our proposed local background values
(Table 2) are the average of 15 samples taken in these cores between 15
and 30 cm depth. We are confident in the fact that these levels are not
affected by recent contamination, and this is confirmed by the low
variability of these values (see Standard Deviation on Table 2). These
cores may not represent the entire Algerian coast, but there are more
adapted than the crustal or shale values, and the chemical contamination indices discussed below are calculated using these local backgrounds.
6.2. Heavy metals levels and relationships
The averages, standard deviations, and ranges of heavy metals determined in the sediments from each station are given in Table 2. The
order of abundance of these metals is: Fe > Mn > Zn > V > Cr >
Pb > Ni > As > Cu > Co > Cd. The average levels along the entire
Algerian coast are lower than the mean shale or crustal values (Taylor,
1964 and Taylor and McLennan, 1985, 1995), except for V, Pb, Zn and
As, and the difference for As is very important.
Viers et al. (2009) estimated the average concentrations in the
suspended sediment of various rivers in the world (Table 2). These
values do not correspond to pre-anthropogenic backgrounds and they
rather give an indication of the global contamination over the world.
Our sediments are clearly below these values and can be considered,
again, as non contaminated.
However, the most realistic comparison as discussed above must be
done with the local reference background. Fe, Mn, V, Cr, Co, Ni, Cu and
Cd concentrations are all below this local background and only the
concentrations of Zn and especially As (factor two) are above it.
The maximum levels of Fe, Mn, Pb, V, Co, Ni, Zn and Cd were found
at station S20 in the region of Beni Saf in the Western part of the coast.
I. Ahmed et al.
Marine Pollution Bulletin 136 (2018) 322–333
326
i is called the
monomial potential ecological risk factor, and it corresponds to the
potential ecological risk of a given contaminant.
The local background (see Section 6.1) is selected as the reference
baselines in this study. The calculated RI values can be categorized into
four classes of potential ecological risks: low risk (< 150), moderate
(150–300), considerable (300–600) and very high (> 600).
4.3. Sediment quality guidelines method (SQGs)
Various sediment quality guidelines are used to protect aquatic
biota from the harmful and toxic effects related with sediment contaminants (McCready et al., 2006). These guidelines evaluate the degree to which the sediment-associated chemical status might adversely
affect aquatic organisms and are designed for the interpretation of sediment quality. They can be used to rank and prioritize both sites and
chemicals of potential concern (Dias De Alba et al., 2011).
4.3.1. Effects range-low and effects range-medium values
The “Effects range low” (ERL) and “Effects range medium” (ERM)
are sediment quality guidelines developed by Long and Morgan (1990)
to relate metal concentrations with their effects. These concentrations
have been separated into three categories: scarcely observed or predicted effect (below ERL), occasionally observed effect (ERL – ERM)
and frequently observed effect (above ERM) (Long et al., 1995a,
1995b). The mean ERM quotient (m-ERM-Q) method is applied, to
determine the possible biological effect of a group of heavy metals, by
calculating a mean quotient (Long and Morgan, 1990; Carr et al., 1996;
Long et al., 2000) using the following equation:
∑
−
− =
m ERM Q
[C /(ERM )]/n
i
i
(6)
where Ci is the concentration of metal i in the sediment, ERM i the
guideline values for the metal and n is the number of metals. The
toxicity probabilities of m-ERM quotient of < 0.1, 0.11–0.5, 0.51–1.5,
and > 1.50 are 9%, 21%, 49%, and 76%, respectively (Long et al.,
2000).
4.3.2. Threshold effect level (TEL) and probable effect level (PEL) approach
The TEL and PEL levels (also referred as TEC and PEC) are other
sediment quality guidelines widely used to assess the ecotoxicology of
sediments. This approach is also based on the relation between measured concentrations of metals and observed biological effects, such as
mortality, growth or reproduction of living organisms. The Threshold
effect level (TEL) refers to the concentration below which adverse effects are expected to occur only rarely, whereas Probable effect level
(PEL) indicates the concentration above which adverse effects are expected to occur frequently (Long and MacDonald, 1998 and Mac
Donald et al., 2000).
