3.1 Physico-chemical properties of Sela river water
The results of Physico-chemical parameters for water samples were analyzed following the standard method (APHA, 2005), which are presented in Table 3. The temperature of the Sela River water ranged from 23.0 to 24.6 °C. The average temperature (24.06 °C) throughout the river was assumed to be quite warm, which could enhance the growth of microorganisms and thus develop unpleasant tastes/odors if used for drinking. Prescott et al. (Prescott L.M. et al. 1999) reported the optimal temperature for the growth and survival of mesophilic bacteria including human pathogens as (20–45 °C). Dissolved oxygen (DO) in water plays an important role in the aerobic respiration and survival of fish communities and aquatic life lived in a water system (Wetzel 2001). For example, the concentration of DO is inversely proportional to that of the heavy metals present in the water reservoir and thus increases the toxicity of heavy metals for aquatic life. Additionally, lower DO may adversely affect the life of aquatic biota at higher water temperatures (Ojekunle et al. 2016). The criterion of DO for warm and cold water biota is 5-6 and 6.5-9.5 mg/L respectively (WHO 2011). In this study, the DO of the studied water was found from 4.5 to 9.5 mg/L (Table 3). The average DO is 5.89 mg/L whereas the sites, W10 and W1 present the lowest and highest DO status respectively. Also, lower DO (< 5 mg/L) was observed at sites W3, W5, W9, and W10 (4.5 ~ 4.8 mg/L) with temperatures 23.8, 24.5, 24.4, and 24.5 °C respectively. As the solubility of DO decreases with the increasing temperature of water (Dodds &Whiles 2010), the studied river has a lower capacity to hold oxygen at most of the sites with warmer temperatures. On the other hand, higher concentrations of DO was also observed at some of the sites due to some physical processes like wind mixing or increased diffusion from the air (Dodds &Whiles 2010). The diversity of a water system and the growth of bacteria in water bodies is enormously controlled by pH. Prescott et al. (Prescott L.M. et al. 1999) reported that microorganisms habitually produce acidic/basic metabolic wastes and thus the pH of the water bodies may be changed with time. The pH of the studied river for all water samples was slightly alkaline (7.45-8.57), which were within the permissible limit range for drinking, irrigation, and aquatic life as 6.5– 8.5 (WHO 2011), 8.5 (Ayers &Westcot 1994) and 6-9 (McPherson et al. 1999) respectively. However, the pH of the site, W2 was slightly higher (8.57) than the recommended value for drinking and irrigation (Table 3) indicating the presence of free carbon dioxide or bicarbonate in the respective sample site (Wetzel 2001).
The electrical conductivity (EC) of a water system represents the degree to which the water body can conduct electricity. EC values mainly depend on the number of ions present in water, mobility of ions throughout the water reservoir, the valence of ions, and temperature. The range of EC values for Sela river water was from 3200 to 3800 µS/cm, which was about 3 to 4 times higher than the recommended standard for drinking water (1000 μS/cm, WHO, and NAFDAC) indicating frequent microbial activities in the analyzed water (USDA-NRCS 2014).
Total dissolved solids (TDS) in water bodies are the measurement of anions, i.e., CO32-, HCO3-, Cl-, SO42-, PO43-, NO3-, and cations, i.e., Ca2+, Mg2+, Na+, K+, Fe2+ and Mn2+ present (Wetzel 2001). The TDS recorded in the studied water was 2250 ~ 2690 mg/L, which was about 4 and 2 times greater than that of the recommended value for drinking water and protection of aquatic life (USDA-NRCS, 2014), respectively. As the amount of EC and TDS are related to each other, the increased EC and TDS values of the studied water suggest that the studied water samples are significantly rich in different ions, which may result from various sources, i.e., weathering of the rocks and soil minerals, sewage, agricultural run-off, wastes from the water treatment process and other industries (Ojekunle et al. 2016, Wetzel 2001).
