PAH concentration in soils and plant tissue
Out of the 16 PAH component chemicals analysed from soil samples, measurable concentrations were detected for 14 PAHs (Table 1: fluorene (FL) and acenaphthene (ANA) were below levels of detection in all soil samples). There was considerable variation in both the composition and concentrations of PAHs making up the total PAH load in samples from each area, with values for individual PAHs varying from 0.1 μg kg-1 (e.g. benzo(g,h,i)perylene (B[ghi]PERY) in a sample from Sabah Al-Ahmad Protected Area), to 1010.0 μg kg-1 of chrysene (CH) and phenanthrene (PH) in a soil sample from Sabriya Oilfield. Chrysene (CH) was usually the predominant PAH, especially in Sabriya Oilfield (where it was found in 30 samples). Next commonest were phenanthrene (PH) and pyrene (PY) in soils collected from different conditions (fresh/dry oil lakes, and deposit piles of contaminated soil at several locations). Other PAHs were generally either present in lower concentrations or were below detection levels.
The highest values of total PAH contents in soils were seen in samples from two northern areas (Sabriya Oilfield and Bahra, the latter being located close to two major oilfields and strongly affected by oil flows and hydrocarbon deposition in 1991). Elevated total PAH values (though much lower than the values seen in Sabriya soils) were also found in soil samples from two other oilfield areas (Burgan and Um Al-Aish, with some individual samples exceeding 200 µg kg-1 soil dry weight for total PAH), but not in the Um-Ghadair Oilfield, where the total PAH in soil was close zero to very low values usually detected in unpolluted areas such as Sabah Al-Ahmad and Umm Al-Rros.
PAHs in plant leaf tissues collected from the two areas with the highest total PAH soil contamination (Um Al-Aish and Sabriya Oilfields: Table 2) showed a different pattern from the soils. There were significant between-area differences in values for some individual PAHs (e.g. fluorene) but there was no significant difference in total PAH content, with high values found in leaf samples collected from both areas. Mean values in plant tissues were also higher than in soils for some individual PAHs, suggesting substantial uptake by the plants. Phenanthrene (PH), one of the most abundant PAHs in soils, also showed the highest mean concentration in plant tissues, with the highest concentration detected in Cyperus conglomeratus . Suprisingly, fluorene (FL), which was detected in high concentrations in leaf samples taken from Sabriya, and to a less extent in samples from Um Al-Aish, was never detected in soil samples. Other individual PAHs with moderate to high leaf tissue concentrations were fluoranthene (FLAN), naphthalene (NA) and pyrene (PY). The remaining PAHs showed relatively low concentrations or were not detected in plant tissues. It is interesting to note that despite the high concentrations of chrysene (CH) found in soils, this PAH compound showed very low uptake by plants. Similar phenomenon was also observed for pyrene (PY). There was a small but significant difference in plant tissue moisture content between leaves sampled from plants growing in the two areas (Table 2).
Comparison of environmental and vegetation variables between TWINSPAN sample groups
A total of 42 species (Supplementary File 1 ) were observed from the sample-sites, with the highest area diversity (SA) found in the Sabah Al-Ahmad Protected Area, followed by the Um Al-Aish Oilfield area. The results of multivariate classification of the samples x species matrix provide strong evidence (Table 3) that variation in the vegetation present was not solely associated with geographical location. Seven vegetation sample groups supporting vegetation (labelled Groups A – G) were identified by TWINSPAN at level 2 of the sample classification, and an eighth group (H) contained samples entirely lacking vegetation. Sample groups tended to be separated with high eigenvalues in the TWINSPAN analysis, indicative of substantial between-group differences in terms of plant assemblage present in their component samples.
Indicator species were identified for each sample group supporting plants (other than the three major focused plants). Results of one-way ANOVA comparisons (or non-parametric Kruskal-Wallis Test, for non-normal variables) between TWINSPAN-defined sample groups for individual environmental and vegetation variables (Table 3) provided evidence that Groups C and D had significantly elevated levels of soil PAHs compared with the remaining sample groups, whilst Group B had an intermediate PAH loading. Samples making up these groups also tended to be allocated high ODS values. The soil samples making up these two groups was very dry, and the mean species diversity of the two groups also tended to be low. However, there was no evidence for suppression of vegetation height in Groups C and D, with mean values falling in the intermediate range of the eight sample groups. These three groups were characterised by the presence of one or more species of Haloxylon salicornicum ,Cyperus conglomeratus and Rhanterium epapposum , whilstZygophyllum qatarense (Zygophyllaceae), Salsola imbricata(Amaranthaceae) and Stipagrostis plumosa (Poaceae) were also indicators for elevated PAH soil conditions in at least one of these three sample groups.