Accumulation and ecotoxicological risk assessment of heavy metals in surface sediments of the Olt River, Romania

Heavy metal pollution of river freshwater environments currently raises significant concerns due to the toxic effects and the fact that heavy metal behavior is not fully understood. This study assessed the contamination level of eight heavy metals and trace elements (Cr, Ni, Cu, Zn, As, Pb, Cd, and Hg) in the surface sediments of 19 sites in 2018 during four periods (March, May, June, and October) in Olt River sediments. Multivariate statistical techniques were used, namely, one-way ANOVA, person product-moment correlation analysis, principal component analysis, hierarchical cluster analysis, and sediment quality indicators such as the contamination factor and pollution load index. The results demonstrated higher contents of Ni, Cu, Zn, As, Pb, Cd, and Hg, with values that were over 2.46, 4.40, 1.15, 8.28, 1.10, 1.53, and 3.71 times more, respectively, compared with the national quality standards for sediments. We observed a positive significant statistical correlation (p < 0.001) in March between elevation and Pb, Ni, Cu, Cr, and Zn and a negative correlation between Pb and elevation (p = 0.08). Intermetal associations were observed only in March, indicating a relationship with river discharge from spring. The PCA sustained mainly anthropogenic sources of heavy metals, which were also identified through correlation and cluster analyses. We noted significant differences between the Cr and Pb population means and variances (p < 0.001) for the data measured in March, May, June, and October. The contamination factor indicated that the pollution level of heavy metals was high and significant for As at 15 of the 19 sites. The pollution load index showed that over 89% of the sites were polluted by metals to various degrees during the four periods investigated. Our results improve the knowledge of anthropogenic versus natural origins of heavy metals in river surface sediments, which is extremely important in assessing environmental and human health risks and beneficial for decision-maker outcomes for national freshwater management plans.


Material and methods
The surface sediment samples were collected from 19 sites located in the middle and lower reaches of the Olt River during four field trips in March (T1), May (T2), June (T3), and October (T4) 2018 (Supplementary  Table 1). Eight significant HMs and trace elements (Zn, Cr, Cu, Ni, Pb, As, Cd, and Hg) were analyzed. The data are presented as the mean (µ), range of elements, standard deviation (sd), coefficient of variation (cv), and interquartile range (q3-q1) to estimate the trend and pattern of variation among sampling sites and periods. One-way ANOVA and post hoc Bonferroni tests were used to assess differences in seasonal concentrations of metals in sediments. In addition, the Levene test for equal variance was evaluated. Finally, multivariate analyses, including correlation analysis (CA), principal component analysis (PCA), hierarchical cluster analysis (HCA), and linear regression (LR), were used to determine possible intercorrelation between elements and the element group. PCA was used to assess contamination sources predicted by correlation analysis using varimax rotation with Kaiser normalization. Factor loadings for HMs in sediments were extracted based on engine values higher than 1. HCA was performed to find similarities among metals using Euclidian distance. The statistical analyses were carried out using SPSS 27 (IBM SPSS Statistics, IBM Corp., USA) and R v4.1.0 27 .

Results and discussions
Sediment quality. Overall, the average concentration of the HMs had a decreasing order of As > Zn > Pb > Ni > Cr > Cu > Cd > Hg. HMs become more toxic to the Earth spheres (biosphere, atmosphere, hydrosphere, and lithosphere), and the effects are more evident, especially on human health 28 . Only a limited number of governments have shown concern for minimizing HMs in ecosystems 29 . The World Health Organization (Joint FAO/ WHO Expert Committee on Food Additives, JECFA) 30 has recommended strict limits for exposure to, or the lifetime of, consumer products that contain a certain degree of contamination. Several HMs are monitored and banned (Hg, Pb, Cd, As), as their level of toxicity is very high 31,32 . In comparison, other elements (e.g., Cr and Zn) do not have very strict regulations 30 . Recent studies demonstrate that constant exposure, even at lower levels, increases the risk of respiratory tract cancer 33 and cardiovascular or neurodegenerative diseases 34,35 .
