Occurrence, distribution, and ecological risk assessment of emerging and legacy contaminants in the Kadicha river in Lebanon

The Kadicha river basin in Northern Lebanon is an illustrative example of multiple pressures encountered in the Mediterranean region: it is a small coastal river affected by rapid urbanization, population growth (drastically impacted by the influx of Syrian refugees), and a chronic default of wastewater treatment. In this context, multiple classes of contaminants may attain the river accumulating in sediment. However, very little information is available in the literature on the contamination status in such stressed Mediterranean contexts. This study proposed a first contamination evaluation of a small Mediterranean river submitted to multiple pressures. Two sediment sampling campaigns along sites impacted by increasing urban gradient within the Kadicha river basin were performed to determine the occurrence and the environmental risks of both emerging and legacy contaminants. The results revealed the detection of the 41 studied compounds. The highest concentrations were attained by PAHs and polycyclic musks (up to 311.79, 94.22, and 81.13 ng/g of dry weight for PAH, cashmeran, and galaxolide, respectively). The discontinuous urbanized upstream area and the estuary were the most contaminated areas of the river. An environmental risk assessment showed a hazard quotient (HQ) higher than 1 for both legacy and emerging compounds (EHMC and 4-MBC), indicating a potential risk to benthic species. Monitoring campaigns and implementation of wastewater treatment plants should be encouraged as the anthropogenic pressure on small Mediterranean rivers will increase over the years.


Introduction
The Mediterranean basin region has been affected for millennia by human activities. Today, the Mediterranean basin suffers from disturbances derived from agriculture, livestock, industrial practices, and human population growth. Anthropogenic activities within the Mediterranean rivers' watersheds have resulted in changes in uvial geomorphology, organic matter dynamics, and increases in pollution and salinity (Cooper et al., 2013). The large rivers constitute the major contributors to pollution input into the sea (Ludwig and Probst, 1998;Roy et al., 1999;Abril et al., 2002). However, small Mediterranean rivers contribute to the pollution, especially where population density is high, causing environmental and health risks in river watersheds and bringing to the sea the contamination occurring in the coastal area (Kolpin et al., 2004;Nicolau et al., 2012). All these anthropogenic activities within small Mediterranean watersheds were associated with the releasing of different pollutants, known as persistent organic pollutants (POPs) and emerging pollutants. Historically, POPs (e.g. polycyclic aromatic hydrocarbons (PAHs) and polychlorinated biphenyls (PCBs)) have been monitored and regulated in most parts of the world for the last decades, that's why they are referred to as "legacy contaminants" (Jones and Study area and sampling The Kadicha river in north Lebanon also known as "Abou Ali" is a small Mediterranean river affected by rapid urbanization and population growth. The river land use watershed is representative of modern anthropization in most of the Mediterranean rivers: mixing urban, discontinuous urban, and rural areas with agricultural, recreational activities and small industrial facilities. The residential wastewaters are discharged directly to the ood channel without prior treatment (Naja and Volesky, 2013). Besides, an open area dumpsite at the estuary of the river receiving urban solid waste releases approximately 24 000 metric tons of leachate per year in the estuary (or the lower part) of the river (Naja and Volesky, 2013). Recently, the downstream of this watershed have shown a population growth after the arrival of Syrian refugees (AEMS, 2017) which may lead to an increase in anthropogenic pressure.
The Kadicha river is characterized by a Mediterranean climate with a moderately warm and dry summer and autumn, and moderately cold, windy, and wet winters with almost 80-90% of total precipitation occurring between November and March (Massoud et al., 2006). The Kadicha river ows into the Mediterranean Sea, with a basin draining a total of 484 km 2 (SOER, 2001). The river basin has a length of 44.5 km and an average annual discharge of 262 million m 3 (SOER, 2001). The sampling sites are presented in Figure 1. The land use of the river basin mapped out using the QGIS software showed that the upstream of the river is dominated by agricultural surfaces with some discontinuous urban tissues while the downstream is mainly urbanized. The river's waters are mainly used for domestic supply, hydroelectric production, irrigation, and recreational activities.
Groundwater is the main source of drinking water for residents within the basin.
Twelve sampling sites were selected to realize the possible compromise between representativeness of potentially polluted sites and operational feasibility. The selected sites are described in Table 1 in terms of geographic location and anthropogenic pressures. Three sites (Koussba (Ks), Bshennine (Bs), and Bkeftine (Bk)) were in the rural upstream area designated RU. Four sites (Meryata (Myt), Ardeh (Ar), Al Merdechyeh (Mr), and Zgharta (Zgh)) were in the discontinuous urbanized upstream area designated DUU. These 2 areas are in uenced by agricultural activities with DUU area gathering a higher density of population than RU area (Table   1). Five sites were in the coastal plain occupied by the city of Tripoli: 3 sites (Abou Samra (T-Ab1 and T-Ab2) and Al Marjeh (T-Mj)) in the urbanized downstream area designated UD and 2 sites (EAA-1 and EAA-2) in the estuary designated EST. Tripoli is highly urbanized with a population estimated to 500000 inhabitants (UNEP, 2009). This population has increased after the Syrian crisis. According to UNHCR, the number of Syrian refugees in Tripoli registered as of 30 June 2016, including 7 percent of non-registered displaced Syrians living in informal settlements, reached 52350 (or 13644 families) (AEMS, 2017). The sewage of Tripoli city is discharged into surface water streams or directly through short sea outfalls without prior treatment (Naja and Volesky, 2013 Surrogates acenaphthene d10, phenanthrene d10, chrysene d12, 4-MBC d4, and 4n-NP-2,3,5,6-d4,OD were spiked at 50 ng/g to sediment samples 24 h before extraction to account for losses during the extraction procedure.

