Composition and physico-chemical properties of bottom sediments in the southern part of the Bratsk Reservoir (Russia)

The paper presents the results of studies of bottom sediments taken from the southern part of the Bratsk Reservoir. The following analyses have been conducted: trace element analysis, particle-size analysis as well as chemical analysis of water, hydrochloric acid and alkaline extracts for 18 samples of the bottom sediments. The granulometric analysis has identified the predominance of fine silt and silty-clayed sediments. The data on the content of trace elements in the bottom sediments of the Bratsk Reservoir is presented in comparison with their content in the natural environment of the Baikal region; the anomaly ratio was used to characterize the excess for trace elements. The chemical analysis of extracts has showed “saline contamination” of mud (silt), high concentration of carbonates in it, as well as the presence of mobile (free) forms of aluminum oxide. In this research, an attempt was made by using a correlation analysis to evaluate the impact of various physical and chemical characteristics of the bottom sediments, such as the content of clay fraction, organic carbon, carbonates, and water-soluble salts on the accumulation of trace elements.

The ecological state of the entire reservoir geosystem, as well as its individual components (water mass, bio-organisms, aquatic vegetation, sediments (bottom deposits)), depends on a number of natural and anthropogenic factors. As a result of contamination, the trace element content in all components of the geosystem can vary considerably [1][2][3][4][5] . A dangerous trend in this case is the accumulation of toxic elements, such as heavy metals [6][7][8][9][10][11][12][13][14][15][16][17] . Most of them are biochemically active, are not subject to biodegradation and have the ability to accumulate intensively in such "comfortable" environment as bottom sediments 18,19 .
The accumulation of trace elements in bottom sediments occurs along with complex physical and chemical mechanisms of absorption, which depend on the peculiarities of the composition, structure and properties of the sediment, as well as on the properties of the absorbed compounds 20,21 . The accumulation might occur through interaction with the solid phase, following the mechanisms of ion exchange and complex formation with humic acids, adsorption by clay and other minerals, iron and manganese hydroxides, etc. 22,23 . Therefore, the accumulation and distribution of trace elements is greatly influenced by different physical and chemical characteristics of the sediment, such as granulometric composition (particle-size distribution), particularly the content of silt (clay) fraction, the content of organic carbon, carbonates, water-soluble salts, and amorphous oxides 24 . It is known that these components are involved in the formation of structural bonds of dispersive soils and, depending on the composition and amount, are able to weaken or strengthen these bonds, thus affecting the deformation and strength properties 25 . These components also have a significant effect on the physical and chemical activity of the sediments. For example, elevated concentrations of organic carbon lead to a decrease in the content of inorganic carbon and hydrogen index and, as a consequence, to the change of oxidation-reduction potential, leading to the release of metals from the sediments into the aquatic environment 26 . Due to the leaching of water-soluble compounds, the density of the texture of subsoils, their cohesive properties and permeability change. In the process of dissolution and leaching of salts, the composition of absorbed cations, which significantly affect the properties of sediments, may change 27 .

Sampling and Analysis
Sediment samples were collected from 15 different locations of Bratsk Reservoir in 2010 and 2011 using a sediment core sampler operated from a boat. The sampler was submerged on a steel line using a hand-operated winch. In total 18 samples from a sediment layer of a thickness of about 10 cm were collected. Sampling depth varied from 2.5 m to 28.0 m.
The selected samples were dried to air-dry state in vitro, and then each sample was divided into 2 parts. After the homogenization process, one part was packed in a plastic bag and sent to ACTLABS (Activation Laboratories Ltd.) in Canada to be analyzed for the trace element contents. Laboratory analyses in 2010 were conducted using TD-ICP (total digestion-inductively coupled plasma) method with respect to Cu, Mo, Ni, Pb and Zn. A 0.25 g sample aliquot is digested with HClO4-HNO3-HCl-HF at 200 °C to fuming and is then diluted with aqua regia 34,35 . This leach is partial for magnetite, chromite, barite and other spinels and potentially massive sulphides. The other trace element contents were determined using the INAA (instrumental neutron activation analysis) method. A 1 g aliquot is encapsulated in a polyethylene vial and irradiated with flux wires and an internal standard (1 for 11 samples) at a thermal neutron flux of 7 × 10 12 n cm −2 s −1 . After a 7-day decay to allow Na-24 to decay the samples are counted on a high purity Ge detector with resolution of better than 1.7 keV for the 1332 keV Co-60 photopeak. Analysis results are given in ppm 34,35 .
