Little Akaki River Sediment Enrichment with Heavy Metals, Pollution Load and Potential Ecological Risks in Downstream, Addis Ababa, Ethiopia.

Background: Little Akaki River (LAR) passes through Addis Ababa City, receives inorganic and organic pollutants from various sources. The objective of this study was to investigate the pollution level of LAR by selected heavy metals and evaluate sediment quality using contamination indices. Methods : sediment samples were collected from 10 stations along LAR, processed, digested and heavy metal content was analyzed using ICP-OES. Enrichment factor (EF), geo-accumulation index ( Igeo ), contamination factor (CF), pollution load index (PLI), ecological risk index (RI) were determined. Comparison was made with standard sediment qualities (SQGs) to evaluate ecological and toxicological implication. Results : the mean concentrations of heavy metals in LAR sediment were: Zn (78.96-235.2 mg / kg); Cr (2.19-440.8 mg / kg); Cd (2.09-4.16 mg / kg) and Pb (30.92-596.4 mg / kg). EF values indicated that LAR sediments were moderate to significant enrichment with Zn and Cr; moderate to very high enrichment with Pb, and very high enrichment in all sampled sites with Cd. I geo and CF values indicated that the sediments were moderate to very high contamination with toxic Cd and Pb. PLI and hierarchal cluster analysis revealed that highest pollution load occurred at sampling site (S9), in lower course of the river mainly due to anthropogenic metals inputs from industrial wastes, municipal wastewater treatment plant and agrochemical wastes; hence, its quality was deteriorated and depicted polluted site. The decreasing order of PLI in downstream was: (S9) > (S4) > (S8) > (S3)> (S6) > (S10) > (S5) > (S2)> (S7) > (S1). Pearson correlation indicated that Zn and Cd were generated from common sources of pollution. The ecological risk (RI =350.62) suggested that the contaminated LAR sediment can pose considerable ecological risks of pollution. Conclusions: The concentrations of Zn, Cr, Cd and Pb in LAR sediment were surpassed sediment quality guidelines (USEPA) and eco-toxicological guideline limit values of USEPA (TEC) and CMME (ISQGs). Thus, the contaminated sediments can occasionally pose adverse biological effects on sediment dwelling organisms and impairs the quality of river water. Thus, monitoring and addressing sediment contamination becomes necessary to sustain beneficial uses of river water for various development purposes.

In the past, studies on LAR sediments were limited to assess the concentrations and distribution of some selected heavy metals (Gizaw,2018;Melaku, 2005;Nigussie et al., 2013;Prassie et al., 2012;Tolla, 2006), and their distribution and ecological risk in the sediments of akaki catchment (Berhanu et al, 2018). Some of these studies were based on few samples with limited use of sediment quality indices. The concentration of heavy metals in LAR sediments were not adequately expressed using various sediment quality indices such as enrichment factor (EF), geoaccumulation index (Igeo), contamination factor (CF), pollution load index (PLI) and potential ecological risk index (PERI) and sediment quality standards. The sediment pH, particle size and textural composition, which influence the accumulation of heavy metals in sediment and physical status of the sediment, were not properly investigated. An attempt was no made to identify heavy metal loaded areas, and polluted and unpolluted sites. More importantly, up to now, there have been no exhaustive study on the potential ecological risk and eco-toxicological implications of toxic metals accumulation in sediments. Hence, this study is intended to fill some of these gaps and give full insight about the quality of sediment and its pollution profile.
Thus, the objectives of the study were: (i) to determine the level of selected heavy metals (Zn, Cr, Cd and Pb) in LAR sediment; (ii) to evaluate sediment contamination using various quality indices such as EF, Igeo, CF and PLI, and (iii) to assess the potential ecological and ecotoxicological risk of sediment pollution and its implications for aquatic life and river ecosystem.