3.2 Quantification ofmetals in water
Metal concentration in surface water samples was measured and presented in Table 4 along with the recommended and reference values. The ranges of concentrations in Sela River water were followed as: Pb (20 ~ 148 μg/L), Cd (18.2 ~ 53.6 μg/L), Cr (23.3 ~ 53.3 μg/L), Co (95.6 ~ 317.6 μg/L), Cu (15.3 ~ 37.5 μg/L), Ni (79.7 ~ 217.8 μg/L), Fe (15 ~ 114 μg/L), Mn (5.57 ~ 104.26 μg/L), Zn (17.06 ~ 37.54 μg/L), Ca (80.6 ~ 193 mg/L), Mg (220.8 ~ 721.14 mg/L), Na (28.78 ~ 354.6 mg/L) and K (83.16 ~ 223.15 mg/L). The mean concentrations for heavy metals, i.e., Pb, Cd, Cr, Co, Cu, Ni, Fe, Mn, and Zn were found to be 83.71, 30.7, 34.60, 160.4, 23.6, 148.1, 73.4, 34.98, 28.01 μg/L respectively, whereas the concentration for light metals, i.e., Ca, Mg, Na, and K was found to be 133.55, 456.72, 153.82 and 148.02 mg/L respectively. Besides, the concentration of other heavy metals, As and Hg in river water was found to be less than the detection limits (3.0 and 0.3 μg/L for As and Hg respectively). This study revealed that the mean concentrations of heavy metals Pb, Cd, Co, and Ni, were about 8, 10, 20, and 2 times higher than that of the permissible limit for potable water set by the World Health Organization (WHO 2011). Therefore, the most polluted metals in the studied river water were contaminated with the following descending order: Co > Cd > Pb > Ni respectively (Table 4). It is an alarming situation because the concentrations of Pb and Cd in this study are about 3 and 15 times higher respectively than that of reported concentrations for Pasur River, Bangladesh (Ali et al. 2018). Also, Bhuyan et al.(2019) even noted lesser Cd and Cr concentrations, i.e., 30 and 3 times respectively, in the Brahmaputra River than the present study(Bhuyan et al. 2019)(Bhuyan et al. 2019). It might be happened due to the reason that a huge amount of Pb, Cd, Ni, V, Zn come into surface water/environment from any oil accident (Mustafa 2015), which was happened in the study area in 2015. Furthermore, the Sela River water was a matter of warning for aquatic life due to the elevated concentrations of six heavy metals (out of 11 metals), i.e., Pb, Cd, Co, Cu, Ni, and Hg (Table 4). Moreover, the heavy metals Pb, Cd, Co, and Ni are considered to be toxic in this study. Fig. 2 shows the spatial distribution of toxic metals (Pb, Cd, Co, and Ni) found in Sela River water. The most polluted sites are found to be W3, W6, W7, and W8. The major reasons for water pollution of the studied river caused by toxic metals (Pb, Cd, Co, and Ni) may be anthropogenic activities, i.e., untreated industrial effluents released into the river, municipal wastes, etc in addition to the oil spelling accident.
The concentrations of light metals Ca, Mg and K in surface water were about 1.77, 9, and 12 times higher than the recommended values for drinking, with a descending order of K > Mg > Ca respectively. For irrigation purposes, the concentration of metals: Cd, Co, Mg, and K were 3.07, 3.2, 7.6, and 74 times higher than that of the recommended value (Ayers and Westcot, 1994) where the descending trend is K > Mg > Co > Cd respectively. Reconstitution into insoluble secondary minerals during weathering, the transformation from rocks, and runoff fertilizers from agricultural soils may be the major source of K in the surface water (Mallick, 2017). The elevated concentration of Mg should not impose any health hazard except unpleasant taste and corrosion occurrence (Wetzel, 2001). Also, the concentration of Ca was too low than the required value, which could result in the deprivation of soil from essential nutrients for effective irrigation.
3.3 Assessment of water quality indices
3.3.1 Water quality index (WQI)
It is well known that the water quality index (WQI) is one of the important water pollution assessment tools for representing the overall water quality status of a water system (Brown et al. 1972). Fig. 3(a) illustrates the computed WQI values of Sela River water for different purposes, i.e., drinking (blue column), irrigation (red column), protection of aquatic life (black column). The selected parameters for portable water quality were included different Physico-chemical parameters: pH, temperature, DO, EC, TDS; major cations: Ca+2, Mg+2, Na+, K+ and as well as a number of heavy metals: Pb, Cd, Cr, Co, Cu, Ni, As, Hg, Fe, Mn, and Zn respectively. While the parameters like temperature and DO were omitted for irrigation use. For aquatic life criteria, the excluded parameters were Ca+2, Mg+2, Na+, K+, and Fe. The results show that the ranges of computed WQI values were 391.80 ~ 933.74, 641.38 ~ 4353.91, and 225.50 ~ 286.33 concerning drinking water, irrigation water, and aquatic life protection respectively (Tables S2, S3, and S4). The average WQI values for drinking water, irrigation, and aquatic life were found to be 613.08, 2798.59, and 254.29 respectively (3(a)). According to the criteria of Index 1 in Table 2, higher WQI values (WQI > 100) obtained at all sample sites suggests that the water is unsuitable for any intended use. Among all purposes, the lowest water quality was found for irrigation uses, especially at sites W3, W5, W6, W8, and W9 respectively (Fig. 3(a)). This is because out of 18 parameters, TDS and K has a significant contribution to inferior water quality in the case of irrigation (Table 3, Table 4, and Tables S2, S3, S4). This study also revealed that the concentrations of Ca and Na were found to be within the recommended limit for irrigation, the water quality was further evaluated by irrigation indices, i.e., alkalinity hazard (SAR) and Na% in the latter section (Fig. 3(b)).
