International Journal of Marine Science, 2017, Vol.7, No.23, 214-228
222
The spatial distribution of Ni was more homogenous between ports with significant high levels in fishing berth of
Hurghada (28.96±3.0 to 61.96±28.0 µg/g) and Qusier old port and fishing basin (38.1±5.0 to 73.3±12.0 µg/g).
Salem et al. (2014) stated that the locations recorded high levels of Ni and partially Co related to anthropogenic
discharges. The levels of Ni in most studied ports were higher than the LEL (16 µg/g) but lower than SEL (75
µg/g) except Ø5 (91.40 µg/g) fraction that exceeding SEL limit at Hurghada shipyard. The overall concentrations
of Ni much higher than the Trade Harbours of South Korea (Choi et al., 2012), within the range recorded in the
Sydney Harbour (Irvine and Birch, 1998) and Hamilton Harbour (Poulton et al., 1996) and were lower than
Victoria Harbour (Wong et al., 1995; Tang et al., 2008) and Ceuta Harbour (Guerra-Garcia and Garcia-Gomez,
2005) (Table 2).
The concentration of Co recorded significantly very low concentrations (p
≤
0.01) at Hurghada passenger port and
Hurghada shipyard, meanwhile the highest availability were observed in Qusier old port and fishing basin (Figure
3). There were no standard guidelines for Co in the marine sediments, but the overall Co levels in current study
were below the standard background of average shale (Förstner et al., 1982) and below the recorded values in
other contaminated sites such as in Sydney Harbour (Irvine and Birch, 1998) and Darwin Harbour (Padovan et al.,
2012).
Cadmium showed significantly low concentrations (p
≤
0.05) at Hurghada ports and was below the lowest effect
levels (LEL), meanwhile Safaga and Qusier ports recorded Cd concentrations higher than LEL (0.6 µg/g) but still
below SEL value (10 µg/g) (Figure 3). Comparing to the levels of Cd with worldwide Harbours (Table 2), Qusier
old port and fishing basin recorded levels similar to Hamilton Harbour (Poulton et al., 1996) and Victoria Harbour
(Wong et al., 1995; Tang et al., 2008) but it is lower than the levels in Sydney Harbour (Irvine and Birch, 1998).
2.3 Associations and sources of the leachable heavy metals
Correlation matrix and Principal Component Analysis (PCA) estimated the statistical relationship among heavy
metals as well as between sediment characteristics and the heavy metals. Additionally, PCA was used to infer the
hypothetical sources of heavy metals contamination (Dou et al., 2013; Qiao et al., 2013; Fujita et al., 2014; Yang
et al., 2015).
The correlation matrix for the different heavy metals in sediment fractions showed strong association between
heavy metals in the mud fraction Ø5, followed by Ø3, and the least association was in Ø4. Two significant
associations were observed; the first association was strong positive correlation between metal pairs of; Cu, Zn,
Pb and Ni in Ø5 to lesser extents of; Cu, Zn, Pb in Ø3. The second association was common in all fractions,
competed Mn with Fe and Cd in strong correlations coincided with a negative correlations of carbonate and with
Fe and Mn (Table 3).
As shown in Figure 4, components in Ø3, Ø4 and Ø5 between metals, the obtained results of the PCA showed
wide accordance with the correlation matrix. In Ø3 (Figure 4A), two main components with accumulative account
for 55.44% of the total variance were found. In the first component (31.98% of the total variance), Fe, Mn and Ni
were grouped with positive loading and carbonate content with negative loading. The second component (23.46%
of the total variance) grouped positive loading of Cu, Pb and Zn. In Ø4 (Figure 4B), two main components were
identified, accounted for 45.32% of the total variance. The first PCA (27.5% of the total variance) was associated
with positive loading of Mn, Fe, Cd and Co accompanied with negative loading of carbonate. The second PCA
(17.83% of the total variance) showed positive loading between Cu and TOM as well as negative loading of Cu
with Cd. In Ø5 (Figure 4C), three components (with total account of 72.49%) were recognized. The first
component (32.73% of total variance) included strong positive loadings of Cu, Zn, Pb and Ni. The second
component (24.35% of total variance) showed positive loading between Mn and Fe associated with negative
loadings toward carbonates. The third PCA component (15.42% of total variance) showed positive loading of Cd
with Co and negative loading of Cd and Co toward TOM.