In 2011 analyses were conducted using the Aqua Regia -ICP/MS method. A 0.5 g sample is digested in aqua regia at 90 °C in a microprocessor controlled digestion block for 2 hours. Digested samples are diluted and analyzed by Perkin Elmer Sciex ELAN 6000, 6100 or 9000 ICP/MS 34,35 .
Another part of each sample was analyzed in the Analytical Centre of the Institute of the Earth Crust (Irkutsk, Russia) using a special method developed for analysing clay and loessial soils 25 . Clay and loess deposits are complex natural formations consisting of the primary particles and aggregates. The presence of aggregates affects the results of particle size (granulometric) analysis, as it reduces the content of fine particles and increases the content of coarse ones. This distorts the correct understanding of the massive material, the content of certain particles in it and its classification. The particle size analysis shows the limiting (extreme) dispersion, but in order to obtain information about the natural dispersion, it is necessary to know the number of secondary micro-aggregate particles. This method allows obtaining such information without isolating the micro-aggregate composition or performing additional micro-aggregate analysis of the soil. Namely, the granulometric analysis was made by pipette method, using 3 processes of specimen preparation 35 : (1) aggregate process (shaking up of the specimen in water causes the destruction of coarse water-unstable aggregates); (2) the semi-disperse process (boiling of triturated specimen in presence of ammonia), and (3) disperse preparation (boiling of triturated specimen with sodium pyrophosphate leads to maximum destruction of aggregates). The semi-dispersed technique is a standard one and it is recommended for determining the particle size distribution of clay soils and for naming the sample within the framework of the classification.
By water, muriatic and alkaline extracts the components and structural bonds such as water soluble salts, carbonates, amorphous sesqui-oxides and mobile Al 2 O 3 forms were determined. To a certain extent, water extract imitates the material composition of the soil solution, so it can help get an insight into the content and composition of the most mobile soil components that are more capable of migration. The principle of the water extraction method is the short-term (3 min.) interaction of a soil sample with distilled water, at the ratio 1:5. At the same time, the dissolution of readily-soluble salts occurs, while medium-and low-soluble salts are partially dissolve. Analysis of the extract begins with determining the carbonate and total alkalinity, as well as the pH of the extract, then the amount of dry residue is determined (the content of water-soluble salts is determined as the sum of anions and cations) after evaporation in a water bath. When analysing the hydrochloric acid extract, the contents of R 2 O 3 , SiO 2 , CaO, MgO, Fe 2 O 3 , SO 3 as well as the loss on ignition and the value of the mineral insoluble residue are obtained. Then, using special coefficients, oxides are recalculated for the content of carbonate salts (CaCO 3 , MgCO 3 , FeCO 3 ) and their sum is determined, which characterizes the total carbonate content of the soil. Analysis of the alkaline extract allows us to determine the content of mobile Al 2 O 3 forms 25,36 .
The humus content was determined according to I.V. Tyurin's method based on the oxidation of carbon of organic matter by excess potassium bichromate. The oxidation occurs in a highly acidic medium and is accompanied by the reduction of hexavalent chromium to the trivalent. Excess bichromate in the solution after oxidation of organic matter is titrated with a solution of Mohr's salt. Based on the difference of potassium bichromatebefore and after oxidation -the content of organic carbon in the soil is found 37 .

Results and Discussion
The material composition of the bottom sediments at the site under study refers to the terrigenous genetic type. The terrigene sediments are formed from detrital or pelitic (lutaceous) material, originating from the banks, due to various exogenous processes 38 . As a result of mechanical differentiation under the action of waves, currents, underwater gravitational processes, etc., coarse deposits usually prevail in the area of submerged shoals: pebbles, gravel, and sand. With distance from the shoreline, they are replaced by fine sands and silts.