3.3.2 Alkalinity Hazard (SAR)
SAR is known to be an assessment tool for determining the suitability of water for irrigation use by measuring the number of alkali metals like Na+, Ca2+, and Mg2+ ions (Kawo &Karuppannan 2018, Subramani et al. 2005). The sample points of the studied river water are categorized according to SAR formula (Table 2) and shown in Fig. 3(d) and Table S5. The range of SAR at all sites was observed from 0.25 to 4.29 meq/L with an average value of 1.49 meq/L, whereas the highest and lowest SAR was found for sites W10 and W9 respectively. According to the criteria defined by SAR (Table 2, Index 4), 100% of the studied water samples obtaining SAR < 10 were excellent water quality types for irrigation uses (Table S5). It was reported that the hydraulic conductivity of soil irrigated by water with high contents of Na, K, Ca and Mg showed less negative K effects than those for Na as well as less positive Mg effects than those for Ca. As the deleterious effect of K is considered to be about one-third of that for Na, it is necessary to contain a greater Mg concentration than that of Ca have the same beneficial effect (Smith et al. 2015). Unlike WQI, the calculated SAR values suggest that the Sela River water is suitable for irrigation usage.
3.3.3 Percentage of sodium (Na %)
Na% is also known to be an assessment tool for evaluating the suitability of water by measuring the amount of alkali metal like Na+ ion present in a freshwater system if used for irrigation (Kawo &Karuppannan 2018). The high amount of Na+ ion can cause reduced soil permeability and limited circulation of air/water (Saleh et al. 1999). The percentage of (Na%) in Sele River water for this study was calculated (Table S5) and was graphically shown in Fig. 3(b). This study revealed that Na% ranging from 8.65 to 40.67 indicates the level of water quality from excellent to permissible for irrigation purposes (Table 2, Index 5). The average Na% (19.22) of Sela River water suggests that the study river water is categorized as excellent quality. It was observed that 60%, 30% (W2, W5, and W8), and 10% (W10) of the river water samples fall into excellent, good, and permissible category respectively (Fig. 3(b) and Table S5). As the majority of the samples in the study area are of excellent quality water type, the studied river water is considered to be perfectly fit for irrigation.
3.3.4 Metal Pollution Index (MPI)
Another effective water pollution assessment tool for measuring the contamination status of multiple metals retained in a water system is the metal pollution index (MPI) (Department of civil engineering 1970). In this study, MPI is calculated for 11 heavy metals (Pb, Cd, Cr, Co, Cu, Ni, As, Hg, Fe, Mn, and Zn) according to the formula of Index 2 (Table 2). The results are presented in Fig. 3(c) and Table S3. The calculated MPI showed a different degree of metal pollution in the studied river water for different uses (Fig. 3(c)). The order of MPI values for drinking purposes: Co > Cd > Pb > Ca > Ni > Cr > Mn > Fe > As > Hg > Cu > Zn, for irrigation purpose: Co > Cd > Ni >Cr > Mn > Cu > Hg > As > Pb > Fe > Zn and for aquatic life protection: Co > Pb > Cd > Cu > Ni > Hg > Mn > Cr > Zn > Fe > As respectively. According to MPI assessment criteria (Table 2: Index 2), the metals Co (MPI:31.45), Cd (MPI: 14.56), and Pb (12.02) are assumed to be responsible for heavier pollution (MPI >10) if the water is intended to use for drinking and thus marked as toxic metals due to their adverse effects on human health (Department of civil engineering 1970). Besides, lighter pollution was noticed for Ni (MPI: 2.66). However, the metals like Cr, Cu, As, Hg, Fe, Mn, and Zn had no effect on the pollution for drinking purposes. In the case of irrigation, no heavier pollution is found for metals (Fig. 3(c) and Table S6) although moderately and lightly polluted metals are Co (MPI: 5.03) and Cd (MPI: 4.37) respectively. For the protection of aquatic life (Fig. 3(c) and Table S6), the Sela River water is considered to be heavily polluted due to the presence of toxic metals like Pb (MPI: 14.14) and Co (MPI: 31.45), which may result in a serious threat to the survival of biota living in the studied river. Furthermore, the metals that exhibited lighter pollution effects on aquatic life are Cd, Cu, Ni, Hg, and Mn. The other metals like Cr, As, Fe, Zn are of good quality metals in the studied river water, which do not cause any threat to aquatic life.
3.3.5 Metal quality index (MI)
The metal quality index (MI) is a spatial measurement of metals found at different sites throughout the study area (Caeiro et al. 2005). MI has been calculated by Index 3 (Table 2) and the results were presented in Fig. 3(d). The mean MI values of Sela River water for drinking, irrigation, and aquatic life were found to be 42.18, 7.87, and 42.61 respectively (Table S5). As MI value >1 is a threshold of warning (Gülfem Bakan et al. 2010), the obtained higher MI values (> 1) of the river water at all sample sites indicate a “very low-quality water” for all purposes. According to another criterion for MI values (MI > 6.0, Index 3, Table 2), all the sites through the studied river water are considered to be seriously polluted with metals. In the case of drinking, the sites W1, W3, W6, W8, and W9 are marked as highly polluted zones compared to other sites (Fig. 3d). In addition to the highly polluted zones found for drinking, the sites, W1 and W5 have also greater metal pollution for irrigation use. For the survival of aquatic biota, the sites W3, W6, W8, and W9 are also more affected sites than others. The computed MI values agreed well with that of the WQI values discussed in the previous section, where the greater pollution was also observed at sites W3, W6, W8, and W9 for different utilizations.
