SEDIMENT TRANSPORT FROM SOURCE TO SINK IN THE LAKE BAIKAL BASIN

Jan Pietroń

Sediment transport from source to sink in the LakeBaikal basin

Impacts of hydroclimatic change and mining

Jan Pietroń

©Jan Pietroń, Stockholm University 2017 ISBN print 978-91-7649-842-2ISBN PDF 978-91-7649-843-9ISSN 1653-7211 Cover art: 'Selenga' by Balzhinima Dorzhiev, permission for reproduction granted by the Director of Khankhalaev Gallery in Moscow, RussiaDivider photographs: Paper I (Swedish team at a train station in Ulan Ude, 2014) and IV (Fieldwork in the Selenga Delta, 2016) by Jan Pietroń, Paper II (Placer mines along the Tuul River, 2014) by Jerker Jarsjö, and Paper III (Water dis- charge measurements at the Tuul River, 2012) by Josefin Thorslund. Typeset with LATEX using Department of Physical Geography dissertation templatePublished articles typeset by respective publishers, reprinted with permissionPrinted by: Universitetsservice US-AB, Stockholm 2017Distributor: Department of Physical Geography, Stockholm University

In memory of mygrandparentsTeresa and Stanisław Litak

Abstract

Different magnitude, intensity and timing of precipitation can impact runoff, hillslopeerosion and transport of sediment along river channels. Human activities, such as damconstruction and surface mining can also considerably influence transport of sedimentand sediment-bound contaminants. Many river basins of the world are currently sub-ject to changes in climate at the same time as pressures from other human activitiesincrease. However, because there are often complex interactions between such mul-tiple drivers of change, it is challenging to understand and quantify contributions ofindividual drivers, which is needed in predictive modelling of future sediment andcontaminant flows. This thesis considers sediment transport in the Lake Baikal basin,which is hydrologically dominated by the transboundary Selenga River of Russia andMongolia. The Selenga River basin is, for instance, subject to climate change and in-creasing pressures from mining, but process complexity is reduced by the fact that theriver basin is one of few large basins in the world that still is essentially undammed andunregulated. A combination of field measurement campaigns and modelling methodsare used in this thesis, with the aim to: (i) identify historical hydroclimatic trends andtheir possible causes, (ii) analyse the spatial variability of riverine sediment loadingin the mining affected areas, and (iii) investigate sediment transport and storage pro-cesses within river channels and in river deltas. Results show that, during the period1938-2009, the annual maximum daily flow in the Selenga River basin has decreased,as well as the annual number of high flow events, whereas the annual minimum dailyflow has increased. These changes in discharge characteristics are consistent with ex-pected impacts of basin-scale permafrost thaw. Both field observations and modellingresults show that changes in magnitude and number of high-flow events can consid-erably influence the transport of bed sediment. In addition, the average discharge hasdecreased in the past 20 years due to an extended drought. Under conditions of lowflow, metal-enriched sediment from mining areas was observed to dominate the riverwater. If discharge will continue to decrease in the Selenga River (or other mining-impacted rivers of the world), further increases in riverine metal concentrations mayhence be one of the consequences. Furthermore, under current conditions of extendeddrought, less sediment may have been distributed over the floodplain wetlands in theSelenga River delta. Present estimates, however, show that sediment can still be trans-ported to, and deposited within, the banks and water bodies located in the backwaterzone of the Selenga River delta. This can aid bank and levee stabilization, support thedevelopment of wetlands and foster net sedimentation.