GC-MS/MS analysis
The separation, identi cation, and quanti cation of the 41 molecules were performed using gas chromatography coupled to triple quadrupole mass spectrometry. Capillary gas chromatography analysis was carried out on a Macherey-Nagel column (30 m × 0.25 mm i.d. × 0.25 m lm thickness), keeping the carrier gas ow (helium) at 1.7 mL.min -1 , and the transfer line and the temperature of the source at 300 and 250 ° C respectively. The column temperature ramp was as follows: 78 °C for 0.1 min, increased at 13°C/min to 140 ° C, then at 8°C/min to 180°C then increased at 5˚C/min to 220˚C and then at 3˚C/min to 300˚C held for 10 minutes.
Injection volume was 2 µL in splitless mode and the solvent delay was set to 5 minutes. The mass detector was operated in multiple reaction monitoring (MRM) mode using electron ionization (EI) source set at 70 eV and argon as collision gas (1.5 bar). MS/MS parameters were optimized by injecting standard solutions, using full scan mode (m/ z 50-650) on a rst step to select precursor ions (Q1) that were later fragmented into product ions (Q3) testing different collision energies (CE) (from 10 to 30 eV). All the data were processed using the Xcalibur software. The noise type selected was the root mean square (RMS). The parameters of the multiresidue method are shown in Table 2. Table 2 Molecules, precursor, quanti cation and con rmation ions, the collision energy (CE), and retention times. 2 isomers of the compounds CYP, DELT, and PER. Quality assurance/quality control Quality assurance/quality control procedures were applied to ensure that results are reliable. Method blanks (solvent) were extracted and analyzed as a control in the same way as the samples and no target compounds were detected in the blanks. A standard solution of target compounds was analyzed at the beginning and after each sample sequence to monitor the instrumental and potential contamination during GC-MS/MS detection.
Calibration curves were constructed in heptane for each compound in the range of 0-1000 ng/mL for PAHs and 0-100 ng/mL for the other compounds. Instrumental limits of detection (iLOD) and quanti cation (iLOQ) for each target compound were calculated based on the signal to noise ratio of 3 (iLOD) and a signal-to-noise ratio of 10 (iLOQs) near the target peak by using calibration curve solutions in the range of 0-50 ng/mL. The method limits of detection (MDL) were estimated by multiplying the iLOD by the volume of the nal extract of sediment (1mL) and then dividing it by the mass of extracted sediments (2 g). The method limits of quanti cation (MQL) were estimated by multiplying the MDLs with a factor of 10/3. These parameters are reported in Table 3.
The calculation of the matrix effects is essential to quantify molecules at trace levels in the environmental matrix. The matrix effects of the analytical method were evaluated using solvent and matrix-matched calibration curves. The matrix effects were negligible (between -20 and 20%) for 11 compounds. However, the other molecules presented a high matrix effect (between -100 and 191%). Due to these high matrix effects, the quanti cation of all analytes was done with matrix-matched calibration curves using ve orders of magnitude (2.5, 12.5, 25, 50, and 100 ng/g and up to 1000 for PAHs). Table 3 Instrumental (pg) and method (ng/g dry weight) limits of detection and quanti cation of the target compounds.

