ASSESSMENT OF OVERBANK SEDIMENTATION RATES AND ASSOCIATED POLLUTANT TRANSPORT WITHIN THE SEVERNYA DVINA RIVER BASIN

It is now widely recognized that significant proportion of pollutants in rivers is transported with suspended sediments. This paper presents a combination of reconstruction of recent floodplain sedimentation rates based on detailed description of sediment sections and 137 Cs stratigraphy with geochemical analysis of overbank deposits at selected sites on rivers of the Severnaya Dvina River basin. Overbank sedimentation rates for sections sampled on floodplains of the Severnaya Dvina and Vychegda Rivers are characterized by noticeable decrease from ≈1.5–4.0 cm year –1 between 1954 and 1963 to <1.0 cm year –1 at present. It can be explained by the natural evolution of the floodplain segments sampled. In contrast, highest modern floodplain aggradation rates (≈1.8 cm year –1 ) observed for the relatively small Toshnya River are definitely associated with human impact – locally intensive agriculture. Evaluation of geochemical properties of overbank sediments has shown that general levels of the sediment contamination by heavy metals are low.


INTRODUCTION
Fluvial systems are dynamic geomorphic cascades where sediment and associated compounds are moved from sources to sinks and eventually delivered into the receiving sedimentary basins (lakes, seas, oceans, less frequently no-drainage continental depressions). Main natural controls of the spatial and temporal patterns of fluvial sediment transport are climate, geological, geomorphic and landscape settings of drainage basins and hydrological regime [Makkaveev, 1953;Schumm, 1977;Milliman & Syvitski, 1992]. Deforestation, agricultural development, urbanization, hydropower development, river regulation, mining and other human activities can profoundly affect the processes of erosion and deposition in river systems and may cause co nsiderable changes, both in sediment redistribution within river basins and in sediment export into the ocean [Milliman & Meade, 1983;Dedkov & Mozzherin, 1984, Syvitski et al., 2005]. It is now widely recognized that significant proportion of pollutants in rivers is transported in association with fine-grained suspended sediments [Carter et al., 2003;Salomons & Brils, 2004]. Hence, overbank deposition of suspended sediments on river floodplains creates a sedimentary record which can represent a very informative sampling medium for evaluating pollution levels from various periods [Ottesen et al., 1989;Bogen et al., 1992]. In addition overbank sediments of larger rivers are believed to be representative for their drainage areas since they integrate contribution of numerous smaller tributaries [Ridgway et al., 1995;Xie & Hangxin, 2001]. By sampling sufficiently deep cores in overbank sediment it is possible to characterize recent levels of sediment-associated pollution, its variation over the period of industrial development, and geochemical composition of sediment under pre-industrial pristine conditions [Ottesen et al., 1989;Bogen et al., 1992;Carter et al., 2003]. If reliable dating of overbank sediment cores is available in order to build up recent sediment stratigraphy, it becomes possible to quantitatively determine the changes that have taken place and, possibly to indicate future trends. If the sediment transport in a particular river system of interest has been monitored over sufficient period of time, the geochemical analyses of overbank sediment can be converted into fluxes of various chemical elements.
In this paper we present results of the investigation of floodplain sedimentation rates at selected sites on rivers of the Severnaya Dvina River basin combined with geochemical analysis of depth-incremental sediment samples. Such a combination allows evaluating the level of sedimentassociated contamination of the river basin area, determining point sources of pollution and making some conclusions on delivery of particle-bound pollutants into the Arctic Ocean [Holmes et al., 2002]. Sampling sites were located in different parts of the Severnaya Dvina River basin, in most cases upstream and downstream from the cities with important industrial objects.