Description of the study area
The study was conducted on LAR sediment. The river starts from Geferssa Reservoir which is located at foot of Entoto Mountain, and flows through Addis Ababa City, the capital city of Ethiopia, and finally joins Aba-Samuel Reservoir (Fig. 1). The city is located at 9° 2' N and 38° 42' E. LAR flows through varied altitudes that range from 2464 m.a.s.l. at around Geferssa Reservoir in the north to 2048 m.a.s.l. at the merge to Aba-Samuel Reservoir in the south. The river drains a total catchment area of about 540 km 2 (Kebede et al, 2013). In the upper catchment, the river flows through deep gorge, on rocky bed with turbulences where as in the lower catchment, it flows in a gentle slope landscape that surrounded by irrigation farm and grazing lands. There are two types of soil dominantly found around LAR: vertisol which is commonly found on top of gentle slope lands and fluvisol at bottom of slope lands and on adjacent to the Akaki River banks (Itanna et al., 2003).
LAR is one of the major rivers crossing through the city and most used for socio-economic developments. Urban and pre-urban farmers are using this pollutants loaded river water for irrigation to grow varieties of vegetables around the river banks and supply the City with fresh vegetables. Moreover, pre-urban communities are largely dependant on the river water for drinking and domestic uses, cattle drinking, washing cattle, and sand mining during dry season.
Fishing is also undertaken in the lower parts of the river course and in Aba-Samuel Reservoir. Fig.1 shows the map of the study area.

Sampling site selection and sample collection
A preliminary reconnaissance survey was undertaken to select sampling points along LAR and thus, ten sampling stations were selected based on: accessibility, nearness to point source of pollution such as industrial and municipal waste discharge point, non-point sources (irrigation farms), permanently identifiable features and purpose of the study. Accordingly, samples were collected at the following sampling sites: Gefersa Reservoir (S1-control sample) , Soramba, just after merge of Burayu and Wongate stream (S2), Kolfe Bridge (S3) , below Kera bridge in proxy to Addis Ababa City abattoir (S4), below Mekenissa bridge, where large irrigation practiced (S5), below Gofa bridge (S6), Bihire-Tsige vegetable farm area (S7), in proxy to Akaki Kalti industrial area (S8), below Gelan Guda Kebele Bridge at the middle of irrigated farm land (S9), and Aba Samuel Reservoir, at merge of LAR and the reservoir (S10) (see; Fig.1). The exact Geographical location and altitude of each sample station recoded using GPS (GARMIN, GPSMAP62st). Sediment samples were collected from these selected stations from April 6-9, 2018 between 9:00-11:00 AM, during the dry season when river flow was minimal, following the procedures described in USEPA (2001). At each sampling station, three grab sediment samples of nearly the same amount were randomly collected using clean plastic scoop (grasp sampling technique) from depth of 0-10 cm, starting from most downstream sample along straight section of the river with least disturbance. The grab samples were thoroughly mixed to form a homogenized composite sediment samples. At each station, the physical status of the sediment was noted based on OhioEPA( 2001). The sediment samples of 1500-2000 gm per site were placed in dense polyethylene bags, sealed, labeled and immediately transported to Addis Ababa University, Center for Environmental Science laboratory and kept the sediment in refrigerator at 4 C o until they were farther processed.

Determination of pH and particle size composition
Following the procedures described in Mohiuddin et al., (2010), the sediment and water were mixed at ratio of 1:2.5, thoroughly stirred for 30 min. and suspension was kept to stay overnight.
The pH of sediment was measured using pH-16, Bench pH meter, Model PHS-3CB ACC-Deg-0.01.
The particle size compositions of the sediments were determined following the procedures described in Ozkan (2012). The sediment particle sizes distribution was grouped into four textures classes on the basis of the sieve result as: Clay<0.002mm; Silt = 0.002-0.063mm; Sand= 0. 063-2mm;and Gravel >2mm (Hu et al., 2013). Each sieve result was carefully collected and weighted using electronic balance (Model: JD210-4 CE).
The percent of grain size (%) was computed using the formula described in Uwah et al (2013), which is expressed as: % Grain Size= (Sieve weight/ total weight) x 100. The sediment composition /texture/ classes were determined based on ternary diagram of flok's classification.

