Seven sampling sites were selected along the ULRB. Figure 1 shows the geographic distribution of these sites. The characteristics for each sampling site can be described as follows:

Site 1: Oulaik spring (Roads with high traffic density; farms and industry for dairy products).

Site 2: Houch El Rafika village.

Site 3: Bedanyl village (Intensify agricultural land: potatoes and wheat; farms; industries for slaughter of poultry; coffee shops and tourist activities).

Site 4: Bednayl 1 village (Electricity power station; water industry above 400 m).

Site 5: Berdawnil 2 (Industries; tourist activities and restaurant; in front Lebanese France hospital).

Site 6: Dair Zanoun village (End of intensive agricultural land; potatoes and wheat; farm; road with high traffic density (Lebanon-Damascus); paper mills; tanneries).

Site 7: Massabiki village (Road with intensive traffic density; Massabiki hotels).

The selection of these sites was based, firstly, on the practicability of collecting sediments; especially accessibility, and secondly their location with respect to typical activities of industries like dairy and poultry farms, tanneries, paper mills, wheat and potatoes, tourist and high vehicle movements located around the ULRB and their effect in the reject direct of pollutants.

Approximately 2 kg of sediment were collected from each of the sampling points at the sediment-water interface (i.e. surface river bed sediments) using a polyethylene bags. Water samples were collected simultaneously with sediments at each sampling site using two 1 L polyethylene bottles. Each bottle was supplied by 2% of nitric acid to acidified (pH < 2) and stored in portable coolers before transport them to the laboratory.

The pH, Electrical Conductivity (EC) and Total Dissolved Solids (TDS) were determined in-situ, and water samples subsequently filtered through 0.45 μm membrane in the laboratory. Sediment samples were dried at room temperature and sieved; where the sediment size less than <65 μm size was retained.

The concentrations of Pb, Cd, Fe, Mn, Cr, Cu, and Zn were measured using the Inductively Coupled Plasma-Optical Emission Spectrometry (ICP-OES) with Ultra Sonic Nebulizer (USN) (Model: Perkin Elmer Optima 3000). The samples were filtered by filtration system through membrane filter with pore size 0.45 μm before the analysis using Standard Methods (APHA, 1992). Sediment samples were digested using microwave digestion techniques as reported by Littlejohn et al. (1991) in which 0.5 mg of sample was placed in Teflon vessel with 5 ml HNO₃ (65%), 2 ml HF (40%) and 2 ml H₂O₂ (30%) by using Microwave digestion system (model: MILESTONE mls-1200 mega). An aliquot of the filtration of the samples was taken (i.e. about 100 ml). Digestion solutions were measured for the total heavy metals using ICP-OES (APHA, 1995). Digestion extracts were used to analyze the total metal concentrations by Atomic Absorption Spectrophotometer (AAS). Blanks containing all the components except sediments were analyzed to determine background interferences. All measurements were performed in triplicate and average values were reported.

Approximately 10 g of air-dried sediments was suspended in 50 mL of deionized water and manually agitated for 5 min. The suspension was allowed to rest for about 1 hour with occasional shaking until the pH, TDS and EC were measured.

Cation Exchange Capacity (CEC) was measured after it was exchanged with cobalt hexamine (Co(NH₃)₆Cl₃) and dosage of its residual concentration in the equilibrium solution (Morel, 1957; Mantin & Glaeser, 1969). About 1.5 g of sample was dispersed and shaken during two hours at 30˚C in 30 mL of cobaltihexamine solution (16.6 × 10⁻³ mol/L). Samples were then centrifuged for one hour at 46.251 g. Supernatants analyzed on a UV–Visible Spectrophotometer, using cobalt absorption band at 472 nm to derive CEC from residual concentration in cobaltihexamine. Measurements were always carried out in duplicate to check for further reproducibility.