Granulometric analysis (particle-size distribution) of bottom sediments. The granulometric analysis of the sediments from the south of the Bratsk Reservoir has determioned the predominance of fine silt and silty-clayed muds. The sediment received its name according to the particle-size distribution of the sample with semi-dispersed preparation. Figure 2 shows a batch of the ten most representative samples from the groups of finely-aleuritic (B-12-11, B-8-11, B-7-11, B-3-11, B-2-11) and aleuritic-clayed (B-11-11, B-6-11, B-5-11, B-4-11, B-5-10) silt 35 . The particle-size distribution of each sample is presented in two versions of sample preparation: a -semi-disperse, and b -dispersed ones. The graphs demonstrate how the ratio of particle contents by fractions significantly changes towards the increased number of more micro-fine particles with maximum destruction of aggregates in the case of dispersed sample preparation.
www.nature.com/scientificreports www.nature.com/scientificreports/ In fine silt muds, the content of 0.05-0.01 mm fraction is mainly up to 50-70% (Fig. 2). The fractions of 0.25-0.05 mm and 0.010-0.002 mm have approximately the same ratio. The total content of silt with particle size of <0.01 mm does not exceed 40%. Aleuritic-clayey silts generally contain no more than 50% of 0.05-0.01 mm particle size (but when the dispersed sample preparation is used, their content decreases sharply several times in favor of fine-grained particles <0.001 mm in size), and more than 40% of clay particles <0.01 mm in size. The content of the sand fraction (>0.25 mm) in almost all samples does not exceed 1%.
trace element composition of bottom sediments. The average, minimum and maximum trace element contents in the sediments from the south of the Bratsk Reservoir are presented in Table 1 22,35,[38][39][40][41][42] . As a rule, environmental assessment of the levels of content of chemical elements in the natural environments is carried out by comparing their actual concentrations versus regulatory (standardized) parameters. However, to date there are no regulatory requirements for the maximum permissible trace element concentration for the sediments in Russia. Therefore, in order to assess the quality of bottom sediments in the Bratsk Reservoir, Table 1 shows the values of percentage abundance (clarks) of trace elements in the earth's crust, according to Rudnick and Gao 42 , as well as the average trace element contents in the bedrock of the Baikal region 22,39 and abyssal silt of Lake Baikal 40,41 (for certain elements). The studies show that for such trace elements as Co, Cr, Cu, Mn, Ni, Pb, Th, V, Zn, the exceeding in the geochemical background of bedrock in the region is observed. It may be noted that the average trace element contents in the deep (abyssal) silt of Lake Baikal for the most part can be checked against the maximum contents in the bottom sediments of the Bratsk Reservoir or exceed them.
Thus, in order to determine the excesses of the aforementioned trace elements (Co, Cr, Cu, Mn, Ni, Pb, Th, V, Zn), anomaly ratio (AR) was used, which is calculated as the ratio of the content of a chemical element in the studied natural object to the geochemical background 43 . So, the highest excess of the background (among the aforementioned) is recorded for Co, Ni and Zn (AR is 2.0-2.18). The average content of other elements exceeds the geochemical background less than 1.5 times.
Structure-forming components. According to the results of chemical analysis of aqueous extract, silts in the Bratsk Reservoir are characterized by varying degrees of salinity: from medium saline (Water-soluble salt contents, S ws % range within 0.329-0.944%) 35 -to extremely highly saline (S ws is up to 2.080%); the average value is 1.042% (Fig. 3). The salinity quality is predominantly of carbonate-sulfate type; it is caused by a high degree of water solubility of the chloride salts, which practically do not accumulate in the sediments.
The total content of carbonates in the sediments varies widely, from 15.08 to 31.32% 35 , on average it is 24.36% (Fig. 4). CaCO 3 and MgCO 3 as prevail in the content of carbonate salts. According to Ryashchenko 38 , the constant presence of CaCO 3 dominated carbonates is a regional feature of quaternary sediments (loessial and clayey soils) in the south of Eastern Siberia, as well as salinity, however, at a weaker level (S ws varies within 0.06-0.42%)  Table 2 presents Pearson's correlation coefficients between the contents of Co, Cr, Cu, Mn, Ni, Pb, Th, V, Zn (trace elements for which the excess of geochemical background of bedrock in the region was marked), such physical and chemical parameters as the content of pelitic (clayey) particles (<0.01 mm) in a  www.nature.com/scientificreports www.nature.com/scientificreports/ sample with different variants of sample preparation, as well as the content of calcium and magnesium carbonates, organic carbon and water soluble salts.

Influence of physical and chemical characteristics of bottom sediments on the distribution of trace elements.
