Basic characteristics of river sediments
A considerable variation was found in the distribution of clay (81 to 48.4 g kg−1), silt (145 to 656 g kg−1), and sand (38 to 821 g kg−1) particles among sediment materials. The associated coefficient of variations (CV) was 57, 59.5, and 41%, respectively. Statistical data related to the physicochemical properties of sediments and their main elements are reported in Table 2. The variations in particle size distribution located sediment material in seven textural classes ranging from loamy sand to silty clay. The high variability in particle size distribution suggests that different sets of geogenic and anthropogenic processes are enacted in the development and distribution of sediments in the rivers. The pH and CCE ranged from 7.4 to 8.2 and 31 to 251 g kg−1, respectively, indicating the dominancy of alkaline-calcareous condition. None of the sediment samples exhibited salinity conditions (EC > 4 dS m−1) with EC in the range of 0.3 to 1.4 dS m−1. A relatively low range of OM was found in all samples ranging from 7 to 61 g kg−1 with a mean value of 19 g kg−1. This range of OM coincides with the corresponding values in regional soils47. Except for pH, other sediments properties demonstrated above 35% of CV illustrating a wide range of variability in sediments’ physicochemical properties across the study rivers.
The highest concentration among major elements was observed in SiO2, varying between 37.5 and 55.2%, with a mean percentage of 44.9%. This element followed in magnitude by Al2O3 (8.9–15.9%), CaO (5–14.3%), Fe2O3 (4.8–10%), MgO (2.4–17.2%), K2O (1.2–3.1%), Na2O (0.68–2.7%), SO3 (0.01–4.8% g kg−1) (Table 2). Considering the semi-arid climatic condition of the study region, higher levels of SiO2 and lower levels of Al2O3 may indicate that the silicate minerals forming the sediments of the area have not been subjected to severe weathering processes. Likewise, the Na2/K2O ratio was greater than 1 in the majority of sediment samples, implying an enrichment of potassium feldspar and the relatively intense weathering of Na-bearing minerals in the region48,49. The CIA value was in the range of 64.9 to 85.7% with a mean percentage of 72.9%, representing a moderate chemical weathering intensity of lithological materials (65% < CIA < 85%) in the region36.
Heavy metals concentration in sediments
Heavy metal concentrations except for a few cases shown in Fig. 2, did not vary significantly between summer and winter periods, although an overall greater concentration during summer period was evident across all rivers. In summer, sediments contained 6.4–20.2%, 7.1–10%, 9.1–22.6%, 7.2–9.9%, and 5–8.4% higher Zn, Cu, Cd, Pb, and Ni contents than winter. Greater heavy metal concentrations during summer are likely due to the seasonal variations in rivers’ water fluxes as such water recharge to the river is limited in summer, generating less mobility and greater accumulation of heavy metals in the summer. Nevertheless, contradictory observations exist in the literature50. The positive relationship between heavy metals concentration in sediments and the river flow depth and rate has been thoroughly observed in previous studies51,52. The mean annual concentrations of heavy metals except Pb, varied significantly across studied rivers (Fig. 2). The greatest mean annual Zn, Cu, Cd, and Ni concentrations were observed in Sim, Gad, Naz, and Bar rivers, respectively. The observations indicate the diversity of contaminant resource surrounding the basin rivers. Descriptive statistics of the five heavy elements concentrations in the sediments of the Urmia Lake basin rivers are presented in Table 3. The mean values of heavy metals in the sediments were in the descending order of Ni > Zn > Pb > Cu > Cd which varied largely among the sampling points. The level of Zn, Cu, Cd, Pb, and Ni varied in the ranges of 32.6–87.5, 14.2–33.3, 0.42–4.8, 14.5–69.5, and 20.1–183.5 mg kg-1, respectively, for winter, and 35.3–92.5, 15.6–35.1, 0.47–5.1, 15.5–73.1, 23.2–188.3 mg kg−1 for summer. The obtained ranges are comparable with data found in previous studies in Asia4,54,55,54.
The variability of heavy metal concentrations, characterized by CV ranged from 24 to 77% (Table 3). Given the values of CV, heavy metals were grouped into two levels of moderate (Zn and Cu with 16% < CV < 36%) and high variability (Cd, Pb, and Ni with CV > 36%)55. These results highlight that the heavy metal variabilities were in a moderate to high level in the river sediments of the Urmia Lake basin which is likely associated with the spatial distribution of heavy metal discharge units along the rivers.