Pretreatment and digestion of sediment samples for heavy meal analysis
Unwanted materials such as leaves, debris, shells and coarse gravels were carefully removed and then, sediments samples were air dried at ambient room temperature. The dried sediment samples were powdered using mortar and pestle and mixed; sieved using 45µm sieve. Following the procedures described in Sekaberia et al, (2010), 1.25 g of subsample of sediment were taken from each sample and digested with 20 mL aqua regia (3:1 HCl /HNO 3 ) and then, with 5 mL H 2 O 2 in open beaker using heat plate until the digest reach near dryness. The beaker was rinsed with 10 mL of de-ionized water and the samples were farther digested with 5 mL HCl to near dryness. Finally, the digest were cooled and the beaker was rinsed with 50 mL de-ionized water and were transferred into a small flask. The concentration of heavy metals (Cd, Cr, Pb and Zn) in the sediment samples was determined using inductively coupled plasma optical emission spectrometry (ICP-OES Arcos Spectrophotometer, made in Germany).

ICP-OES operating conditions and calibration of the instrument
All the measuring conditions were configured as follows: plasma power (1400W), average plasma flow rate (6.41 L/min.), pumping speed (30 rpm ), nebulizer flow (0.73 L/ min.), nebulizer pressure (1.96 bar), Argon pressure (6.75 bar), and torch positions and measuring time adjusted.
The calibration and standardization of the spectra method was performed according to the standard protocols set for the instrument. But, standardization is undertaken daily: it is a quick procedure for correcting measuring intensities so that the correct concentrations of element is obtained using the original calibration curve. Culibration curves were prepared using 0. 06,0.11,0.17,0.56,1.12,1.68,2.24 and 2.80 mg / L of Zn;0.03,0.06,0.08,0.28,0.56,0.84,1.12,and 1.40 mg / L of Cr, Cd and Pb. Quantifications of the elements were recorded at 213. 856, 231.604, 267.716 and 220.353 nm, which correspond to the most sensitive emission wave-lengths of Zn, Cd, Cr and Pb, respectively. The sample was nebulize and the concentration was calculated on the linear graph of the standard concentration and the corresponding intensities. The cullibration curve showed linearity, R 2 of 0.999964 for Cd, 0.999874 for Cr, 0.999757 for Pb and 0.999439 for Zn. Thus, There is good correlation between concentration and emission intensities of the analysed elements.

Assessment of levels of sediment contamination
In order to assess the levels of LAR sediment contamination with heavy metals, quantities indices were employed. These include: EF, ( I geo ), CF, PLI, PERI and IR

Enrichment factor (EF)
The EF is often used to assess natural and anthropogenic sources of trace metals and status of sediment contaminations (Zhao et al, 2017). EF was determined using formula described in (Issan and Qanber, 2016) which is expressed as: Where, "C x " stands for concentration of metal in sample sediment, and "Fe" concentration of iron in a given sample sediment. The element "Fe" was taken as a normalizing element, because, its abundance in the earth's crust has not been much influenced by anthropogenic activities (Al Obaidy, et al., 2014).For geochemical background value, the world average shale value for elements were adopted from Turekian and Wedepohl (1961). According to Issa and Qanbar, (2016), the resulting EF value can be categorized into five classes These are (i) category-1: EF < 2, indicates; deficiency to minimal level of enrichment, (ii) category-2: 2 ≤ EF< 5; moderate enrichment, (iii) category-3: 5 ≤ EF < 20; significant enrichment, (iv) category-4: 20 ≤ EF < 40; very high enrichment, (v) category-5: EF ≥ 40; extremely high enrichment.

Geo-accumulation index (I geo )-
This index was employed to evaluate the magnitude of sediment contamination (Rubio et al.,2000). Geo-accumulation index was calculated (Banu et al.,2013)as follows: Where, "C n " represents the concentration of heavy metal in sample sediment, "B n " stands for the world average shale value of metal element "n", while the factor 1.5 was applied for correction of background matrix attributed to lithogenic variations (Ke et al., 2017;Martin et al., 2012).

Contamination factor (CF)
Contamination factor is commonly used to demonstrate the level of contamination of sediment by particular toxic metal at a given sample site (Manoj and Padhy, 2014). It is defined as:

Pollution load index (PLI
Where, "CF 1 , CF 2 , CFn", stands for contamination factor of each element, "n" = number of metals under study. According to Tomlinson et al, (1980) sediment is considered to be polluted,if PLI value >1; otherwise, not polluted for PLI value <1.