Basic statistics for the investigated elements (Cr, Ni, Cu, Zn, As, Pb, Cd, and Hg) in sediments across the middle and lower Olt River showed the following mean concentration levels (µ): 25.19, 42.20, 24.13, 73.78 49 , and Cabezo Rajao, Spain (314.7 mg/kg) 50 . Mining activities appear to be the highest source of As pollution, and As enters river ecosystems by wind, flash floods in the hot season, from water discharges in spring after snowmelt, or from accidental releases 51 . However, the mean concentration of As reported in two studies from the Pra Basin of Ghana, a region with gold mines, showed reduced contamination 38,52 . Both studies showed comparable results in two different sampling periods (0.15 mg/kg and 0.714 mg/kg), indicating no exceptional events causing point source pollution and explaining comparable values in time (low variability). However, in the Olt River case, the difference in the concentration level for samples collected in 2018 and 2019 was significant within the same sampling areas, ranging from a maximum value of 240.14 mg/k in the first year to 8.35 mg/kg in the second year 19 . We noted the highest concentration of As (mg/kg) at site #11 (T1 = 232, T2 = 240.14, T3 = 228.7, T4 = 231.4); for the same location in 2019, the values ranged between 0.238 and 7.79 mg/kg. It is worth mentioning that the As maximum value did not exceed 29.4 mg/kg at sites #3, #8, #12, or #19; at the other sites, the concentration level was high, comparable with that at site #11. Extreme amounts of HMs are deposited on the bottom river layers and remobilized under different environmental conditions 53 .

Spatial and temporal distribution of HMs in sediments.
We noted that the HM concentration fluctuated from upstream sites (#1) until the Olt River overflowed into the Danube (#12). The spatial variation in contaminant levels for each sample site and period is illustrated in Fig. 1. The coefficients of variation (cv) for Cr, Ni, Cu, Zn, As, Pb, Cd, and Hg had respective ranges of 66-90%, 41-66%, 70-135%, 44-63%, 47-48%, 35-79%, 48-65%, and 81-208%, indicating that these elements had uneven spatial and temporal distributions. A high HM concentration range, corroborated with a high coefficient of variation amplitude, showed substantial anthropogenic influence 54 . Statistically, the significance of a low coefficient of variation in As, for example, suggests a relatively stable variation in HM contamination levels in sediment samples and not necessarily a natural background. The analyzed dataset showed significant differences between the Olt River sampling sites from the middle reaches (#1-#10) and the lower reaches (#11-#22) only for Cr and Zn (p < 0.01). This result indicates there are more HMs in the middle reaches than at sites in the lower reaches. The higher level of HM contamination can be explained through various input sources of anthropically induced pollution, such as dams from power plants that create specific environments by restricting the river flow velocity 13 . River hydrodynamics regulate desorption and absorption reactions and can also be responsible for high contaminant level variability 55 . Other authors have attributed the higher concentration of some metals (e.g., Hg and Pb) to the Râmnicu Vâlcea chemical chlor-alkali plant 23 . This rationale does not explain similar or high amounts of pollutants from the sites www.nature.com/scientificreports/ located far upstream, starting with site #1 (Fig. 1). The higher values measured in most investigated locations can result after spring floods and extreme events that transport and deposit HMs released from industrial and municipal waste, mining, and agricultural practices 13 . Chromium is a mutagenic, highly toxic, and carcinogenic element and it has been extensively studied because of its capacity to migrate and transform in surface sediments after long-term contamination. Chromium naturally occurs in two oxidation states: Cr(VI), including the highly soluble oxyanion, and Cr(III), which is less toxic and less mobile. Cr(VI) is involved in redox reactions, absorption-desorption, and oxidation forms, and it control Cr toxicity and mobility 56,57 . In our case, the range of maximum concentrations for chromium measured in each period oscillated between 42.0 mg/kg (T3) and 98.9 mg/kg (T2), and extreme values were achieved at sites #10 (T1), #5 (T2) and #11 (T3, T4). In addition, we observed significant differences between the Cr population means (F = 7.61, p < 0.0001) and population variances (F = 4.97, p < 0.003) of the four periods investigated. Chromium originates from untreated industrial emissions and domestic wastewater discharge released into the Olt River. The mean HM statistics for each sampling period showed higher values in T1 (22.94 mg/kg) and T2 (43.57 mg/ kg) than in T3 (15.14 mg/kg) and T4 (17.84 mg/kg).