3.4 Quantification of heavy metals in sediments
Table 5 presents the concentrations of measured and earth’s crust for 15 metals in Sela River sediments. The ranges of concentrations in Sela River sediments were as follows: Pb (19 ~ 31.61 mg/kg), Cd (0.75 ~ 2.35 mg/kg), Cr (32.11 ~ 49.45 mg/kg), Cu (29.08 ~ 36.93 mg/kg), As (2.23 ~ 4.09 mg/kg), Fe (20240 ~ 40481 mg/kg), Mn (312.83 ~ 572.85 mg/kg), Zn (63.25 ~ 84.08 mg/kg), Ca (326 ~ 4265 mg/kg), Mg (11646 ~ 26550 mg/kg), Na (1485 ~ 75896 mg/kg) and K (3913 ~ 6677 mg/kg). The mean concentrations of Pb, Cd, Cr, Co, Cu, As, Fe, Mn, Zn, Ca, Mg and were 26.58, 1.37, 40.11, 33.70, 3.30, 30255, 476.64, 74.4, 1507.8, 14311, 11269, and 4998 mg/kg respectively in the sediment samples. The metals that exceeded the reference values are Pb, Cd, Cr, Cu, As, Zn, and Mg (Table 5). It was observed that the concentrations of Pb, Cd, Cr, Cu, As, Zn and Mg were found to be 1.56, 13.43, 1.15, 2.36, 1.35, 1.43, and 1.05 times higher respectively than the reference values. Consequently, the concentration order of the elevated metals decrease as follows: Cd > Cu > Pb > Zn > As > Cr > Mg. Thus, it is revealed that Cd is the most dominant heavy metal among the quantized elevated metals in studied sediment samples.
In addition, the most polluted sites contaminated by Cd were the sites from 7 to 11, where the range of Cd concentrations was 2.11 ~ 2.35 mg/kg indicating around 20 times greater elevation of Cd concentration in sediments. Other metals like Fe, Mn, Ca, Na, and K of the studied sediment samples was below the earth’s crust value (Table 5). Besides, the least dominant metal measured in sediment samples was Ca (Table 5), which indicated that Ca is more soluble in the river water rather than present in the sediment. It is important to notice that the elevation in concentrations for Pb is lower but for Cd is higher in sediments than that of surface water due to different mineralogy and absorption capacity of metals into sediments (Haque et al. 2004).
Moreover, the elevated metals within the sampling area are supposed to be deposited from discharge effluent or municipal waste rich in those respected metals. As noticed in Table 5, the metal pollution status of the Sela River is assumed to be not only similar to the near Pasur River but also greater than the Brahmaputra River (Bhuyan et al. 2019). Therefore, it is highly expected to further investigate the pollution status of the study area by using different pollution indices (Table 2).
3.5 Assessment of sediment quality indices
3.5.1 Contamination factor (Cf), Degree of contamination (Cd), and Modified Degree of contamination (mCd):
The contamination factor (Cf), the simplest and traditional approaches for evaluating the contamination status by single metals quantified from sediment analysis (Hakanson 1980). However, the Cf for the studied metals in the Sela River was computed using the equation of Index 6 (Table 2) and depicted in Fig. 4(a). It was found that the average order of Cf values for the metals was Cd (13.42) > Cu (2.36) > As (1.65) > Pb (1.56) > Zn (1.43) > Cr (1.15) > Mg (1.06) > Fe (0.98) > Mn (0.90) > Na (0.44) > K > (0.17) respectively (Table S7). As noticed, the estimated average Cf value of all studied sites was highest for Cd and lowest for Ca (Fig. 4(a)). According to the assessment criteria of Cf (Table 2, Index 6), the Sela river has very high Cf values for Cd, moderate values for Pb, Cr, Cu, Zn, Mg, As, and low values Mn, Fe, Ca, Na, K. It is important to note that Cd gives 100% (S1 ~ S20) as very high contamination due to its average Cf> 6. The most polluted sediment contaminated by Cd (23.04) is in site S8 (Table S7). Also, the Cd profile showed an increasing trend in the middle of the sediment sites, S7 ~ S11, where Cd may not have mobilized and thus get accumulated (Mustafa et al., 2015). Whereas for Mn, Fe, Ca, Na, and K, the average Cf < 1 was recorded suggesting low contamination in Sela River, which might be due to fewer activities triggering the accumulation of the respective metals in the sediment. However, different anthropogenic activities contribute to moderate contamination by Mn, Fe, and Na at 25%, 45%, and 10% sample sites respectively (Table S7). On the contrary, the other metals Pb, Cr, Cu, Zn, Mg, and As represents moderate contamination with average Cf values in the range of 1 to 3 (Fig. 4(a) and Table S7). Meanwhile, Pb, Cu, Zn, and As exhibited 100% as moderately contaminated. For Cr, only 10% of sites (S9 and S11) were low contaminated over the whole profiles and the remaining stations were moderately contaminated (Table S2). The higher Cf in Cr was found at sites, S7 and S8 regarding moderate contamination level. Besides, Mg showed nine sites (S3 ~ S10, S20) that exhibit 45% low contaminations whereas the remaining are accounted as moderate enrichment by anthropogenic or natural input. Therefore, the Cf values of the studied heavy metals in sediments can effectively measure the contamination status for Sela River and thus indicated that the low contaminants of the Sela River sediments are Mn, Fe, Ca, Na and K whereas the moderate contaminants are Pb, Cr, Cu, Zn, Mg and As and the very high contaminant is Cd.