Sediment properties affecting heavy metal adsorption
The Pearson’s correlation analysis was conducted to identify the sediment properties that may exert a control on sediment’s heavy metal adsorption capacity (Fig. 3). The majority of heavy metals showed strong to moderate correlations with Al2O3 (r = 0.4 to 0.8) and Fe2O3 (r = 0.4 to 0.7) content, illustrating that Al and Fe oxides considerably increase the heavy metal retention capacity of sediments. These correlations are consistent with previous observations11,56, which highlighted the substantial increase of heavy metals absorbance in presence of aluminum and iron oxides in mineral particles. Significant positive correlations were found between Cd and Pb (p < 0.01) concentrations in given riverine sediment material whereas less significant correlation coefficient found between Ni and Cu, implying an analogy in transport-accumulation mechanism and sources of these metal pairs57. Similar observations are reported by Ustaoğlu and Islam58 in sediments of a few rivers in Giresun, northeast Turkey, and Shu et al.4 in the surface sediments of tidal flats along the north Jiangsu coast, China. The lack of significant correlation between Zn on one hand and Cu, Cd, Pb, and Ni on the other hand corroborate that this element may not have a common sources and transport mechanism with other metals.
Sediment pollution indices
The summary statistics for PI, PLI, and ER pollution indices are reported in Table 4. The PI values for sediment-associated heavy metals ranged between 0.53–1.4, 0.64–1.43, 0.51–5.48, 0.7–3.31, and 0.32–2.63 for Zn, Cu, Cd, Pb, and Ni in the summer and 0.49–1.32, 0.58–1.36, 0.45–5.13, 0.66–3.14, and 0.28–2.56 in the winter, respectively. These ranges of PI fall within unpolluted (PI ≤ 1) to slightly polluted (1 < PI ≤ 2) categories for Zn, Cu, and Pb, and slightly polluted to moderately polluted (2 < PI ≤ 3) categories in the mainstream of sediment samples according to Muller59. The ANOVA revealed significant differences in PI values among the rivers (Fig. 4). MAH and SIM rivers exhibited significantly greater levels of PI-Zn than BAR and NAZ rivers. Likewise, the GAD river had significantly the greatest PI-Cu among all rivers, even though the PIs for both Zn and Cu remained below the high-risk level (PI < 2). The PI-Pb did not exceed the low-risk level for almost all rivers, except the NAZ river which showed moderate-risk of PI, especially during the summer period. The greatest values of pollution index occurred regarding the Cd and Ni metals where PI-Cd and PI-Ni reached the high-risk level in NAZ and BAR rivers, respectively (Fig. 4). The mean comparisons convey no significant differences in PI during summer and winter periods for most rivers, even though an increase from 6.2–8.2% for Zn, 5.3–10.8% for Cu, 6.8–13.3% for Cd, 5.4–6.1% for Pb, and 2.7–14.3% for Ni in the summer versus winter period was evident (Fig. 5), indicating that the change of seasons can affect the concentration of heavy metals in sediments. It is allied with the greater summer temperature which accelerates the biodegradation of organic matter and releases binding metals58,60.
The ANOVA revealed a significant difference among rivers in the PLI index (Fig. 5), even though the values did not exceed the range of the slightly polluted category (PLI < 1). The magnitude of PLI values among studied rivers were in the order of BAD > GAD > NAZ > MAH > SIM > SHA > ZOL (Fig. 5). PLI varied from 0.5 to 1.54 and 0.43 to 1.41 with a mean value of 0.91 and 0.82 in the summer and winter sediments, respectively, reflecting the range of unpolluted (PLI < 1) to the lower limits of slightly polluted (1 < PLI < 2) (Table 4). Nearly 73.5% and 26.5% of the summer sediment and 79.4% and 20.6% of the winter sediment showed PLI-no pollution and PLI- moderate pollution, respectively. The mean value of PLI was greater than 1 for BAR and GAD rivers in both summer and winter sediment samples, reflecting slight pollution in the sediments of these two rivers while the sediments of other rivers were nonpolluted.
The mean Ei varied significantly among heavy metals. The Ei values for the five heavy metals followed the descending order of Cd > Pb > Ni > Cu > Zn for both summer and winter sediments (Table 4). All metals exhibited a low ecological risk to the environment with Ei values lower than 40 except for Cd. The Ei values for Cd exceeded 80 in more than 52% of the total samples of both seasons, indicating Cd had considerable ecological risks in most sediment samples. Our results on Ei-Cd are comparable to the previous studies where studied rivers were extending along vast urban or agricultural area52,61. It can be attributed to the fact that cadmium concentration in all rivers was way beyond the reference level of 0.3 mg kg−1 found naturally in shale, indicating a strong anthropogenic intervention, particularly in NAZ and BAR rivers which flow along larger farmland areas62,63. The influx of phosphate fertilizers applied to cropland, dumping of oily materials from water vehicles, and leakage of waste leachate from manufacturing units along the rivers are among the common reasons for high Cd concentrations52,64.