STUDY AREA LOCATION AND DESCRIPTION
The Severnaya (Northern) Dvina River basin is located in northern part of the European Russia. The river drains vast areas of predominantly forested and scarcely populated plains and lowlands of northern part of the Eastern European or Russian Plain and flows into the White Sea of the Arctic Ocean (Fig. 1). The Severnaya Dvina River itself begins near the Velikiy Ustyug City at the confluence of its two main sources -the Suhona River sourcing from the Kubenskoe Lake at the southwestern part of the basin and the Yug River with head at the Severnye Uvaly Upland at the basin southern margin. From that point it is named the Malaya (meaning Little in Russian) Severnaya Dvina until it receives from the east its largest right tributary -the Vychegda River (in fact having even larger discharge than the main river in many cases) -about 60 km downstream near the Kotlas City (Fig. 1b).
Total basin area is approximately 357 000 km 2 (5 th largest in the Europe after the Volga, Danube, Dnieper and Don River basins). Total length of the Severnaya Dvina River (with the longer of its two main sources -the Suhona River) is about 1300 km, which makes it only 8 th largest in the Europe (after the Volga, Danube, Ural, Dnieper, Don, Pechora and Rein Rivers). Average annual discharge at the delta apex is about 3420 m 3 s -1 , yielding annual flow into the White Sea of about 108 km 3 year -1 (discharge measurements available from 1881). Peak discharges during the spring flood vary between 11 000 and 36000 m 3 s -1 , with average value of about 22000 m 3 s -1 [Holmes et al., 2002]. Information on suspended sediment yield (SSY) is much more limited (record begins from 1950, with several gaps). Direct measurements of bedload sediment transport were very limited; most of the available information is based on analysis of the bedforms migration [Alexeevskiy, 1998] [Bobrovitskaya, 1996]. There are also glaciofluvial, glaciolacustrine and alluvial Quaternary deposits widespread within the region. Upper till layer with associated glaciofluvial and glaciolacustrine deposits are in most cases covered by up to 2-4 m thick superficial layer of the so-called sheet loams containing no coarse particles. Their exact origin is still questionable, but most likely can be attributed to a complex of processes (frost weathering, aeolian transport and deposition, sheet wash, solifluction) occurring under periglacial conditions after the glacial retreat. These sheet loams represent the most widespread parent material for the regional soil formation.

Geological and geomorphic characteristics
Geomorphology of the Severnaya Dvina River basin is substantially uniform and dominated by lowland undulating plain topography. This general uniformity is interrupted only by hilly upland terrain of the Timanskiy Kryazh Upland along the southeastern margin of the basin, where elevation of the main interfluve between the Severnaya Dvina and  The most important specialization of the region is forest industry (timber harvesting) for constructional purposes (with large proportion of that being exported) and as a primary product for paper and pulp industry (used locally at a few large pulp and paper plants such as those at the Koryazhma, Syktyvkar, Novodvinsk etc.). The most intensive forest cuttings took place during 1970-1980s., mainly within the most developed parts of the basin (surroundings of the Arkhangelsk, Vologda, Syktyvkar, Kotlas) and along major railways. The Severnaya Dvina River is an important transportation route, timber and paper products constituting large percentage of the total cargo volume.

Major sediment and pollution sources
Analysis of general information on physiographic conditions and human activities in the Severnaya Dvina River basin makes it possible to conclude that, in contrast to densely populated and intensively cultivated territories of the Central European Russia, such sediment sources as soil and gully erosion on catchment slopes are unimportant for the basin-scale sediment budget. Their contribution can be considerable only locally in southwestern part of the basin, where percentage of cultivated land is noticeable. Main source of fluvial sediment is bank erosion, which is most intensive on largest rivers of the basin at reaches characterized by free meandering channel types. Consequently, main temporarily sediment sink is represented by vast floodplains of largest rivers characterized by comparatively active reworking as a result of interaction of bank erosion, overbank deposition and lateral floodplain accretion by bedform stabilization. From a point of view of main pollution sources, very low percentage of cultivated land means that pollution of watercourses within the basin is primarily associated with point sources represented by several centers of industrial development with precisely known locations. Certain degree of floodplain sediment pollution can also be caused by airborne contaminants originating from distant sources, which are significantly more difficult to identify [Langedal & Ottesen, 1998]. Additional sediment and (potentially) pollution source specifically for the Severnaya Dvina River basin can be represented by forest cutting areas. However, information for quantification of associated erosion and pollution is at present unavailable.