Occurrence and distribution of contaminants
Frequencies of detection and concentration ranges of the studied compounds are presented in * <MDL: concentration lower than the method detection limit and <MQL: concentration lower than method quanti cation.
The distribution pattern within each class of contaminants occurring in the different areas (RU, DUU, UD, and EST) is shown in Figure 2 for the month of February. The pro le is similar in September except for pesticides. UVFs and UVSs showed a quite homogenous distribution within DUU, UD and EST areas, with OC accounting for almost 40% of the total substances whereas 4-MBC accounted for almost 38% of the RU area. Among PCMs, DPMI showed the highest percentages, accounting for 34% of the studied areas except for DUU, where AHTN was found at the highest concentration. The distribution pattern of pesticides in February showed a strong contrast between the DUU and RU areas, where PER and CYP showed the highest percent, respectively. CYP accounts for 90% of the pesticides in the RU area and PER accounts for 87% in the DUU area. These two pesticides showed equivalent concentrations downstream of the river (UD and EST). In September, PER and CYP were the pesticides found at the highest concentration for the four areas. CYP accounts from 21% in the UD area to 40% in the DUU area. PER accounts from 36% in the DUU area to 66% in the UD area. The contribution of irgarol to the total pesticides was low (≤ 10%). Based on their chemical composition, PAHs were divided into 2 groups including low molecular weight compounds (LPAHs) with 2 or 3 aromatic rings and high molecular weight compounds ( However, the population density value used is derived from estimation and further studies should be carried out to use this variable in explaining sediment contamination. Total organic carbon (TOC) content in sediment samples ranged from 1.64 to 5.47 % during the wet season survey and from 1.77 to 7.14 % during the dry season survey. The percentage of ne particles (<63µm) and TOC content in sediments of the studied sites for the 2 studied campaigns are shown in Table S1. It is interesting to note that Pearson correlation analysis showed no correlation between the percentage of ne particles and any of the studied compounds. For TOC content in sediments, no correlation was found with concentrations of UVFs,

Risk assessment
The hazard quotients (HQ) were calculated for compounds with available ecotoxicology data on benthic species (8 emerging contaminants and all legacy contaminants). T-tests were conducted to investigate differences between the HQ values calculated for the two months' surveys (February and September). Differences between the two months' surveys were not signi cant, thus HQ values corresponding to the month of February were interpreted. The values of HQs for individual compounds at every sampling site are shown in Table S2.
For emerging contaminants, HQ values for EHMC and 4-MBC were higher than 1 for the majority of the studied sites exhibiting values up to 11.3 (within the site EAA-2) and 7.1 (within the site Ar) respectively. However, HQ values for the other emerging contaminants were lower than 1 within all the studied sites. For legacy contaminants, the hazard quotient was calculated for the sum of PCBs, the sum of PAHs with low molecular weight (LPAHs), and the sum of PAHs with high molecular weight (HPAHs). HQs for Legacy contaminants were found to be higher than 1 at all sampling sites: HQs for LPAHs were the highest on the majority of the studied sites (>10) followed by PCBs achieving a maximum value at site EAA-2 (31.1) within the estuary area.
The spatial distribution of HQs of legacy and emerging contaminants in the Kadicha river basin are shown in Figure 5. It is important to note that the sum of hazard quotients is not a quantitative measurement but a way to illustrate the spatial distribution of both groups of contaminants at each sampling area. The distribution of HQs among the four different areas is similar to the distribution of the concentration of most legacy and emerging contaminants in sediment. It is clear that HQs for legacy contaminants are signi cantly higher than for emerging contaminants. Regardless of the group of contaminants, the estuary area shows the highest HQs, while the rural upstream area shows the lowest HQs. This suggests that the estuary presents the highest potential risks while the rural upstream area shows the lowest potential risks for benthic species.