Sediment samples were analyzed by X-Ray Diffraction (XRD) using 1 g of randomly oriented sediment powder, that put on a rotating sample holder and leveled with a glass slice to obtain a flat surface. XRD was also acquired using zincite (ZnO) as internal standard. In this latter case, samples were first mixed with 0.111 g of ZnO and ground in an agate mortar for 5 minutes. XRD patterns were collected on a D8 Advance Bruker AXS Diffractometer equipped with a Lynx Eye fast linear detector using Cobalt K radiation (k = 0.17903 nm) at 35 kV and 45 mA. Intensities were recorded from 3˚ to 64˚ with a 0.035˚ step using a 3-second counting time per step. In the presence of standard, intensity was recorded from 4˚ to 80˚ with a 0.02˚ step using a 3-second counting time per step. Data reduction and analysis were performed with the EVA software (DIFFRAC plus from Bruker) and diffraction peaks were determined by comparison with powder diffraction files.

FTIR spectra were collected on bulk samples powder mixed with KBr (KbrO₃) as a transparent matrix. Diffuse reflectance FTIR were recorded on a Bruker IFS-55 Spectrometer in the range from 4000 to 600 cm⁻¹ with a resolution of 2 cm⁻¹. Spectra were obtained from the average of 200 scans collected during 2 minutes. Data reduction and analysis were performed using OPUS program from Farmer (1974), Marel and Beutelspacher (1976) and Russell and Fraser (1994).

Multivariate statistical analyses, including Pearson’s Correlation Analysis (PCA), Cluster Analysis (CA) and Factor Analysis (FA) were performed in this study to reflect the degree of dispersion distribution of the different metals. PCA was carried out by using SPSS16 of 2007. CA is a multivariate technique, whose primary purpose is to classify the objects of the system into categories or clusters based on their similarities, while, FA was carried out to assess the relationship between different variables of heavy metals using SPSS16.

The resulted pH values of the ULRB range between 7.7 and 8.5, which is due to the intensive agricultural activities and the excessive use of fertilizers such as ammonium phosphates. The conductivity and total dissolved solids were found within the permissible limits: 384 - 1484 µs/cm and 159 - 321 mg/l, but only in sites S3 and S6 (1688 and 1780 µs/cm), were high and this could be due to the presence of intensives industrials activities that often release wastes directly into the river without treatment. The total concentration and the mean values of element concentration in the water samples of the three seasons for two years are shown in Table 1.

The heavy metals were differently distributed in all the selected sites, and this due to the presence of different aspects of activities such as industries, intensify agriculture urbanism and tourist, and thus are mainly responsible for elevated levels of the measured elements in ULRB. Mean concentration of heavy metals were ranging as follows: Mn: 0.49 - 1.17 mg/l; Fe: 0.23 - 1.78 mg/l; Zn: 0.28 - 0.54 mg/l; Pb: 0.11 - 0.91 mg/l; Cr: 0.03 - 0.54 mg/l; Cd: 0.04 - 0.05 mg/l; Cu: 0.10 - 0.20 mg/l; and this allowing to arrange the metals from higher to lower mean content in this area as: Mn > Fe > Zn > Pb > Cr > Cd > Cu.

Figure 2 shows that metals concentration was varied between different seasons. Thus, Pb, Mn, Cu and Cr were high in concentration in dry season (class 3). Fe and Cd, were important in rainy season (class 1), while the Zn was observed in high concentration in mid rainy (class 2).

According to Suresh et al. (2011), if the correlation coefficient between the metals is high, then metals have common sources, mutual dependence and identical behavior during their transportation. The absence of correlation among the other metals suggests that the contents of these metals are not controlled by a single factor during three seasons. However, it could be controlled by a combination of geochemical support phases and their mixed associations. In the present study, Pearson’s Correlation Coefficient (PCC), matrix among the selected heavy metals is shown in Table 2. The highly significant positive correlation (r = 0.612) between the concentrations of Fe and Cd suggests that the association of the two elements originates from a common source of input, and also during transportation or deposition processes and this also confirm with what present in Figure 2.

Significant positive correlations between the contaminants Mn and Cu (r = 0.791), indicates the same source of input. Figure 2 confirms that these metals are associated and concentrated in the dry season.