The ER values throughout the study region were in the range of 24.6–189 with a mean value of 64.4 during summer and 21.6–177.1 with a mean value of 59 during winter (Table 4), suggesting low to moderate ecological risk in 97.1% and 2.9% of both summer and winter sediments, respectively. The mean ER values were significantly different across the sediments of seven rivers and showed the descending order of NAZ > BAR > GAD > MAH > SHA > SIM > ZOL (Fig. 5). Similar to PLI, the mean ER values of the summer sediments were higher than the winter period, ranging from 8.2 to 17.8% (Fig. 5). The magnitude of ER values across rivers are comparable with their corresponding sediment metal concentration during both seasons.
The toxic unit (TU) was calculated to assess the impact of sediment pollutants on the aquatic organisms45. The values of TU decreased in the order of Ni > Cd > Pb > Zn > Cu during the summer and winter periods (Table 5), which aligns well with the previous studies52. The sum of individual TU values across heavy metals was less than 4 in 85 and 88% of sediment samples during summer and winter, respectively, indicating a low toxicity level of heavy metals to benthic organisms for most samples65. BAR and GAD rivers deduced significantly greater values of ∑TU than other rivers (Fig. 5). For both summer and winter samples, the contributions of each heavy metal to ∑TU were highly variable, with the highest contribution from Ni (66.1–66.9%), followed by Cd (15.7–16.2%), Pb (8.77–8.86%), Zn (6.73–6.83%), and Cu (3.9–4%). These results surmise a considerable possibility of the adverse effect of Ni and to a lower extent Cd on freshwater organisms. The high contributions of Ni and Cd may be relevant to their low PEL values, as suggested by several authors4,66. Analogous to the other pollution indices, ∑TU was 5 to 12% higher in sediments collected during summer than winter periods, indicating that the potential acute toxicity of heavy metals was affected by seasonal changes in flow characteristics.
Human health risk assessment
Non-carcinogenic heavy metal risk
The health risk of Zn, Cu, Cd, Pb, Cd to induce non-carcinogenic diseases via ingestion, inhalation, and dermal contact was calculated for children and adults separately (Table 6). The HQ values of ingestion pathway was maximum at the range of 0.0001–0.267 for all metals and both age groups which followed by HQder (a range of 2.18E−08) and HQinh (a range of 4.35E−05–0.167). This result implies that the non-carcinogenic risk of diseases caused by the ingestion of all heavy metals from the riverine sediment of the region is much more hazardous than that by dermal contact and inhalation. These observations are analogous to the findings of the studies conducted by Proshad et al.52 and Şimşek et al.2. The HQ ranking of metals demonstrated the descending order of Pb > Ni > Cd > Cu > Zn for three exposure pathways and both age groups and seasons. In accordance with the previous studies8,58, the observed trend indicates a greater role of Pb in the incidence of non-carcinogenic risk than the other metals.
For both children and adults, the ranking of the HI values followed the order Pb > Ni > Cd > Cu > Zn during summer and winter―similar to the order observed for HQ values (Table 6). The findings confirmed that Pb and Zn had the maximum and minimum contribution to non-carcinogenic health risk, respectively. However, the HI values of all heavy metals as well as THI for children and adults were < 1, illustrating no significant non-carcinogenic risk in the analyzed sediments. In all sediment samples, children appeared to have significantly higher potential for non-carcinogenic health risk, HQ (via ingestion, inhalation, and dermal contact), HI (4.8 to 10.5 time), THI (10.2 to 10.3 time) compared with adult. The higher susceptibility of children to the health impacts of heavy metals is due to their behavioral characteristics―e.g., more outdoor activities and oral and finger practice67. Heavy metals can cause various health issues in children e.g., skeletal injuries, cardiovascular disease, or respiratory system and productivity68. For both summer and winter sediments, the mean values of THI across the seven rivers followed a descending order as BAR > NAZ > GAD > SHA > MAH > SIM > ZOL and BAR > NAZ > GAD > MAH > SIM > SHA > ZOL for children and adult, respectively (Fig. 6). The THI calculated based on summer sediments where higher than those of winter, ranging from 3 to 28% and 6.6 to 10.4% for child and adults, respectively.