Discussions
Occurrence, distribution, and sources of legacy and emerging contaminants The anthropogenic pressures currently exerted within the small Mediterranean river watersheds are well represented in the Kadicha river with high human pressure along the river. The discontinuous dwelling place observed in the mountainous rural areas becomes denser in the downstream part of the river with the densely urbanized area in the estuary. Higher anthropogenic pressure is re ected in higher sediment contamination in more densely populated areas. The discharge of leachates and non-treated wastewater e uents from urban areas contributed to the highest concentrations of UVFs, UVSs and PCMs obtained in the estuary and also within DUU independently of the season. HHCB and UV-236 concentrations then appeared correlated signi cantly with population density and might represent promising markers of human activities in future studies within the watershed. Among the anthropogenic pressures, pesticides are used in general as markers of agricultural activities. We selected pyrethroids instead of other pesticides groups because of their important use in mosquito control by residential of the urban area during the summer season. Consequently, concentrations of pyrethroids during September's survey were higher than during February's survey. Moreover, contamination of sediments from the discontinuous urbanized upstream was higher than the other groups of sites. These ndings suggest that the mixed land use, combining agricultural, urban and residential zones can be considered as the prominent source of insecticide contamination in the Kadicha river basin. Among pyrethroids, PER can be used as a marker of human activities due to its correlation with population density. On the opposite, the low levels of concentrations of irgarol, used as a booster biocide added to antifouling paints (Thomas et al., 2002), measured in the river sediments can be explained by the absence of boat activity within the river.
Alkylphenols are products of degradation of polyethoxylated alkylphenols (Giger et al., 1984)  Fate in the aquatic environment: legacy vs. emerging contaminants The presence of the legacy and emerging contaminants in the Kadicha river is due to anthropogenic pressures discussed previously and depends on the fate of these contaminants in sediment. Half-lives of legacy and emerging contaminants reported from the literature are summarized in Table S3. The half-lives of the studied PAHs in anaerobic soil reported from the literature range between 124 (Phe) and 455 days (BghiP) (Table S3) (except for Naph, Ace, and Flu with half-lives < 65 days). After their deposition in sediments, PAHs are less subjected to photochemical or biological oxidation, especially in anoxic sediment (Nemr et al., 2007). Thus, PAHs tend to be persistent in sediments, where they may accumulate at high concentrations (Witt, 1995). This can explain their correlation with TOC content in the studied sediments. The biodegradation of PCBs was reported to be very slow in sediments with half-lives between 3650 days (PCB 52) and 13750 days (PCB 180) (Beurskens et al., 1993;Sinkkonen and Paasivirta, 2000;Magar et al., 2005). In this study, PCBs concentrations did not show a correlation with TOC content. Our ndings are in accordance with the study of (Vane et al., 2007). The correlation PCBs-TOC can be discriminated by other geochemical processes such as sediment composition of organic matter that may play a greater role than TOC content (Landrum and Faust, 1991;Vane et al., 2007). The half-lives of the studied emerging contaminants were shorter and ranged between 1.9 (AHTN) and 141 days (UV-326) except for BIF (425 days), PER (197 days), and MTC (443 days). This suggests that biodegradation processes for the majority of emerging contaminants except the aforementioned compounds can occur in sediments more easily than for legacy contaminants. However, the biodegradation processes should be interpreted with caution as the half-lives of contaminants can differ from a matrix to another.
While biodegradation of legacy contaminants has been studied for decades, there is a lack of studies on the biodegradation of emerging contaminants in soils and sediments, and these molecules need to be more studied to understand their fate in the environment.

Ecotoxicological implications
The lipophilicity of the studied compounds enhances their accumulation in sediments and as a consequence their accumulation in benthic species (Vimalkumar et al., 2018). It has been reported that the ecological risk of highly lipophilic chemicals (Kow>5) tends to be ampli ed through bioaccumulation and trophic transfer ( Tang

