SUSPENDED AND DISSOLVED MATTER FLUXES IN THE UPPER SELENGA RIVER BASIN

We synthesized recent fieldbased estimates of the dissolved ions (K Na Ca Mg Cl S 2– 4 O HC –3 O ), biogens (N – 3 O , N – 2 O , P 3– 4 O )(C, mg/l), heavy metal (Fesum, Mn, Pb) and dissolved load (DL, kg/day), as far as suspended sediment concentration (SSC, mg/l) and suspended load (SL, kg/day) along upper Selenga river and its tributaries based on literature review and preliminary results of our 2011 field campaign. The crucial task of this paper is to provide full review of Russian, Mongolian and English-language literature which concern the matter fluxes in the upper part of Selenga river (within Mongolia). The exist estimates are compared with locations of 3 main matter sources within basin: mining and industry, river-bank erosion and slope wash. The heaviest increase of suspended and dissolved matter transport is indicated along Tuul-Orkhon river system (right tributary of the Selenga River where Mongolia capital Ulanbaatar, gold mine Zaamar and few other mines are located). In measurement campaigns conducted in 2005, 2006 and 2008 the increase directly after the Zaamar mining site was between 167 to 383 kg/day for Fe, between 15 and 5260 kg/day for Mn. Our field campaign indicated increase of suspended load along Tuul river from 4280 kg/day at the upstream point to 712000 kg/day below Ulaanbaatar and Zaamar. The results provide evidence on a potential connection between increased dissolved and suspended matter fluxes in transboundary rivers and zones of matter supply at industrial and mining centers, Sergey R. Chalov, Aleksandr S. Zavadsky, Ekaterina V. Belozerova, Mariya P. Bulacheva, Jerker Jarsjö, Josefin Thorslund, Jambaljav Yamkhin 1 Research associate, Faculty of Geography, M.V.Lomonosov Moscow State University, Moscow, Russia; Leninskie gory, 1, 1199911, Tel. +7 495 9391552, E-mail: srchalov@rambler.ru * Corresponding author 2 Senior scientist, Faculty of Geography, M.V.Lomonosov Moscow State University, Moscow, Russia; Leninskie gory, 1, 1199911, Tel. +7 495 9391233, E-mail: az-mgu@rambler.ru 3 Postgraduate student, Faculty of Geography, M.V.Lomonosov Moscow State University, Moscow, Russia; Leninskie gory, 1, 1199911, Tel. +7 495 9391533, E-mail: ekv.belozerova@gmail.com 4 Postgraduate student, Faculty of Geography, M.V. Lomonosov Moscow State University, Moscow, Russia; Leninskie gory, 1, 1199911, Tel. +7 495 9392576, E-mail: mbulacheva@mail.ru 5 Associate professor, Department of Physical Geography and Quaternary Geology, Stockholm University, SE-106 91, Stockholm, Sweden; Tel. +46 8 164958, E-mail: jerker.jarsjo@natgeo.su.se 6 M.Sc. student, Department of Physical Geography and Quaternary Geology, Stockholm University, SE-106 91, Stockholm, Sweden; Tel. +46 7 03659249, E-mail: thorslund.josefin@gmail.com 7 Senior scientist, Institute of Geography, Mongolian Academy of Sciences; Ulaanbaatar, Mongolia, Amariin gudamj 1, Tel. +986 11 262247, E-mail: jambaljav@gmail.com SUSPENDED AND DISSOLVED MATTER FLUXES IN THE UPPER SELENGA RIVER BASIN gi212.indd 78 15.06.2012 12:48:17 79 EN VI RO NM EN T along eroded river banks and pastured lands. The gaps in the understanding of matter load fluxes within this basin are discussed with regards to determining further goals of hydrological and geochemical surveys.

ABSTRACT.We synthesized recent fieldbased estimates of the dissolved ions (K + Na + Ca 2+ Mg 2+ Cl -S 2- O )(C, mg/l), heavy metal (Fe sum , Mn, Pb) and dissolved load (DL, kg/day), as far as suspended sediment concentration (SSC, mg/l) and suspended load (SL, kg/day) along upper Selenga river and its tributaries based on literature review and preliminary results of our 2011 field campaign.The crucial task of this paper is to provide full review of Russian, Mongolian and English-language literature which concern the matter fluxes in the upper part of Selenga river (within Mongolia).The exist estimates are compared with locations of 3 main matter sources within basin: mining and industry, river-bank erosion and slope wash.The heaviest increase of suspended and dissolved matter transport is indicated along Tuul-Orkhon river system (right tributary of the Selenga River where Mongolia capital Ulanbaatar, gold mine Zaamar and few other mines are located).In measurement campaigns conducted in 2005, 2006 and 2008 the increase directly after the Zaamar mining site was between 167 to 383 kg/day for Fe, between 15 and 5260 kg/day for Mn.Our field campaign indicated increase of suspended load along Tuul river from 4280 kg/day at the upstream point to 712000 kg/day below Ulaanbaatar and Zaamar.The results provide evidence on a potential connection between increased dissolved and suspended matter fluxes in transboundary rivers and zones of matter supply at industrial and mining centers,