Even for Pb, significant differences between populational means (F = 29.93, p < 0.0001) and populational variance (F = 4.94, p = 0.003) were calculated. The Pb contaminant level varied from 28.1 mg/kg (T2) to 86.6 mg/ kg (T3, T4), with maximum values at sites #3 (T1), #14 (T2), and #13 (T3, T4). The sampling periods can be associated with spring (March, May), summer (June), and autumn (October). Each period is characterized by specific climatic conditions, including temperature, precipitation, and evapotranspiration, which can induce a specific regime of organic matter, pH, and the association between metals 58 . The population means and variance of the four periods investigated did not vary significantly in the cases of Ni, Cu, Zn, As, Cd, or Hg, for which the contaminant level (mg/kg) ranges were 66.93-80. 63 . The Ni concentration was higher than the national safety guideline of 35 mg/kg at 45 randomly distributed sites along with spatial position and sampling periods, indicating natural and human-induced sources. The natural presence of Ni in the river environment can be associated with watering soils and rocks and atmospheric deposition. A strong correlation was presented with acid-volatile sulfide (AVS), total organic carbon (TOC), Fe, and Mn exchange capacity (CEC) 59 . The Ni geochemistry is not fully understood, and according to Rinklebe and Shaheen 60 , serious attention must be paid to this essential nutrient that, at high concentrations, can adversely affect organisms. Coprecipitation with Fe and Mn (hydr)oxides can be associated with Ni's potential to mobilize during changes in pH, increasing phase partitioning, and toxicity in water sediments 61 . Ni's high background content in the Carpathians was conditioned by river hydrological characteristics, and it is most often bound in immobile phases 62 .
Copper exceeded the prescribed limits at sites #2 (T1) and #12 (T3 and T4), indicating a low contamination risk. The interquartile ranges (q3-q1) in T1 = 26.94 mg/kg, T2 = 30.13 mg/kg, T3 = 24.15 mg/kg, and T4 = 26.2 mg/kg demonstrated low midspread variability regardless of season. Copper deposits naturally in hot waters associated with volcanism, explaining the high concentrations in sites #4 and #5 in all sampling periods (T1-T4). Waters in the Călimănești region (#4, #5) can be cold (< 20 °C), hypothermal (20-34 °C), thermal (42-43 °C), or geothermal (> 87 °C) and can contain sulfurous, chlorinated, sodium, calcium and magnesium. Copper in the Danube River (#12) can be associated with anthropogenic sources, given that the values there were the highest measured in this study. This trace element is characterized by a high degree of remobilization from water to sediments. Under extreme events (floods), it can precipitate with sulfides in an anoxic environment under extreme events (floods) that induce resuspension 63 .
In our study, cadmium, a highly toxic metal responsible for environmental and human severe diseases, was at the upper limit in sites #4 (T3, T4), #10 (T1, T3, T4), #14 (T1, T3, T4), and #20, #21 (T2, T3, T4). The mean values (mg/kg with standard deviation) were T1 = 0.42 ± 0.27, T2 = 0.49 ± 0.32, T3 = 0.53 ± 0.26 and T4 = 0.53 ± 0.28, with a minimum interquartile range (q3-q1) in T1 (0.23) and maximum in T4 (0.55). Recent literature indicates a strong relationship between tourism and cadmium occurrence, which affects the whole globe, except Antarctica 64 , and this pattern can explain cadmium's presence in areas with intense tourism (Olănești, Căciulata, Ocnele Mari). Cadmium presence in the Olt River due to anthropogenic activities can be associated with industrial and municipal waste. Point sources such as industrial platforms (e.g., helicopter producer IAR Ghimbav) were frequently reported as sources of cadmium release into the water ecosystem of the Olt River basin 19 . Even so, Ni-Cd batteries, fossil fuel combustion, phosphate fertilizer, and waste incineration are responsible for Cd discharges in the environment 65 . The UNEP 66 indicates that natural processes (e.g., dust storms, volcanic activities, climate change, erosion, and wildfire) are responsible for significantly higher Cd deposition than are anthropic processes. According to this study, Romania is included on a red list of high Cd contamination hotspots worldwide.