Local and international context
In Lebanon, if there is some data on legacy contaminants, there is a lack of research on emerging contaminants (Tables S4, S5, S6, S7, and S8). Only one study has been conducted on the occurrence of UV lters (EHMC, OC, and ODPABA) in river sediments (Amine et al., 2012) while no studies have been conducted on UV stabilizers in Lebanese river waters. Among UV lters, the concentrations of EHMC and OC were in the same range of the previous study of Amine et al., (2012) conducted on the estuary of the Kadicha river (Table S4). Whereas the concentration of ODPABA was slightly lower in the present study. Moreover, the ubiquity of OC in the present study is in accordance with the study of (Amine et al., 2012) which is probably due to its wide usage in personal care products and cosmetics (ECHA, 2019). The predominance of OC in the estuary area can also be explained by its multiple sources: OC can occur due to direct discharge of the leachates in the estuary. In addition to its point source, OC is highly stable to photodegradation (Rodil and Moeder, 2008)  No studies were conducted on the occurrence of fragrances in Lebanese rivers. Among polycyclic musks, HHCB and AHTN were the most studied in riverine sediments (Table S5)  (2017a) who found that DPMI and ADBI were the fragrances with the lowest concentrations among 5 studied compounds (HHCB, ADBI, OTNE, AHTN, and DPMI) in coastal sediments from the Cadiz Bay in Spain while the HHCB was the predominant contaminant. This difference can be due to the preference of usage of fragrances by country.
The majority of the studies regarding the occurrence of pyrethroids were conducted in the USA and China (Table  S6). Thus, there is a lack of studies conducted within the Mediterranean basin region concerning these contaminants. This is a novelty for the present study. In Lebanon, two studies were conducted on the occurrence of bifenthrin in riverine surface water samples, groundwater, and drinking water (Kouzayha et al., 2013;Youssef et al., 2015). High concentrations of pyrethroids were found in sediments from Nhue and Lich rivers in Vietnam (Duong et al., 2014)

Conclusions
The present work is the rst study to assess the distribution of both legacy and emerging contaminants in a Lebanese watershed. All of the studied compounds were detected in sediments with legacy contaminants detected more frequently than emerging contaminants. Both groups of contaminants showed an increase of concentrations from the upstream to the downstream of the river associated with greater anthropogenic pressures. Different anthropogenic sources of contamination could be identi ed: contamination with UVF, UVSs, and PCMs was linked to direct discharge of raw wastewaters and leachates in the water body while contamination with pyrethroids and PAHs were related to pest control and vehicle tra c respectively. Moreover, several molecules (PAHs, PER, HHCB, and UV-326) have shown a signi cant correlation with population density and might represent promising markers of human activities in future studies within this watershed.
Biodegradation of emerging contaminants in sediments could occur more easily than for legacy compounds. The most elevated potential risk for benthic species was observed within the estuary area. Anthropogenic activity in this highly urbanized area contributed to the observed levels of contamination, combined with the lack of actual treatment of the wastewater produced. Moreover, several emerging and all the studied legacy compounds may pose adverse effects for both benthic and aquatic species in the Kadicha river basin. In general, concentrations of the studied contaminants did not exceed concentrations found in the literature and were lower than levels found in highly urbanized and industrial sites.
As all of the studied compounds were detected in this watershed, monitoring programs of these compounds are encouraged speci cally within the urbanized areas of the river to make future public policies on water quality in Lebanon. Moreover, signi cant efforts should be made to collect and treat wastewater in this watershed to improve the situation of the coastal environment receiving these waters. The assessment of organic contaminants within this watershed may help to understand the potential contribution of this small Mediterranean river to the contamination of the coastal area.
Declarations Figure 1 Location of the sampling sites in Kadicha river basin. The sites are presented by their code: Ks for Koussba, Bs for Bshennine and Bk for Bkeftine in the rural upstream area (RU), Myt for Meryata, Ar for Ardeh, Mr for Al Merdechyeh and Zgh for Zgharta in the discontinuous urbanized upstream area (DUU), T-Ab 1 and 2 for Abou Samra and T-Mj for Al Marjeh in the urbanized downstream area (UD), EAA-1 and 2 for the estuary in the estuary area (EST). Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.

Figure 2
Page 33/35 Distribution patterns of contaminants shown in mass percentage for each class within the 4 studied areas: rural upstream (RU), discontinuous urbanized upstream (DUU), urbanized downstream (UD), and estuary (EST). The pesticides show a seasonal variation between the 2 surveys. studied variable from one sampling area to the three other areas at the 95% con dence level (p-value < 0.05*, pvalue < 0.01**).