Potential ecological risk index (PERI) and risk index (RI)
PERI and RI were used assess an overall potential ecological risk of heavy metals in sediment,

PERI of metal element (E
Where, "C i " stands for the concentration of metal in sample sediment, "C o " represents background concentration, "T i r " toxicity response factor of single element, "E i r " potential ecological risk of each metal element under the study. According to Hakanson (1980), the PERI value indicating the severity ecological risk of sediment pollution can be grouped into five classes: class-1: PERI value < 40, indicates low pollution risk; class-2: PERI value between = 40-80, moderate risk; class-3: PERI value between = 80-160, considerable risk; class-4: PERI value between= 160-320, high risk; and class-5: PERI value >320, very high risk of pollution. Similarly, the computed value of RI are categorized into four classes as: class-1: RI value <150, low risk; class-2: RI value between150-300, moderate risk; class-3: RI value between 300-600, considerable risk; class-4: RI value >600, high risk.

Assessment of eco-toxicological effects
To evaluate toxicological adverse effects of contaminated sediments on aquatic life and ecosystem, the concentration of heavy metals in the sediments were assessed in relation (TEC and PEC) of USEPA (2002)

Statistical Analysis
Descriptive statistical analysis was run to determine the concentration of metals, EF, Igeo, CF, PLI and PERI, and the results were presented in tables and graphs using Microsoft Excel 2010.
Pearson correlation and multivariate (Hierarchal cluster analysis, (HCA) applied to evaluate sources of metal elements and to group sample sites that exhibiting similar pollution profile in downstream. Hierarchal cluster analysis was undertaken using R-software (Version 3.3.2).
Pearson correlation was determined using SPSS software,Version 20.

PH and particle composition of LAR sediment
The accumulation of heavy metals in sediment can be influenced by sediment pH and particle size composition ( Ohio EPA, 2001). The pH of LAR sediment samples range from 6.04-8.19 with mean value of 7.56. The lowest pH value (6.04) occurred at sample site (S1) which showed slightly acidic sediment may be due to decomposition of organic matters such as grass and plant   (2000) have described that varying mixture of sand, silt and clay fraction in the sediment is a result of eroding and materials transport capacity of river water. Clay Muck ( black, extremely fine, completely decomposed organic materials); cohesive; has odor, S10 Clay Muck ( black, extremely fine, completely decomposed organic materials); cohesive; has odor, Figure-3 The percentage of particle size composition of LAR sediment

3.2Concentration of heavy metals in LAR sediment
The concentration of trace metals in LAR sediment samples presented in Table- Table-2), revealed that concentrations of Cr in three sampled sites(S3, S8, and S10) and Pb at two sampled sites (S4 and S6) were exceeded the World River Sediment Average Value. This implies that the sediments at these sites were highly enriched , may be due to varying amount of local inputs, sources and other environmental factors influencing metal concentration (Qian et al., 2015). The concentrations of Cd in the sediment at all samples sites surpass the World River Sediment Average Value, exhibiting elevated concentration of Cd in LAR sediment.
To understand the level of pollution, a comparison has also been made with other studies in the country and other developing countries' and presented in In general, the comparsion made indicated that the concentrations of Cd and Pb in LAR sediments were relatively higher and require serious attention.

Enrichment factor (EF)
Enrichment factor is widely applied to quantify the abundance of metals in sediments, the levels of enrichment and to distinguish sources of metals, whether they are derived from anthropogenic or natural source (Al Obaidy et al., 2014;Kong et al., 2018;Zhao et al., 2017).

Pollution load index (PLI)
The PLI of LAR sediment was ranged between 0.95-4.46, average value 2.79. According to Tomlinson et al, (1980) interpretation of the PLI values, except at sample site (S1) which is control sample site(PLI value <1,that exhibiting not pollution), in all other sampled sites, the sediments were contaminated (PLI >1) with heavy metals.