Fig. 4(b) and Table S7 depicted the degree of contamination (Cd) of all the studied sediments for Sela River by summing the above-depicted Cf values (Table 2, Index 7). The estimated Cd values ranged from 19.51 to 37.55 and according to the classification of Index 7 (Table 2), the Cd values indicated from considerable (S1 ~ S6, S12, S17 ~ S20) to a very high (S7 ~ S11) degree of contamination in the Sela River sediments. The maximum degree of contamination was found for site S11 (Cd: 37.55). The average Cd value of all sites is 25.18, which is above 24 denoting a very high degree of contamination of Sela river sediments. Additionally, the Cf and Cd values (Table S7) demonstrated that the higher accumulation of Cadmium (Cd) metal in the sediments may contribute to the most contaminated sites (S7 ~ S11, S13 ~ S16) in the studied area.
To figure out the effect of the number of contaminants on the pollution status of the studied area, the modified degree of contamination (mCd) (Abrahim &Parker 2002) was calculated and presented in Fig. 4(c) and Table S4. As mCd is the modified version of Cd, all sediment sites finally indicate a moderate to low degree of sediment contamination (Table 2, Index 8: 2.0 ~ < 4). However, the contamination pattern of all sites obtained from the mCd profile agreed well with that of the Cd. For example, like Cd values, the five most deteriorated sediment sites in the Sela river were also observed to be S7 ~ S11 and S13 ~ S16 having the highest mCd values, well above the threshold for a moderate degree of contamination (2 ~ < 4).
3.5.2 Pollution Load Index (PLI)
Fig. 4(d) shows the distribution of PLI values for all sample sites in the Sela River. As observed in Fig. 4(d), the PLI values varying from 1.34-16.57 (Table S7) were significantly higher than 1 for 100% of the sites demonstrating metal pollution of all sites to some extent according to the classification of PLI (Table 2, Index 9). The sample sites, S11 ~ S16 and S18 were marked as the most polluted sites having PLI values ranged from 14.63 ~ 16.57. On the other hand, the less but not the least polluted sites were S4, S5, and S9 having PLI values of 1.59, 1.46, and 1.34 respectively. Comparatively, higher sediment pollution was found for the rest of the sites. The higher PLI values of Sela River sediments might be originated from different anthropogenic inputs, i.e., non-treatment industrial waste discharge, agricultural runoff, etc. polluted by the studied heavy metals. Besides, the contribution of individual metals to the PLI values can be evaluated from the Cf values. As notified from Fig. 4(a), the Cf of Cd went beyond the maximum limit (Cf > 6) denoting very high concentration, whereas the Cf of Pb, Cr, Cu, Zn, Mg, and As was in the range of moderate contamination. Consequently, the metal Cd was mainly responsible for major pollution, and the other metals like Pb, Cr, Cu, Zn, Mg, As may cause moderate pollution in the Sela river sediments. On the contrary, the Cf values of Mn, Fe, Ca, Na, and K are low (< 1) for most of the sample sites specifying them as non-pollutants in Sela river sediments. Thus, the increased rate of non-treated industrial waste discharged into the Sela river may cause probable environmental pollution of the river sediments especially with dangerous heavy metals like Cd, Pb, Cr, Cu, Zn, and As. It was reported that Pasur River’s sediments also faced metal contamination (Table 5) due to the elevated concentrations of heavy metals like Pb, Cd, Cr, As (Ali et al. 2018). It is important to note that the difference in indices may result due to the difference in sensitivity of these indices towards the sediment pollutants (Kayode O. Adebowale et al. 2009, M. S. Praveena et al. 2007).
3.5.3 Enrichment factors (EF)
The computation of EF was performed by geochemical normalization of the measured concentrations of heavy metal to a "conservative" element like Fe (Rule 1986) for reducing metal variability due to the similar geochemistry of Fe to that of many trace metals and uniform natural concentration (Daskalakis &O'Connor 1995). To further evaluate the anomalous metal contributions in sediments of the studied river, the enrichment factors (EF) of studied metals were computed and the results are shown in Fig. 5(a).