Overbank sediment sampling locations and strategy
A number of sampling locations has been selected at floodplains of different rivers within the Severnaya Dvina River basin with aim to characterize variability of overbank deposition rates, geochemical properties of floodplain sediment and to determine possible negative impacts of the major industrial centers as point-sources of water and sediment pollution. By now, 10 sections have already been sampled, of which laboratory analyzes have already been completed for six (Fig. 1b) (MSevDV-1), 2 sections on the Vychegda River lower reach downstream (VYCH-1) and upstream (VYCH-2) from the largest paper and pulp plant at the Koryazhma town and 1 section on the relatively small Toshnya River located in headwaters of the Sukhona River (one of the Severnaya Dvina River main sources) in surroundings of the Vologda City characterized by relatively high degree of agricultural development. Detailed visual description of each of the selected sections was undertaken prior to the sampling procedure. Sampling was carried out from naturally undercut floodplain banks where visual analyses showed that at least 0.5 m of fine laminated overbank sediment has been deposited above the underlying sandy cross-bedded channel sediment. The entire sediment column was sampled uniformly with individual samples taken at depth increments of 5 cm from fixed surface area of 10x10 cm 2 . In most cases 1 m of sediment was sampled, covering overbank fine sediment layer and top of the underlying sandy channel deposits. In a few cases, however, longer or shorter columns were sampled, depending on results of section description.

Gamma-spectrometric analysis of 137 Cs content in overbank sediment samples
Representative subsamples with a weight above 100 g were isolated from 10x10x5 cm depth-incremental sediment samples taken for geochemical analysis as described above. Subsamples were dried at 105°C, ground in order to destroy large aggregates and thoroughly mixed to homogenize the material. Particles exceeding 2 mm in diameter were removed by sieving samples through the 2 mm mesh sieve. of the global fallout, Chernobyl fallout and lateral input of sediment-bound 137 Cs into the total point inventory; iii) interpretation of the 137 Cs depth distribution peaks' location in the sediment profile and correspondent activity concentrations, in combination with the available field data including detailed visual description of a sediment profile and floodplain morphology at a sampling location; iv) the final attribution of the available depth profile peaks to corresponding years of maximum fallout (possible options are 1959,1963,1986) and subsequent construction of the recent sediment stratigraphy (additional time mark is provided by the 137 Cs commencement in 1954); and finally v) calculation of average sedimentation rates, with accounting for possible vertical migration of the 137 Cs fallout peaks if such information is available.

Geochemical analyses of overbank sediment samples
Geochemical analyses of the floodplain sediment samples were conducted at the Geological Survey of Norway facilities in Trondheim. After drying, the samples were sieved and the <0.062 mm fraction analyzed for the total contents of 30 elements and an acid soluble fraction of 29 elements. Preparation of an acid soluble fraction involved acid extraction in 7 N HNO3 in autoclave according to Norsk Standard -NS 4770 (1 g sample extracted with 20 ml 7N HNO3 in autoclave for 30 min under 120°C, cooled to room temperature and left overnight before filtration and diluted to 100 ml with distilled water). Analytical techniques applied included the inductively coupled plasma optical emission spectroscopy (ICP-OES) performed using the Perkin-Elmer Optima 4300 Dual View instrument, and the graphite furnace atomic absorption spectroscopy (GFAAS) performed using the Perkin-Elmer SIMA 6000 instrument.

RESULTS AND DISCUSSION
Overbank sediment stratigraphy based on 137 -1 and TOSH-1, Fig. 2a, f ), 2 (section MSevDV-1, Fig. 2e) or 3 (section VYCH-1, Fig. 2c) clearly identifiable peaks of 137 Cs at certain depths, and another one has maximum located at the top of the section (sections SevDV-2, Fig. 2b). These peaks can be attributed to one of the years characterized by maximums of the 137 Cs atmospheric fallout -1959, 1963 or 1986.