Hierarchal Cluster Analysis (HCA)
The cluster analysis undertaken after standardization of the measured concentrations of heavy metals in the sediments using mean transformation as suggested in Rencher (2002). The hierarchical cluster analysis was undertaken following squared Euclidean distance as measures of similarity to group ten sample sites into three clusters (Fig -6). Each cluster described in terms of location, sources and concentrations of heavy metals, and pollution load (Sojka et al., 2018;Zhao et al, 2012). Cluster-1: consists of sample station(S10), (S6), (S1) and (S7). Sample site (S1) and S7 were closely resemble each other as they were mainly subjected to non-point source of metals inputs mainly from agrochemicals as they were surrounded by crop and irrigated vegetable farms, hence they had low PLI (S1=0.95; S7 =1.44). Similaly, sampling site (S10) and (S6) were also closely resemable eventhough sampling site (S10) located in lower end of LAR, while sampling site S6 located in mid-course of the river These two sampling sites receieved metal load from varied sources of inputs, but, they had similar PLI, S6 (3.04) and S10 (2.80);hence they were clustered. sample site (S5) as they had comparable geo-accumulation index for Zn and Cd. The CF for Zn at sampling site (S3) and S5 were also comparable, hence closely clustered. In general, cluster analysis sites with similar contamination levels and sample sites found in proximity to industrial site were exhibiting high PLI.

Correlation
Pearson correlation analyized at confidence limits (CL) of 95% was run to identify the association and sources of trace metals, and the result was presented in table-5. A strong positive correlation recorded between element pairs: Zn -Cd (r = 0.478), indicated that they may originated from common sources such as industries. But, low correlation observed between Zn -Cr (r = 0.260), Zn -Pb (r =0.017), and Cr -Cd( r= 0.053). Absence of strong association among trace elemnts suggested that metals didn't have common sources as their inputs are controlled by a combination of different factors such as geo-chemicals and their mixed associations (Ren et al., ,2015). Negative associations between Pb and Cr (r = -0.194); Pb and Cd (r = -0.133) suggested that these elements deposited in sediments were not associated with each other, and they were derived from diverse and different sources (Chatterjee et al., 2007) Bierhanu et al., (2018) has also reported negative correlation between Cr -Cd, and Pb -Cd for the sediment of the Akaki River catchment The concentrations of Zn, Cr and Cd were negatively correlated with sediment pH, indicating that the pH may be the main factor affecting their distribution in LAR sediments (Ke et al., 2017).

3.3Ecological risk assessment
The potential ecological risk index (PERI) and Risk index (RI) value of LAR sediment presented in Table 6. According to Hakanson (1980) classification, the PERI values for Zn , Cr and Pb were categorized in calss-1, (E i r <40), exhibiting low ecological risk. However, in sample sites(S4) and (S6) the PERI values for Pb was found to be in class-3, (E i r = 80-160) indicating considerable ecological risk of pollution to river water The ecological risk of Cd at seven sampling sites (S3, S4, S5, S6, S7, S9, and S10) found to be in class-4, (E i r =160-320), indicating high ecological risk,while at three sampling sites (S1, S2 and S8) exhibited very high (E i r >320)risk High risk of Cd is partly explained by higher geo-accumulation and high EF, and partly due to its higher toxicity response factor as compared to other elements under study (Ghaleno et al., 2015).

Eco-toxicological effects assessment.
Threshold effect concentration (TEC), probable effect concentration (PEC) and interims sediment quality guidelines (ISQGs) were applied to assess toxicological effects of sediment pollution to aquatic life. According to Burton (2002), when the concentrations of metals in sediment found to be below the threshold effect concentration (TEC), adverse biological effects are only rarely happened. The concentration metal in sediment that equals or exceeds the TEC, but lower than the probable effects concentration (PEC) indicate that biological adverse effects occasionally happened. However, the concentration of trace metals equals or exceeded the PEC reflects that adverse biological effects frequently occur (Ke et al., 2017). Especially, the concentrations of Cr and Pb which exceeded PEC limits of (CCME), and the 4 concentration of Pb that slightly surpassed CB-PEC limit indicated in MacDonald et al, (2000),

5
probable cause adverse effects to aquatic life frequently.
The sediment LAR can be categorized in to three textural classes: clay, sandy-clay-loam and 1 sandy-loam. This textural composition is mainly attributed to hydrological behavior of river flow 2 and geological setting of the river and surrounding landscape.

Abbreviations
The data used in this manuscript was originated from field sediment samples collection and 23 laboratory analysis and are available from the corresponding author upon request.   (2017)