According to the classification of EF in Table 2, most of the metals are significantly enriched in the studied sediments except for Ca and K (Fig. 5(a)). The EF values for Pb range from 1.23 to 2.3, Cd from 5.99 to 29.98, Cr from 0.93 to 1.76, Mn from 0.58 to 1.31, Cu from 1.88 to 3.76, Zn from 1.21 to 2.08, Ca from 0.01 to 0.18, Mg from 0.85 to 2.14, Na from 0.06 to 3.2, K from 0.15 to 0.21, As from 1.2 to 2.29 (Table S8). Overall, the order of average EF values for the metals is Cd (14.18) > Cu (2.46) > As (1.71) > Pb (1.61) > Zn (1.48) > Cr (1.19) > Mg (1.10) > Mn (0.94) > Na (0.44) > K (0.18) > Ca (0.06) (Table S8). The concentration ranges of metals Ca and K exhibited EF < 1 (Table 2, Index 10) suggesting their abundance in sediments from mainly crustal materials or natural weathering processes and as a result no significant contribution from non-crustal sources. Although the EFs (average) of Na and Mn < 1.0, those metals still showed enrichments at some sites from non-crustal sources. For example, minor and moderate Na enrichments were at sites S12 (1.09) and S11 (3.2) respectively whereas minor Mn enrichments were at sites S3 (1.06), S5 (1.31), S7 (1.00), S8 (1.27), S17 (1.00) and S20 (1.00).
Among the anthropogenic-contributing metals, Pb, Cr, Mn, Cu, Zn, Mg, Na, and As showed evidence of minor modification (EF, 1.0 ~ < 3) in their enrichments although Cu was moderately enriched at only two sites, S5 (3.76) and S8 (3.23). The only severely enriched metal (EF, 10 ~ < 25) in sediments was found to be Cd (average EF, 14.18), which yield vibrant enrichments (moderately to very severe) along the studied area as shown in Fig. 5(a). As observed, most of the sites suffered from severe Cd enrichment except the sites S12, S14, S17 ~ S20 having moderate Cd modification. It is a matter of great concern that the most enriched site in Cd was S8 (EF, 29.98), which was marked for very severe enrichment (EF, 25 ~ <50). Consequently, Cd is assumed to be the most toxic sediment-bound metal in this study. In addition, it was noticed from Fig. 5(a) that the sites S8 ~ S11 are comparatively more polluted than the other sites as they are located at the inner part of the Sela River. It is well known that the geological formations and soil leaching could be partially responsible for the occurrence of heavy metals in sediments, mainly at the inner part of the river (Ohta et al. 2007, Simeonov et al. 2003). As the metals showing enrichment (EF >1.0) are considered to be indicative of human influences, the enrichment of sediments with toxic elements (Cu, Cd) may also be due to the streams, uncontrolled wastewater discharges, and long transportation phenomena (Nolting et al. 1999). The major cause for observation of a high amount of Cd might be due to the deposition of toxic metals from the oil spill that occurred at Sela River on 9th December 2014 (BBC 2014, Krishnendu Mukherjee &Chakrabarty. 2014). The other minor contributions might be various anthropogenic intrusions in the coastal region of Bangladesh, i.e., automobile industries, ship breaking activities (Raknuzzaman et al. 2016).
3.5.4 Geoaccumulation index (Igeo)
The environmental assessment based geoaccumulation index (Igeo) of Muller’s expression (Table 2, Index 11) for the studied area was calculated (Table S9) and was graphically shown in Fig. 5(b). The estimated Igeo values for the metals of environmental interest are (-0.20 ~ 0.02) for Pb, (0.69 ~ 1.19) for Cd, (-0.21 ~ -0.03) for Cr, (-0.40 ~ -0.14) for Mn, (0.13 ~ 0.24) for Cu, (-0.36 ~ -0.0) for Fe, (-0.09 ~ 0.03) for Zn, (-2.13 ~ -1.03) for Ca, (-0.24 ~ 0.12) for Mg, (-1.41 ~ 0.29) for Na, (-1.04 ~ -0.10) for K and (-0.13 ~ 0.73) for As. The average Igeo values tabulated in Table S9 are Pb (-0.06), Cd (0.92), Cr (-0.12), Mn (-0.23), Cu (0.20), Fe (-0.19), Zn (-0.02), Ca (-1.57), Mg (-0.16), Na (-0.85), K (-0.94), As (0.04). As noticed, the pollution intensity order of average Igeo values for the metals is as follows: Cd (0.92) > Cu (0.2) > As (0.04) > Zn (-0.02) > Pb (-0.06) > Cr (-0.12) > Mg (-0.16) > Fe (-0.19) > Mn (-0.23) > Na (-0.85) > K (-0.94) > Ca (-1.57). Among the analyzed parameters of sediments in the studied river, only three metals Cd, Cu and As having average Igeo > 0 were classified as uncontaminated to moderately contaminated metals. Whereas, the other metals, i.e., Pb, Cr, Mg, Fe, Mn, Na, K, Ca were below the practically uncontaminated status (average Igeo < 0) suggesting that the area is not significantly polluted by these metals. Depending on each metal and sample sites, some deviations in Igeo values were also observed. For example, 25% of the sample sites (S7 ~ S11, Igeo: 0.71 to 1.19) reflects a ‘moderately Cd-contamination’ level (Igeo: 1 ~ 2). Additionally, 25 % of the sites (S2 ~ S4, S9, and S19) were considered to be practically uncontaminated (Igeo < 0) by As. Besides, the Igeo value > 0 was observed at some sample sites in case of metals having negative average Igeo values, i.e., Pb (S7 = 0.005, S8 = 0, S10 = 0.001, S11 = 0.02, S13 = 0.02, S15 = 0.0003); Zn (S13 = 0.03, S14 = 0.03, S15 = 0.02, S17 = 0.003, S19 = 0.001, S20 = 0.02); Mg (S11 = 0.12) and Na (S11 = 0.29). Therefore, it was assumed from these deviated Igeo values that the studied metals of Sela river sediments may be possibly prone to more gradual contamination if proper measures are not taken into consideration in the future.