Cs depth distribution and reconstruction of recent floodplain aggradation rates
Significant variability of total 137 Cs inventory is observed between the sampled sections. For the 3 sections (SevDV-1, VYCH-1 and MSevDV-1, Fig. 2 a, c, e) total 137 Cs inventory at the sampling point is close to the regional values reported (Atlas.., 1998) suggesting that additional lateral input of the isotope with deposited fine sediment has been very limited. For the other 2 sections (SevDV-2, VYCH-2, Fig. 2 b, d) total isotope inventory is substantially lower than the regional values. In case of the section SevDV-2 it can most likely be explained by removal of the upper part of the overbank sediment layer by erosive action of the floating ice during one of the recent floods, as depth penetration of the 137 Cs is very low and the peak is found on top of the sampled section (Fig. 2b). In contrast, for the section VYCH-2 it is obvious that the sampling depth was insufficient to reach the base of the 137 Cs depth distribution (Fig. 2d). This section is characterized by low thickness of the fine overbank sediment layer (<0.5 m) underlain by cross-bedded sandy channel deposits. Thus it can be suggested that this particular floodplain segment is very young and formed after the isotope fallout has already ceased. Generally low concentrations of 137 Cs in the sediment prove that it has been originated not from the atmospheric fallout, but mostly from lateral input and redeposition of material from other floodplain segments eroded upstream (Fig. 2d)  The last section (TOSH-1, Fig. 2f ) is characterized by very high 137 Cs inventory, exceeding by about 3 times the regional fallout values reported (Atlas.., 1998), even though in this section it is also obvious that the sampling depth was insufficient to reach the base of the 137 Cs depth distribution. It must be noted that, despite certain fluctuations, no systematic upward decrease of the isotope concentration in deposited sediment occurs in the section TOSH-1 (Fig. 2f ). It is obvious from very high 137 Cs inventory, shape and depth penetration of the isotope distribution curve that the main sediment source in this case is not bank erosion upstream but soil erosion on arable valley slopes.
From the presented 137 Cs depth distribution curves, reliable sediment stratigraphy and reconstruction of overbank deposition rates can be carried out for the 4 sections with clearly distinctive peaks (SevDV-1, VYCH-1, MSevDV-1 and TOSH-1, Fig. 2). For the section SevDV-1 (Fig. 2a)  Summarizing the above data, it can be stated that comparatively high spatial-temporal variations of sedimentation rates are observed between the sampled floodplain sections within the Severnaya Dvina River basin. It can be suggested that floodplain deposition rates for sampled reaches of the Severnaya Dvina and Vychegda Rivers (except for the youngest floodplain segment at VYCH-2) generally decreased from ≈1.5-4.0 cm year -1 between 1954 and 1963 to ≈0-1.0 cm year -1 at present. Certainly, it is not reliable to use data from 3 individual points (sections SevDV-1, MSevDV-1 and VYCH-1because sections SevDV-2 and VYCH-2 do not provide estimates of deposition rates) for any large-scale extrapolations of floodplain sedimentation rates. Nevertheless, they provide some insight on floodplain overbank sedimentation rates for relatively young floodplain segments located in proximity to the main channels. It is believed that the observed tendencies reflect normal natural evolution of floodplain segments where sampling sites were selected. Gradual floodplain aggradation results in less frequent and prolonged inundation and, as a consequence, in the decrease of overbank deposition rates. Representativeness of the point-estimated values for the entire floodplain segments is supported by the fact that all of the sections sampled are characterized by uniform structure of upper laminated fine sediment layer, without any abrupt changes in lamination along its strike. Main source of fine sediment in that case is erosion of floodplain banks located further upstream.
The highest average overbank deposition rate for the period of 1963-2008 (≈1.8 cm year -1 ) was detected at the Toshnya River floodplain (Fig. 1b, 2f ), which represents the smallest of the rivers sampled. It is located in the territory with relatively high area of cultivated lands, which can locally represent an important sediment source. Such floodplain aggradation rate is very significant for such a relatively small river. However, it must be borne in mind that most of the sediment originating from soil erosion on cultivated land and delivered into small or medium rivers becomes redeposited in their valleys. For example, for the Volga River basin it has been estimated that only less than 10% of catchment-derived sediment reaches valleys of rivers with a drainage basin area exceeding 10000 km 2 [Sidorchuk, 1995;Sidorchuk & Golosov, 2003]. In addition, in the case of the upper tributaries of the Suhona River (including the Toshnya River), most of the suspended sediment delivered from these tributaries into the main river is intercepted and redeposited in the area where the Suhona River valley crosses the former glacial lake bottom and has wide inherited floodplain. Therefore, floodplain deposition data for the Toshnya River are applicable only for small and medium rivers of the most densely populated and agriculturally developed southwestern part of the Severnaya Dvina River basin. This sediment source is almost completely disconnected from the further parts of the fluvial sediment cascade.