The concentrations of Zn during the winter closure period were lower than those after the mid rainy period (Table 1). This is attributed to the increased consumption of this element by the others compound present in water. The negative correlation between Cu and Zn concentrations (r = – 0.617) it could be a result of precipitation of the elements as hydrous metal oxides (El-Sayed & El-Sayed, 1980), whereas the concentrations of Zn increase in mid rainy season and Cu more concentrated during the winter closure period. The major sources of Zn are the liquid and solid domestic and municipal wastes followed by dumping and atmospheric deposition (Issa et al., 1996).

The relationship between EC and TDS varies with the concentration of salts in the water and the proportions of various present salts, while the correlation between TDS and EC (r = 0 .988) was positive and significant.

*.correlation is significant at the level (2-tailed); **.correlation is significant at the 0.05 level (2-tailed).

Measured of pH in sediments was shown to be higher than that in the water, which is ranged between 7.00 up to 8.45, it is slightly alkaline and this could be partly contributing to the increased concentration of heavy metal ions in the sediments due to precipitation of dissolved metals. EC and TDS are related together and they were increased. EC was ranged from 319 to 1484 um/cm and TDS varies from 171 to 884 mg/l (Table 3).

Pearson’s Correlation Coefficient matrix among the selected heavy metals, physical parameters such as EC, TDS, PH, and CEC; and the concentration of relative distribution of Quartz (Q), Calcite (C), Dolomite (D) and kaolinite (K) are shown in Table 9. Pb tends to remain tightly bound to sediments under oxidizing conditions. Positive correlation coefficients are observed among the metals Pb and Mn (r = 0.549), and this is normal, because, the Pb in sediments is primarily found in association with manganese hydroxide (Nehmeh et al., 2019). There is significant correlations between the contaminants of Pb and Cu (r = 0.796), Cr (r = 0.848), Zn (r = 0.562); and TDS (r = 0.449); and this is due to the same source of contamination, such as wastewater, industrial and agricultural activities residues. Cd and Cr (r = 0.445); Mn and Cu (r = 0.582); Cu and Cr (r = 0.721); Cu and Zn (r = 0.649); TDS and EC (r = 0.988).

The major minerals of quartz, calcite, dolomite and koalonite are poorly correlated with the studied metals. However, sand mineral (i.e. quartz) has positive correlation with EC (r = 0 .450) and TDS (r = 0.506); CEC and pH (r = 0.848) have positive correlation and significant, and this was observed in all sites. CEC and dolomite are positively correlated (r = 0.515), and this is due to the competition between Ca²⁺ and Mg²⁺ as cation substitution to incorporated in CEC. Calcite and dolomite are significant positively correlated (r = 0.775), because the majority of rock lithologies in Lebanon is principally formed from carbonate rocks (limestone and dolomite).

Q = Quartz; C = Calcite; D = Dolomite; K = Kaolinite.

Factor analysis (FA) is carried out on the data set (16 variables as in the above analyses). FA yielded three factors with explaining 63.44% of the total variance (Table 10). The first factor accounted for 30.44 % of the total variance, and it is mainly characterized by positive loading of concentrations of Pb, Cr, Cu, Mg, CEC, pH, TDS, EC quartz, kaolinite and calcite.

Factor 2 accounted for 18.29% of the total variance, which mainly consists of positive loading of Fe, Mn, and dolomite mineral, whereas factor 3 accounted for 14.71% of total variance and showed positive loading of Cd. From the overall factor analysis, it was found that calcite is the major factor and quartz is the least important factors to increase the heavy metals concentrations.

Figure 6 allows distinguishing that the factor of seasonal variations is not important, because there is combination between the sampling sites.

Journal of Geoscience and Environment Protection

The first class (Class 1/3) has found in the positive right side of the axis F1 and F2, and it contains the sites of S1r, S6r and S7r during rainy season. This class characterized by the presence of heavy metals of Pb, Cu and Zn. The third class (Class 3/3) corresponds to spring (mid rainy) and has found in the positive side of C1 and C2. This means that no variation occurs between two seasons. The

second (Class 2/3) corresponds to mixed of samples in summer season, whereas the samples during rainy season has found in negatives side of the axis C1 and C2 and have the influence of the metals Cr, Mn and Fe.