3.5.5 Sediment quality guideline (SQG)
Sediment quality guideline value (SQGV) is the criteria of tolerable concentrations of analyzed metals in sediments, which eventually predict the toxicity for protecting the aquatic biota living in that area (Burton 2002, McCready et al. 2006). The two commonly used SQG approaches reported are effective range low/median (ERL/ERM) trigger values and threshold/probable effect level (TEL/PEL) (Burton 2002). The recommended SQGVs (ERL/TEL) represents the threshold value below, which infrequent adverse effects would be found upon living biota in or near sediments. On the other hand, the upper SQGV (ERM/PEL) represents the high probability of effect on living organisms of the studied area (Burton 2002).
In this study, the distribution of analyzed eight metals (Pb, Cd, Cr, Mn, Cu, Fe, Zn, and As) bound to sediments throughout the study area was achieved according to SQGV (Table 6) and shown in Fig. 6(a) and 6(b). As noticed, metals like Pb, Zn, and As were distributed in 100% of the sediment samples that did not exceed the ERL/TEL values indicating rare toxic effects to the river biota. However, the concentration of Cr was < ERL for 100% samples but < TEL for 15% samples. In addition, 45 and 50% of sediment samples had Cd and Cu concentrations respectively, in the range of ERL ~ ERM indicating occasional adverse effects to the river biota. Furthermore, the concentrations of metals Cd, Cr, Mn, Cu, and Fe falling in the range of TEL ~ PEL were observed for 100%, 85%, 70%, 15%, and 95% samples respectively, where occasional adverse biological effects are predicted. At last, but not least only 5% of samples had Fe distribution less than PEL. However, it is well known that the concentration of Iron (Fe) may not be considered as a serious threat to the aquatic organism and human beings' health (Burton 2002). Moreover, the concentration of metals like Cd, Cr, Mn, Cu, and Fe measured at different sites is a matter of future concern as the distribution of these metals is already in the middle range (ERL-ERM/TEL-PEL) of toxicity toward living organisms. If proper steps are not taken in time, the concentrations of these heavy metals may exceed the limit of ERM/PEL, which would present a significant risk to the lifestyle of biota lived near the metal-bound sediments.
For evaluating the combined toxicity level of different metals retained in sediments and identifying the affected sediment sites, the more realistic and effective SQG measurement, which is mean SQG quotients, i.e., Effective range median mean quotient (ERMQ), Probable effect level mean quotient (PELQ) were also used in this study (Table 2 Index:12] (Chapman &Mann 1999). The results are presented in Fig. 6(c) and Table S10. According to Table 2 (Index 12), all the sediment sites of the Sela River showed medium-low toxicity (ERMQ: 0.11 ~ 0.5; PELQ: 0.11 ~ 1.5) to aquatic biota as the average ERMQ and PELQ is 0.12 (30% toxicity) and 0.36 (25% toxicity) respectively. Therefore, the computed medium-low toxic SQG quotients suggested that it is a proper time for routine monitoring of heavy metals pollution nearby Sela River and assuring the survival of aquatic ecology in the future so that the living organisms of that area could be safe from further exposure of toxic metals.
3.5.6 Environmental Toxicity Quotient
Environmental Toxicity Quotient, another important ecological risk index for assessing sediment toxicity, was calculated in this study and presented in Fig. 6(d) and Table S10. According to the ETQ based toxicity level (Table 2, Index 13), the average ETQ value (268.68) confirmed that the sediments of the Sela River are very highly toxic to the aquatic biota. It is important to note that extremely high toxicity would be observed at the sites S12 ~ S14, S17, and S20. Besides, the sediment sites, S2, and S12 had the lowest and highest toxicity respectively (Table S10). It should be mentioned that the sites S12 ~ S14 were also found to be highly contaminated in PLI computation in previous sections (Fig. 4(d)).