Geochemical properties of overbank sediments
Geochemical properties of overbank deposits can now be evaluated basing on results of the laboratory analyses of 90 samples from the same 6 floodplain sections discussed above. Average and maximum values of concentrations of the 9 selected heavy metals in the analyzed samples are presented in the Table 1, where those are also compared with the global clarke concentrations, global clarke for clayey deposits (as dominant surface materials  Fig. 3).
However, substantially increased concentrations of certain heavy metals (in particular Mn, Zn and As) exceeding the MAC values have been detected in individual sediment layers at sampling sites located immediately downstream from industrial centers. Those can point either to relatively short-term pollutant dumpings, for example, associated with failures of the industrial waste treatment systems, or to more prolonged periods of continuous input of pollutants into the fluvial system. For example, Fig. 3 shows comparison of Mn, Zn and As content in overbank sediment of the Severnaya Dvina River lower reach downstream (section SevDV-1 - Fig. 3a-c) and upstream (section SevDV-2 - Fig. 3d-f) from the largest industrial center -Arkhangelsk City. At the upstream section no excess concentrations of the selected heavy metals above the MAC values have been observed, while downstream there are 4 successive samples with concentrations of As and 1 sample with concentration of Zn exceeding the MAC values (Fig. 3b, c). All these 5 samples are located below the 137 Cs maximum penetration depth and, thus, can be attributed to the first half of 20 th In general overbank floodplain deposits sampled within the Severnaya Dvina River basin are characterized by substantial variability of concentration of various chemical elements, both within a single section and between sections, as can also be exemplified by Fig. 3. Nature of this variability is most likely not uniform. For example, practically identical patterns of depth distribution of Mn, Zn and As in section SevDV-2 ( Fig. 3d-f ) can most likely be explained by variation of finer fractions of sediment to which these heavy metals are preferentially bound. On the other hand, the observed peaks of Zn and As concentrations in section SevDV-1 (Fig. 3a-c), located in <45 km downstream can be explained only by anthropogenic impact. Further investigations may also allow us to determine different geochemical signals in overbank sediment originated from geologically different parts of the Severnaya Dvina River basin, thus making it possible to evaluate relative contribution of different sediment sources and their temporal variability [Carter et al., 2003;Collins & Walling, 2004]. Combination of 137 Cs-based stratigraphy, geochemical analysis and quantitative information on sediment fluxes will allow evaluating the level of sedimentassociated contamination of the river basin area, determining point sources of pollution and making some conclusions on delivery of particle-bound pollutants into the Arctic Ocean.

CONCLUSIONS
Most of the Severnaya Dvina River basin area is characterized by very low level of agricultural development. Therefore, sediment production from soil and gully erosion on arable land has very minor contribution into the basin-scale fluvial sediment budget. Much more important contributors into sediment yield of main rivers within the basin are bank erosion directly on these rivers and sediment export from their smaller tributaries. The latter is also mainly originated from bed and bank erosion, as small river catchments are mostly forested. Main temporarily sediment sink within the Severnaya Dvina River basin fluvial system is represented by vast floodplains of largest rivers characterized by comparatively active reworking as a result of interaction of bank erosion, overbank deposition and lateral floodplain accretion by bedform stabilization.
Application of the 137 Cs radioactive tracer has allowed establishing recent overbank sediment stratigraphy and reconstructing aggradation rates for several floodplain sections for 1-4 time intervals since 1954. It has been shown that the sampled relatively young floodplain segments of large rivers experience natural tendency of decreasing aggradation rates associated with declining frequency and magnitude of their inundation as a result of gradual increase of surface elevation above the channel. Main source of fine sediment in that case is erosion of floodplain banks located further upstream. In contrast, floodplain of the relatively small Toshnya River located in area characterized by intensive agricultural development experiences continuous aggradation without noticeable tendency of decrease, with main sediment source being soil erosion on cultivated valley slopes. Evaluation of geochemical properties of the Severnaya Dvina River basin overbank sediments for the same floodplain sections has shown that general situation is much better than in the Central European Russia, in particular as regards the sediment pollution by heavy metals. In most of the samples concentrations of pollutants were below the MACs and comparable to the clark concentrations for clay fractions. However, substantially increased concentrations of certain heavy metals (in particular Mn, Zn and As) exceeding the MAC values have been detected in individual sediment layers at sampling sites located immediately downstream from industrial centers. Those can point to relatively short-term pollutant dumpings, for example, associated with failures of the industrial waste treatment systems, or to more prolonged periods of continuous input of pollutants into the fluvial system.
Further research perspectives will be in quantitative assessment of contribution of different sediment sources on one hand and different pollution sources on the other hand. Main attention will be paid to collection of measurements of bank erosion rates for largest rivers of the basin, expanding database on recent floodplain aggradation rates and searching for independent sources of this information.
The established geochemical properties of overbank sediments will be compared with those of main sediment and pollution sources. Long-term data on suspended sediment yields will be used to quantify fluxes of major particle-bound pollutants and compare their contribution into the Arctic Ocean pollution with available information on solute fluxes.