The mean PEL quotient (m-PEL-Q) was calculated using:
∑
−
− =
m PEL Q
[C /(PEL )]/n
i
i
(7)
where m-PELQ values of < 0.1, 0.11–1.5, 1.51–2.3 and > 2.3 coincide
with 10%, 25%, 50% and 76% likehood of toxicity, respectively (Long
et al., 1995a, 1995b). Consequently, four relative levels of priority
(highly toxic, medium toxic, slightly toxic and non toxic) have been
proposed to interpret m-ERM-Q and m-PEL-Q values.
4.3.3. Potential acute toxicity
For assessing the extent to which the aquatic organisms may be
influenced by sediment pollution, a potential acute toxicity can be estimated (Swarnalatha et al., 2013; Radhouan et al., 2015 and Soliman
et al., 2015). Potential acute toxicity of contaminants in sediment
samples is the sum of the toxic units (ΣTU) defined as the ratio of the
determined concentration of element i (C i ) to PEL value of element i
(P i ), (Pedersen et al., 1998).
=
TU
C/P
i
i i
(8)
In this study, toxic units (TU) were calculated to normalize the
toxicities caused by various trace metals, which allows the comparison
of their relative effects.
5. Statistical analysis
Pearson correlation analysis was implemented to determine the
relationship between the twelve heavy metals investigated in the surface sediments of Algerian coast. Multivariate analysis (Principal
component analysis (PCA) is an effective tool for providing suggestive
information regarding trace metal sources and pathways (Hu et al.,
2013), and it has been applied on our data set of fifty one stations and
twelve metals. Both correlation and PCA were performed using the
statistical software package STATISTICA version 6.1 for Windows.
6. Results and discussion
6.1. Choice of reference backgrounds
This is not the scope of this paper to discuss the advantages of using
general or local geochemical backgrounds in order to evaluate contamination levels, but the fact is that such choice may drastically affect
the results in term of ecological risk. If the average crustal or shale
values are easy to use, it is also known that they do not represent
correctly the carbonate watersheds (Viers et al., 2009), which are
particularly developed on the Algerian coast. Table 2 reports these
crustal and shale values, and it is clear that the differences observed for
V, Cr, Ni or Co for example will affect the comparison.
We thus used unpublished data to define local reference background
values, in order to get the most realistic evaluation of the contamination. These values reported on Table 2 were derived from three sediment cores collected in the Bay of Algiers, between 40 and 100 m water
depth. These samples were analyzed for heavy metals at the CEREGE
with the same protocol; and our proposed local background values
(Table 2) are the average of 15 samples taken in these cores between 15
and 30 cm depth. We are confident in the fact that these levels are not
affected by recent contamination, and this is confirmed by the low
variability of these values (see Standard Deviation on Table 2). These
cores may not represent the entire Algerian coast, but there are more
adapted than the crustal or shale values, and the chemical contamination indices discussed below are calculated using these local backgrounds.
6.2. Heavy metals levels and relationships
The averages, standard deviations, and ranges of heavy metals determined in the sediments from each station are given in Table 2. The
order of abundance of these metals is: Fe > Mn > Zn > V > Cr >
Pb > Ni > As > Cu > Co > Cd. The average levels along the entire
Algerian coast are lower than the mean shale or crustal values (Taylor,
1964 and Taylor and McLennan, 1985, 1995), except for V, Pb, Zn and
As, and the difference for As is very important.
Viers et al. (2009) estimated the average concentrations in the
suspended sediment of various rivers in the world (Table 2). These
values do not correspond to pre-anthropogenic backgrounds and they
rather give an indication of the global contamination over the world.
Our sediments are clearly below these values and can be considered,
again, as non contaminated.
However, the most realistic comparison as discussed above must be
done with the local reference background. Fe, Mn, V, Cr, Co, Ni, Cu and
Cd concentrations are all below this local background and only the
concentrations of Zn and especially As (factor two) are above it.
The maximum levels of Fe, Mn, Pb, V, Co, Ni, Zn and Cd were found
at station S20 in the region of Beni Saf in the Western part of the coast.
I. Ahmed et al.
Marine Pollution Bulletin 136 (2018) 322–333
326