SUSPENDED AND DISSOLVED MATTER FLUXES IN THE UPPER SELENGA RIVER BASIN
along eroded river banks and pastured lands.The gaps in the understanding of matter load fluxes within this basin are discussed with regards to determining further goals of hydrological and geochemical surveys.
KEY WORDS: mass flow, suspended and dissolved matter transport, transboundary rivers

INTRODUCTION
Matter supply into water systems affects humans and the natural environment world-wide.Both natural processes, such as dissolution of substances from weathering soil and rock, and anthropogenic activities, particularly from the agriculture and industry sectors (UNEP, 2009), can cause suspended and dissolved matter transport.Suspended and dissolved particles transported by river flow can originate either from input into channels from drainage basins (basin-sourced sediment) or from particle detachment within the channels themselves from their beds and banks (channel-sourced sediment).On any river one can find intensively eroded channel banks and recently formed accumulative within-channel bars.The different sources and characteristics of the channel particles, as well as the different river processes, induce heterogeneities in particle compositions and ultimately determine the unequal fall velocities and sediment distributions.Vertical sediment fluxes significantly correlate with transporting capacity.Near-bottom sediment exchange is the result of general laws of matter movement which are governed by the turbulent diffusion equation [Alexeevsky, 2006]: where the different terms reflect impacts of turbulent transport, advection, dispersion, convection and gravity, respectively, and A is the turbulent exchange coefficient, ρ -water density, s -suspended sediment concentration at the point, u, v, w -local velocity vectors.In basin-scale transport assessments, a crucial task is connected with the quantification of the advection and dispersion terms.Key challenges are furthermore to identify the exact locations of main sources in river basins, and the different contributions of natural and anthropogenic impacts.
Understanding of matter fluxes along river system depends on the monitoring system.

METHODS AND MATERIAL
The study area covers the upper part of the Selenga basin within Mongolia (Fig. 1).The Mongolian portion of the Selenga River watershed is composed predominantly of broad alluvial valleys flowing through steppe grasslands with source areas in taiga and mountain ecosystems.Maximum river discharge is driven by the spring melt of the accumulated snowpack.A second peak in river hydrographs is observed in late summer, August or September, during the rainy season.Our field campaign was conducted during the summer runoff increase and thus under the conditions of significant increase in sediment and dissolved matter concentrations.
Existing estimates of suspended and dissolved solids concentration were synthesized and analyzed in terms of spatial distribution in river basin, its temporal variability and possible linking with various sources of matter supply.We used gathered data to assess mass flow in a river which represents the matter mass that passes through a cross section (also referred to as a control plane) of the river per unit of time.Mass flows are products of local concentrations and discharges, according to where C i is the concentration (SSC or C) at the control plane for (given) time t i , and Q i is the water discharge through the control plane for (given) time t i .Time period was applied as 1 day, therefore the MF (both DL and SL) was estimated as The analyzed dissolved and suspended matter concentrations (dissolved and particulate) were synthesized from our 2011 measurements and reviews and reports devoted to measurements campaigns in rivers of the region [Kuznetsov, 1955 O ).This paper does not introduce to the results of our samples analyses of biogens (nitrogen and phosphorous) and heavy metals (such as Zn, Cu, Ni, Cd, Pb) content in water and in sediments (bed and suspended).
In all reported works the discharge measurement and sampling procedures followed standard methods.Depthintegrated water samples were collected with a GR-16M bottle sampler at the midstream.Turbidity was measured on site using portable "HACH" 2100P meter.
To study SSC the samples were filtered through pre-weighed membrane and paper filters with the "Millipore" filtration system.The samples were then ovendried and re-weighed.Discharges were determined from bridges, by boat or by wading in two ways.In the first case, flows were measured with a hydrometric propeller ISP-1 at the one-fifth depths of each width increment.The total water discharges were calculated by multiplying the discharge flow velocities with crosssectional areas of the rivers.The total sediment discharges were calculated by multiplying sediment-load velocities with average SSC-values and cross-sectional areas of the rivers.When the depth was more than 1.5 meters the Acoustic Doppler Profiler (ADP) was used.
For the present study we used information only from the stations, where any relevant hydrochemical measurements have earlier been performed.Observations at 14 stations from total 35 were used for present analyses, comprising data on suspended sediment load (SSC and SL) and concentration of dissolved ions (Ca 2+ , Mg 2+ , K + , Na + , Cl -, S 2-