Principal component analysis and hierarchical component analysis. River inputs and histori-
cal deposition upstream of dams are essential sources of toxic HM sediment contamination 19 . Remobilization between water and the bottom layers of sediments is regulated by water discharges of the Olt River and tributaries. PCA was applied to identify and interpret the metal relationships and discuss similar contamination sources 67 . The PCA results based on the correlation matrix for the four periods investigated are presented in However, the loadings of the principal components in May were positive (As, Hg, Cd, Pb) and moderately positive (Cu, Ni, Cr, Zn). Comparable associations were observed for the loadings of principal components in June and October as positive (Cr, Cd, As), moderately positive (Cu, Pb, Ni, Zn), and moderately negative (Hg). Ni, Cu, Zn, Cr, and Pb are mainly derived from industrial effluents, domestic sewage, and runoff from mining and agriculture. However, they can be partially attributed to a natural geogenic origin from sediment accumulation and parental riverbeds. Cd, Hg, and As have loading factors in the first and second principal components, indicating that there is a mixed source that is less geogenic and more related to anthropogenic activities, a fact sustained even by correlation analysis and coefficients of variation. These metals are observed in different positions and with loadings of various lengths, suggesting one or more pollution sources. High loadings in PC2 were accumulated in this group associated with pesticides and phosphate fertilizers used in agricultural practices. Since toxicity and bioavailability are correlated with geochemical form and concentration, these elements can pose a potential hazard to the aquatic Olt River environment.
Divergent from most reported studies in which PCA and HCA were convergent here, HCA separated four groups according to the Ward cluster method and absolute correlation distance using similarity as the criterion (Fig. 3). Thus, the first group was associated Cr, Ni, and Zn, the second was associated As and Cd, the third was associated Cu and Pb, and the fourth was distinctly associated with Hg. These results suggested various levels of geogenic and anthropogenic origins of the related metals. The sampling location in the middle Olt River, an intramountain area that is mainly touristic due to its thermal waters, was related to one group with the highest mean site contamination level. Downstream, the Olchim Râmnicu Vâlcea chlor-alkali plant was included in a second group with many points indicative of anthropogenic activities associated with industries and municipal waste. The third group was associated the sampling location with a less distinct spatial pattern but with less contamination. Finally, the fourth group included two sites downstream from the Râmnicu Vâlcea municipality that were represented by artificial lakes with a similar pollution level.  www.nature.com/scientificreports/ (r = 0.48) in October (p < 0.05) (Fig. 4). It is worth noting that there was no significant relationship in June between any HM and elevation. Additionally, we observed that Pb was correlated negatively with elevation in both June (r = − 0.41, p = 0.08) and October (r = − 0.40, p = 0.08). Other research indicated a relationship between the decreasing trend of HM content and increases in elevation and river flow rate during the spring 68 (Fig. 4). The correlations between elements indicated similar levels of HM contaminants, similar released sources of pollution, associated dependence during their transport, and remobilization in river sediments 69    Pollution indices. Geoaccumulation index is used to assess the presence and intensity of anthropogenic contaminant accumulation in sediments. I geo indicated a moderate level of Cu in June and October of 2018 (#12), when a reference geochemical crustal background was used. The I geo calculated for Ni revealed sediment samples ranging from nonpolluted to moderately polluted, regardless of site and season. Except at locations #4, #9, #13, and #20, the As content was classified as moderate to heavy pollution in 2018 based on the Müller ranking 74 . Moderate contamination with Hg was observed at locations #10 and #11. The mean and standard deviation I geo values presenting a negative result indicated no contamination for Zn, Cr, Cu, Ni, Pb, Cd, and Hg (− 2.0 ± 1.3; − 3.4 ± 2.0; − 2.0 ± 1.8; − 0.55 ± 0.9; − 2.1 ± 1.3; − 2.2 ± 1.7; and − 1.6 ± 1.2, respectively) and were positive for As (1.4 ± 1.4) in 2018. The I geo for As indicated a change from moderate to strong pollution and 79% of samples ranged between 1.93 and 2.46. However, no significant differences between site locations with high-level metal contamination were detected. Pollution was observed for all sampling periods, suggesting that temperature increase rather than accidental discharge was the main factor associated with the increasing concentration. PI index generally represents the toxicity status of elemental pollution in sediments and varies according to the crustal background used. Heavy pollution (PI > 3) was calculated for all sites except #3, #8, #12, and #19 in 2018. The mean PI values had ranges of 4.1 ± 1.6, 4.2 ± 1.8, 4.1 ± 1.7, and 4.1 ± 1.8 for March, May, June, and October, respectively. According to the RI values, an intermediate ecological risk was observed at locations #10 and #11 in May (216, 228), June (210,195), and October (228, 206). Our study has illustrated more than inefficient use of participatory practices in the past regarding the spatiotemporal alteration of water bodies. It has also indicated possible negative influences via environmental drivers. Mining activities, agricultural practices, sewage discharge, and industrial discharge have exacerbated natural freshwater contamination, especially in west-central and southern Romania.