Carcinogenic heavy metal risk
The carcinogenic risk potential (CR) of Cd and Pb was calculated for children and adult through ingestion (CRing), inhalation (CRinh), and dermal contact (CRder) of sediments (Table 6). For both children and adults and during both seasons, the CR of different exposure pathways were in the order of CRing > CRder > CRinh, a similar rank to those of non-carcinogenic risk (HQ and HI). During summer, the TCR (∑CR) values for Cd and Pb in children and adults ranged from 1.4E−05 to 1.53E−04 and 1.26E−06 to 1.36E−06 for Cd and from 2.64E−06 to 1.24E−05 and 2.33E−07 to 1.1E−06 for Pb, respectively. Regarding the winter, the corresponding ranges of TCR-Cd were 1.26E−05 to 1.43E−04 for children and 1.21E−06 to 1.27E−05 for adult while TCR-Pb ranged from 2.43E−06 to 1.16E−05 for children and 2.18E−07 to 1.05E−06 for adult. These results suggest that both children and adult were almost within tolerable risks (1 × 10−6 to 1 × 10−4) of carcinogenic hazards to human health. The mean value of TCR (Cd + Pb) was in rang of 4.56E−05 over winter to 5.03E−05 in summer and 4.05E−06 to 4.46E−06 for children and adult, respectively. The comparison of these data implies that the risk of TCR (Cd + Pb)-initiated cancer in children (5 case of cancer for every 1,00,000 inhabitants) was 11.2 to 11.3 times higher than in adult (4 to 5 case of for every 1,000,000 inhabitants). This distinction may also be explained by children’s behavioral patterns which encourages the propensity for their skin particularly hand-to-mouth activities67.
For both children and adults, the mean values of TCR across the sediments of seven rivers followed a descending order as NAZ > BAR > GAD > SHA > SIM > MAH > ZOL and NAZ > SIM > MAH > SHA > ZOL > BAR > GAD for Cd and Pb, respectively (Fig. 7). These results reflect a substantial variation in the spatial distribution of TCR as well as heavy metals concentration among riverine sediments of the region which is analogous to other studies in the literature61,67,69. Moreover, the mean values of TCR were significantly higher in the bulk of sediment samples during summer than those during winter (Fig. 7). The difference might be attributed to the slower flow rate during summer leading to heavy metal precipitation as well as intensified input from farmlands during summer61. Therefore, the sediment poses higher health risk during dry season, summer than the wet season, winter.
Principal component analysis for riverine sediments’ heavy metal concentrations and associated ecological and health risk indices has been provided in Fig. 8. The first and second PCs cumulatively explained 81.35% of the variability in the dataset. PC1 was associated with greater coefficients of ER, THI (children and adults), TCR-Cd (children and adults), Cd, TCR-Pb (children and adults), Pb, and PLI while PC2 was associated with greater coefficients of Ni, TU, Cu, and PLI. Pb and Cd exhibited high variability (CV > 36) and a significant positive correlation (r = 0.75, p < 0.01), indicating that Cd and Pb are most likely initiated from the same sources (e.g., discharge of urban wastewater, run-off from cropland or leachate from the waste dumping). Cd and Pb are considered two primary metals with industrial and urban origin (e.g., residues from industrial activities, traffic emission, and fuel combustion) and farm agrochemical (e.g., fertilizers and pesticides)4,52,69,70. Greater coefficients of variables within each PC indicates the collinearity among variables. The collinearity among heavy metals may indicate a mechanistic relationship between certain heavy metals that stimulate the co-occurrence of one another. Zn did not show collinearity with any of the other variables and was mainly explained by PC3 (Table 7), suggesting that Zn is likely to be mainly initiated from lithogenic origins such as natural weathering of Zn-bearing minerals or rocks as found in other studies14,71,72,73. In other studies, Zn is reportedly associated with basic minerals, particularly shales, the compounds that are predominant in the study region11. Regarding rivers, BAR and GAD showed a stronger sign of pollution with Ni and Cu metals and demonstrated a greater potential for acute toxicity, ∑TU. The lowest risk of acute toxicity was relevant to SIM, ZOL, and SHA rivers. Among all rivers, NAZ exhibited the greatest association with Pb and Pb-associated indices. The findings of the loading plot of the PCA and hierarchical cluster analysis (CA) (Fig. 8), which was performed to confirm the PCA data, showed the five heavy metals were categorized into three groups: Cd–Pb, Cu-Ni, and Zn. The findings of the CA were consistent with the PCA results. Overall, the data collected during the study are both representative and practically informative about the pollution and risk health of heavy metals in the river sediments of the Urmia Lake basin.