CONTEMPORARY ASSESSMENTS OF SUSPENDED AND DISSOLVED MATTER FLUXES
The synthesized data concerning suspended sediment load is presented in table 1, whereas table 2 is devoted to dissolved load.Analyses of the contemporary spatial distribution of water quality parameters in the upper Selenga rivers demonstrate, that there are few driving forces in supply of material into river networks.The main sources are primarily associated with mining activities, in-channel processes and slope erosion (Fig. 2).
According to existing estimates, considerable increase of dissolved load is observed along the Tuul river.Mass flows are relatively low upstream, but can increase by orders of magnitudes directly indicating a significant impact from the mining activities, Ulaanbaatar    2008] showed that the increase directly downstream of the mining site (from dissolved concentrations) were found to be between 167 to 383 kg/day for Fe, between 15 and 5260 kg/day for Mn, and between3.6 to 4.6 kg/day for Pb, compared with values upstream the site.Previously [Stubblefield et.al, 2005] reported increase of SSC and total phosphorus along Tuul river.Our field campaign indicated increase of suspended load along Tuul river from 4280 kg/day at the upstream point to 712000 kg/day below Ulaanbaatar and Zaamar (Fig. 3).The following part of the paper is devoted to the analyses of main sources of matter supply within basin.

Mining and industry
In recent years, exploration for natural resources has increased rapidly.Many river basins are used intensively for mining of gold, silver, and coal, supplying also precious stones, gravel, and other natural resources.A total of 784 enterprises are engaged in mining, of which 204 small-scale gold mining companies are operating on 6,065,298 hectares of land [Batimaa, 2011].Some of the gold miners are reported to use mercury in the gold extraction.The surface water inventory revealed that gold mining activity affects 28 rivers in 8 provinces of Mongolia.Many square kilometers of the river terraces are heavily disturbed.Flooding could breach thin strips of land separating dredge pits from river channels, resulting in massive sediment loading.
The most significant evidences of mining impact on suspended and dissolved matter fluxes of upper Selenga rivers exists along the Tuul river where the Zaamar Goldfield is located.The Selenga basin zoning was done on the basis of channel-formation activity of rivers.
According to such a zoning the territory was divided into three districts: the Upper Selenga district, the Orkhon-Selenga district and the Khantay district.Each district has its own specific conditions of channel-forming determined by effective water discharge and character of inundation of floodplains.
In conditions of free development of channel deformations in Upper Selenga and Khantay districts multi-thread channels are widespread due to regular and long-term floodings.In the Orkhon-Selenga district, along with multi-thread channels, meanders forming in both free and adapted conditions are also widespread.These results provide us with understanding of crucial role of channel evolution in matter supply to rivers.At the same time the absence of exact descriptions of channel development is still prevents from a complete assessment of channel erosion contribution to total sediment and dissolved matter fluxes.

Soil erosion
Overgrazing and land-use mismanagement are considered to be main causes of soil erosion and land degradation in the upper Selenga basin rivers.Grassland vegetation has been reported to have decreased recently due to overgrazing by an increasing number of livestock [Chuluun and Ojima, 2001].In general, the soil in a semiarid area has low organic content and a large percentage of silt portions, which results in a high susceptibility to soil crusting.The decrease in grassland vegetation increases the raindrop impact on the soil surface, splash erosion, and susceptibility to surface sealing, thereby causing enhanced water erosion.Therefore, the surface vegetation cover is considered to be the primary reason for soil erosion [Onda et al., 2006].
Increase of erosion process intensity in the first two-thirds of the 20 th century in the Selenga River watershed and a reduction of this intensity in the last third of the century is identified by [Korytny et al., 2003].Changes in the river network structure (the order of rivers, lengths, etc.) as a result of agricultural activity during the 20 th century are regarded to be a driving factor of this increase.The soil erosion rate was observed to be directly proportional to the severity of grazing [Chen et al., 2007].
Exact links between field data available on soil erosion and land degradation processes and matter movement along waterways are limited.Nevertheless some data show evidence that large floods, the rarest of hydrological events, can have the biggest impact, carrying the most sediment and dissolved matter mostly due to surface erosion.Observations made upstream of the Kharkhorin village at the Orkhon river at 29-31 of July demonstrate a significant increase of suspended load due to heavy rains and water level increase (Fig. 4).Our estimates show that the sediment load was about 3000 tons per day, which exceeds the average value for this period 10 times.It should be noted that the upstream reaches of these rivers contain large gold mining areas.Thus, the rates of matter increase could also be affected by mining.Various types of mining activities appear to provide significant changes in suspended or dissolved load.Open-cast gold mines of Tuul and Orkhon river basin lead to significant increase of SSC and SL.Flooding could breach thin strips of land separating dredge pits from river channels, resulting in massive sediment loading, as it was observed in the studied rivers.Elevated loading of dissolved ions was observed below ore mining and ore processing factory of the Erdenet in the Khangal river.At the same time low water discharge of the Khangal river provide fast decrease of dissolved solids concentration downstream along river network (below Khangal river confluence with large Orkon river).Observed increase of dissolved solid load along mining zone could be compared with total water-borne mass flow of metals to surface waters in England and Wales (which is 1509 kg/day for Fe, 13.9 kg/day for Mn and 50.6 kg/day for Pb) [Mayes et al. 2010].