The PLI was investigated to compare the integrated pollution level at different sampling sites, as HMs can significantly vary in various sediment samples. The PLI was assessed as the n th root of the multiplication of contamination factor, where values > 1 qualified the site as polluted 75 . We found the PLI values to have ranges of 0.06-5.73 (cv = 40%) in March, 0.49-6.04 (cv = 44%) in May, 0.33-5.7 (cv = 43%) in June, and 0. 35-5.8 (cv = 45%) in October, indicating that over 89% of the sites were polluted by HMs to various degrees (Fig. 5). The most polluted site was #11 (Drăgășani accumulation lake), located downstream from the Râmnicu Vâlcea municipality. Iordache et al. conducted a study near the Râmnicu Vâlcea upstream chlor-alkali plant and reported a low level of CF for most sites with Ni, Cu, Pb, and Cu 76 ; the authors also calculated the PLI, which indicated no HM enrichment in surface sediments. However, surveys conducted in 2015 located between Călimănești (#4) and downstream Râmnicu Vâlcea, near our site #22, indicated moderate and considerable contamination of Cu, Hg, Zn, Ni, and Cr and serious levels of pollution at 7 of 9 sites 26 . Furthermore, in 2019, the investigated sites were considerably or highly contaminated with As, according to the CF. According to the PLI eight locations from 22 that were analyzed were relatively highly polluted 25 . These results show enrichment in the last ten years of Olt River surface sediments. Thus, according to reviewed reports, there has been moderate pollution found downstream from the chlor-alkali plant over the past ten years. At present, we found high surface sediment contamination in both the middle and the lower reaches of the Olt River.

Conclusions
HM concentrations in southern Romania and the Olt River surface sediment samples were investigated to assess their spatial distribution, seasonality, and contamination levels. The results showed that the concentrations of Ni, Cu, Zn, Pb, Cd, and Hg in surface sediments were generally higher than their respective national background values-up to 8.28 times more in the case of As. The spatial variability pattern had similar trends for all HMs, indicating increasing concentrations in the middle compared to the lower sites. The statistical tests showed a statistically significant difference between the mean and variance only for Cr and Pb when investigating the temporal variability between site measurements. We observed a strong association between HMs in March and May and a strong relationship in March between elevation and Pb, Ni, Cu, Cr, and Zn, demonstrating that river velocity accumulated with spring discharges are conditioning metal levels in the spring. The HMs Cd, Hg, and As had mixed sources that were less geogenic and more from anthropogenic activities. The CF indicated that the Olt River was highly contaminated with As at 15 of 19 sites. A low degree of contamination was also observed in various locations for the investigated elements. PI analysis indicated that the Olt River sediments were highly polluted (PI > 3), and the PLI showed varying degrees of pollution in more than 89% of the sites surveyed, increasing the authorities' need for action. The I geo indicated that Zn, Cr, Cu, Ni, Pb, Cd, and Hg were at a pollution-free level and that As was present at levels ranging from unpolluted to moderately polluted. A higher involvement by national authorities, including monitoring measures to control contamination in industrial areas and municipal waste discharges, is needed to improve the Olt River environment.