3.6 Source identification
3.6.1 Principal component analysis (PCA)
Principal component analysis (PCA) was performed on the river water and sediment quality data using Varimax rotation with Kaiser Normalization, which was used in this study to elucidate the observed relationship of cluster variables in simple ways (Mertler 2005). The calculated factor loadings, together with cumulative percentage, and percentages of variance explained by each factor are shown in Table 7. Table 7 shows that three factors were extracted for both water and sediment quality data sets based on eigenvalues more than 1 (Fig S1), which represented 87.42% and 81.25% of the total variance in the study area. For water quality data, it was observed that PC1 in the data sets explained 57.62% of the total variance, and it was positively loaded with Cr, Co, Ni, Fe, Zn (anthropogenic sources) in association with Mg and K (geogenic sources) in Sela River water (Bodrud-Doza et al., 2016). It might be happened due to the reasons that these potential heavy metals (Cr, Co, Ni, Fe, and Zn) can be assimilated in river water through metal activities/industrial effluents (Saha 2017). Several industries such as ceramics, brick, pottery, petroleum refinery, shipbreaking industries are located in the study area, which is responsible for these heavy metals’ presence in Sela River water (Rahman 2019). PC2 represented 17.86% of the variance and loaded heavily on Pb and Cd, which also indicating the anthropogenic sources of pollution by Pb and Cd (Rahman 2014), as well as hundreds of boats and small ships, are running on the river every day. PC3 explained 11.95% of the variance, showed high (> 0.95) positive loading (Table 7(a)) on Cu and Mn, that is indicating the sources of pollution by Cu and Mn is anthropogenic in the Sela River. Therefore, these three components play a critical role in explaining metal contamination in the surface water of the Sela River, and subsequently, the graphic representation of the three components is shown in Fig. 7(a).
On the other hand, the sediment quality data revealed that three principal components have been taken into consideration and accounted for the total variation of data in order with a cumulative variance of 81.26% in Sela River (Table 7(b)). It was observed that PC1 was accounted for 44.99% of the total variance and had high Cr, Mn, Cu, and Zn in association with the crustal metals (K, Fe, and As), which indicates that anthropogenic toxic metals (Cr, Mn, Cu, and Zn) coming into sediment and trapped with the crustal metals. This PCA data is consistent with the reported result in the literature mentioning that about 43.34% of the total variance explained by the first factor (PC1) is correlated with Ni, Cr, Pb, and Co, which represents the anthropogenic sources originated from domestic/municipal wastewaters and vehicle emissions (Ustaoğlua 2020). PC2 is strongly loaded with Na and Mg and accounted for 23.98% of the total variance, which indicates that abundant of Na and Mg might be lithologic sources (Dietrich 2020). Subsequently, PC3 accounted for 12.33% of the total variance and highly loaded with Cd (> 0.95) and moderately loaded with Pb (< 0.69), which suggested that the sediment of Sela River was loaded with Pb and Cd for anthropogenic activities (Wang 2020) in the study area. However, the relationship between the contaminants based on the first three principal components is illustrated in Fig. 7(b). It can be seen from this Figure that the sediment data associated with different metal sources are indeed separated, so categorization into different pollution sources is meaningful.
3.6.2 Correlation matrix analysis
The relationships between variables for both water and sediment quality data were analyzed by the Pearson correlation matrix, and the results are presented in Table 8. It should be mentioned here that the high degree of correlation between element concentrations suggested either a common or a similar geochemical behavior origin (Mertler 2005). From Table 8, it could be observed that there was a strong correlation of Cd with Ni (r = 0.877), Zn (r = 0.731), Ca (r = 0.696), Mg (r = 0.686) and K (r = 0.699) in water sample, which indicating the sources of metals are anthropogenic and geogenic (Edet 2002). Subsequently, Cr showed the same type of positive correlation with Ni, Zn, Fe, Ca, Mg, and K (Table 8(a)), which was consistent with the principal component analysis (Table 7). On the other hand, Zn showed positive and strong correlation with Cd (r = 0.731), Cr (r = 0.784), Ni (r = 0.849), and Fe (r = 0.895). It may be attributed to anthropogenic sources of metals in surface water (Saha 2017) of the Sela River, which was also observed in PC analysis (Fig. 7(a)). The correlation among the other elements in surface water samples could be found in Table 7(a). However, sediment-water quality data also used to explore the relationship between the elements and presented in Table 7(b). This study revealed that geogenic element: K was positively correlated with Pb, Mn, Cu, Fe, Zn and As; suggesting that the anthropogenic sources originated from domestic/municipal/industrial effluents that come into surface water and eventually trapped into the river sediment in the study area. On the other hand, strong positive correlation was observed between Pb (main emission source of industries + vehicles) and Cd (r = 0.501), Cu (r = 0.469), Fe (r = 0.491), Zn (r = 0.675) and As (r = 0.613). It should be mentioned here that this analysis revealed a significant positive relationship for each metal at 99% and 95% confidence level, which indicated that a significant relationship exists between water bodies and surface sediment in aquatic systems of the Sela River. Therefore, the elemental association may signify that each paired elements have an identical or common source in the aquatic systems (Singh 2002).