CONCLUSION
Mass flows in the Upper Selenga river basin are somewhat different from both suspended (SSC) and dissolved solids concentration measurements.For instance, mass flows of Fe were increasing below Zaamar goldfield and downstream at a much higher rate than concentrations do.This is due to incorporation of local discharges with local concentrations and discharges increase considerably at downstream locations of the river system.Mass flows are relatively low at the mining site, but can increase by orders of magnitudes directly after the site indicating a significant impact from the mining activities.
More generally a main result is that while both suspended and dissolved material is provided from different sources, the total annual mass flow mostly depends on specific hydrological events.Large floods can have the biggest impact, carrying the most sediment from the surface -both from mined and nonmined lands, and reworking channel and floodplain geometry.137Cs inventory for unmined basins showed [Onda et al., 2006] that that the high sediment yields in the former case might be due to the susceptibility to erosion by recent over grazing.Industry accelerates both watershed and channel erosion and thus increases contributions from all sources.Further understanding of sources contribution, as far as understanding of sediment and dissolved matter regional budget could be done on joining largescaled and small-scaled observations.
Potentially important questions that have not been addressed include effects of sediment and dissolved matter fluxes in ground waters.Wastewaters in the area of mining and industrial centers can cause serious alterations in groundwater recharge.Since groundwater is the main drinking water source in the region [AATA, 2008] and pollution of aquifers is often irreversible due to very slow recharge rates [Zandaryaa et al. 2008] -the process should thus be controlled.
Overall, to increase the knowledge about sediment and dissolved matter fluxes spreading and long term effects in this region, hydrological measurement and monitoring need to be extended.Simultaneous monitoring of both discharges and concentrations is essential for the reliability of mass flow estimations.If such extensive monitoring would be implemented, this would considerably decrease uncertainties in future (mass flow) investigations.Addressing of individual (and combined) effects of multiple upstream matter source zones on downstream mass flows would be of great value, from both scientific and pollution management perspectives.

Fig. 2 .
Fig. 2. Main sources of matter supply a -mining (Zaamar), b -in-channel (bank) erosion (Orkhon river upstream from Tuul confluence), с -slope wash (near Kharkhorin village) sector.More specifically, the three considered measurement campaigns conducted in 2005, 2006 and 2008 [AATA, -situ sampling for temporal analyses.At the points where measurements were conducted several time in the 20th and 21st centuries (e.g.T-6, O-8) the significant fluctuations of both SSC and DSC are indicated.At the Tuul river upstream from the confluence with Orkhon river (T-6 station) measurements were done 7 times, SSC varied between 11 mg/l at 7 Oct 1934 to 716 mg/l at 26 Aug 1934.Maximal SSL was 3627,8 • 10 3 kg/day at 4 Aug 1934 indicating the significant impact of water discharge on sediment load.Reported values of major ions concentration evidence up to 5 times change between measurements.For example, at the downstream reach of Orkhon river (above confluence with Selenga river) (O-9 station) major ions concentration varied up to 6 times between Jul, 2009 and Aug 2011.Simultaneously ions concentrations were highest during low water period [Votincev et al., 1965], and decreased significantly during high water periods (e.g.our field measurement during August 2011).

Fig. 3 .
Fig. 3.Estimated mass flows (kg/day) of total suspended and dissolved load at five different sampling points along the Tuul River (numbers according to tables 1-2)

Table 2 . Estimates of dissolved solid
[Karpoff and Roscoe, 2005]mar Goldfield has been active since the seventies[Karpoff and Roscoe, 2005].Today the area has the greatest gold production in Mongolia with 147 tons produced from 1998 to 2007[AATA, 2008].The largest and most recent gold mine in the Zaamar Goldfield is the Big Bend Placer Gold Mining Project (Big Bend).