Many studies have confirmed that particle-size distribution of sediments affects the transport and accumulation of high-density (heavy) metals in soils 45,46 . Huang et al. have found that Cu, Ni and Mn showed a positive correlation with clay proportions because clay particles are more prone to adsorb heavy metals 47 . In this study, the correlation analysis has also showed that the highest value in the accumulation of trace elements in the sediments is their dispersion, i.e., the content of particles <0.01 mm in size. Co, Cr, Cu, Ni, Pb and Zn demonstrated the most significant positive correlation in terms of the content of pelitic (clayey) particles in an aggregated state (r ranges from 0.597 (Pb) to 0.849 (Ni)). When semi-dispersed and dispersed sample preparation is used, i.e., when aggregates are destroyed, this correlation is weakened for most elements, as indicated by a decrease in the correlation coefficient. Thus, when semi-dispersed and dispersed sample preparation is used, only Co, Ni and Pb show the average degree of correlation (0.5 < r < 0.69) with the content of pelitic particles, while Cr, Cu and Zn show the moderate correlation (0.3 < r < 0.49).
The influence of organic matter on the accumulation of heavy metals in the sediments is insignificant. This is indicated by the correlation coefficients of the content of organic carbon and heavy metals in the sediments, varying from −0.313 to 0.459 (the exception is an inverse correlation of С org with Th (r = −0.686)). The weak correlation between concentrations of heavy metals in the sediments containing organic matter in most cases suggests the secondary role of humic substance in the process of accumulation and migration of metals, and it is the evidence of predominantly terrigenous nature of the formation of trace-element composition of the bottom sediments in the Bratsk Reservoir.
The territory under study is intensively used for agricultural purposes. Farms tend to use large quantity of fertilizers (both mineral and organic ones) and pesticides to promote the acceleration of agricultural production. As numerous studies show, this practice often leads to the accumulation of trace such elements as Cr, Ni, Cu, Zn, As, Cd and Pb in soils 45,48,49 . Soil erosion, being an important factor in the redistribution of soil in the reservoir, is probably the main mechanism by which the studied trace elements enter the bottom sediments of the reservoir 50 .
The correlation analysis has showed a average negative correlation of the content of calcium carbonate CaCO 3 with Ni (r = −0.524), as well as a moderate negative correlation with such trace elements as Co, Cr, Cu, Pb and Zn. The content of magnesium carbonate MgCO 3 moderately positively correlates with such trace elements as Cr, Cu, Pb, V and Zn. Besides, there is a moderate negative correlation between the content of water soluble salts and S ws and the content of Mn, Pb and V in the bottom sediments. Interestingly, metals are correlated with each other. A particularly strong positive correlation (r > 0.70) is observed between Co, Cr, Cu, Ni, Pb and Zn. Such a pronounced correlation might indicate that these trace elements have the same source of influent into the reservoir 51 . Mn is the least active in correlation with other trace elements (its correlation is generally very weak, r < 0.19). For Th and V, the correlation with other trace elements mainly varies from very low to moderate.
Thus, the correlation relations confirm the dependencies -both well-known and specific for the reservoir under study -between the analysed features of the bottom sediments. First of all, it is the relationship between particle size and the content of toxic metals, which can be attributed to the conditions of sedimentation of sediments in the reservoir. The accumulation of trace elements in finely dispersed sediments can be viewed as a manifestation of the purifying role, especially in the zone of a variable backwater of the reservoir, but also as an increasing threat of environmental pollution. Significant correlation coefficients of the content of most toxic trace elements in the samples under study suggest the complex nature of the influent of pollutants into bottom deposits. As shown by the correlation analysis, humic substance has the secondary role in the process of accumulation and migration of metals, and it is the evidence of predominantly terrigenous nature of the formation of trace-element composition of the bottom sediments in the Bratsk Reservoir. 5. The content of heavy metals in the reservoir sediments is controlled mainly through the dispersion of bottom deposits -the content of pelitic (lutaceous) particles (<0.01 mm) in the aggregated state. No significant correlation between the content of carbonates and water-soluble salts and the accumulation of trace elements under consideration has been identified. There is a strong direct relationship between many heavy metals, which indicates the same source of their influent into the reservoir.  Table 2. Correlation matrix of the content of heavy metals in bottom sediments of the Bratsk Reservoir.
Explanations: Agg -the content of pelitic particles (<0.01 mm) in samples with aggregate preparation; SD -the same, with semi-dispersed preparation; D -the same, with dispersed preparation.