Sammanfattning

Nederbördens olika magnitud, intensitet och tidpunkt kan påverka ytavrinning, vat-tenerosion och transport av sediment längs flodkanaler. Mänskliga aktiviteter, somdammkonstruktion och gruvdrift i dagbrott kan också påtagligt påverka transport avsediment och sedimentbundna föroreningar. Många avrinningsområden i världenpåverkas för närvarande av klimatförändringar samtidigt som trycket från andra män-skliga aktiviteter ökar. Men eftersom det ofta förekommer komplexa interaktioner mel-lan sådana multipla orsaker till förändring, är det utmanande att förstå och kvanti-fiera bidrag från enskilda orsaker, vilket behövs vid prediktiv modellering av framtidasediment- och föroreningsflöden. Denna avhandling behandlar sedimenttransport i Ba-jkalsjöns tillrinningsområde, som hydrologiskt domineras av den internationella Se-lengafloden i Ryssland och Mongoliet. Selengaflodens tillrinningsområde är exem-pelvis påverkat av klimatförändringar och ökat tryck från gruvdrift, men processkom-plexiteten reduceras av det faktum att tillrinningsområdet är ett av världens få storasom fortfarande väsentligen saknar dammar och flödesreglering. I denna avhandlinganvänds en kombination av fältmätningskampanjer och modelleringsmetoder, i syfteatt: (i) identifiera historiska hydroklimattrender och deras möjliga orsaker, (ii) anal-ysera den rumsliga variationen i flodens sedimentbelastning inom de gruvpåverkadeområdena, och (iii) undersöka sedimenttransport- och retentionsprocesser inom flod-kanaler och i floddeltan. Resultaten visar att det årliga maximala dygnsflödet, lik-som det årliga antalet högflödeshändelser, har minskat i Selengafloden under perio-den 1938-2009, medan det årliga minimala dygnsflödet har ökat. Dessa förändringari flödeskaraktäristika överensstämmer med förväntade effekter av storskaligt tinandepermafrost. Både fältobservationer och modelleringsresultat visar att förändringar ihögflödeshändelsers magnitud och årligt antal kan påverka transporten av bottensedi-ment påtagligt. Dessutom har medelflödet minskat under de senaste 20 åren på grundav långvarig torka. Under lågflöden observerades metallberikat sediment från gru-vområdena dominera flodvattnet. Om flödena fortsätter att minska i Selengafloden(eller andra gruvdriftspåverkade floder i världen), kan således ytterligare ökningar avflodvattnens metallkoncentrationer vara en av konsekvenserna. Under den långvarigatorka som nu råder mängden sediment som fördelats över våtmarkerna i Selengaflo-dens delta ha minskat. Sediment beräknas dock fortfarande kunna transporteras tilloch deponeras inom flodbankar och vattenkroppar i Selengadeltats backwaterområden.Detta kan bidra till stabilisering av bankar och skyddsvallar, stödja våtmarkers utveck-ling och främja nettosedimentering.

Streszczenie

Wielkosc, czas trwania oraz intensywnosc opadów atmosferycznych oddziałuje nacharakter odpływu, erozje oraz transport osadów rzecznych. Równiez ingerencjaczłowieka w srodowisko – np. budowa zapór i zbiorników wodnych, czy górnictwo od-krywkowe – w róznym stopniu moze wpływac na transport osadów oraz powiazanychz nimi zanieczyszczen. Wiele dorzeczy na Ziemi, bedacych pod wpływem obecnychzmian klimatycznych, jest jednoczesnie poddawanych narastajacej antropopresji.W celu przewidywania przyszłych zmian w transporcie osadów i powiazanych z nimizanieczyszczen, potrzeba dogłebnego zrozumienia i oceny wpływu poszczególnychczynników powodujacych te zmiany. Taka analiza jest jednak czesto utrudniona zewzgledu na złozone interakcje pomiedzy czynnikami powodujacymi zmiany.

Niniejsza rozprawa doktorska przedstawia wyniki badan zwiazanych z analizatransportu osadów rzecznych w zlewni jeziora Bajkał, zdominowanej hydrologicznietransgraniczna rzeka Selenga, przepływajaca przez tereny Rosji i Mongolii. Zlewniarzeki Selengi podlega współczesnym zmianom klimatycznym oraz wzrastajacej presjizwiazanej z górnictwem. Złozonosc procesów hydrologicznych jest jednak w tymwypadku ograniczona, poniewaz zlewnia Selengi jest jednym z nielicznych, wzgled-nie duzych dorzeczy na swiecie, którego przepływy sa – jak dotychczas – naturalne,nieuregulowane przez zadne zapory lub zbiorniki wodne. Dla poszczególnych celów:(i) identyfikacji historycznych trendów hydroklimatycznych i ich przyczyn, (ii) anal-izy przestrzennych zmian w transporcie osadów rzecznych w czesci zlewni dotknietejgórnictwem odkrywkowym oraz (iii) badania procesów transportu i magazynowaniaosadów w korycie i delcie rzeki; zostały w pracy zastosowane hydrometryczne danepomiarowe, dane pochodzace z badan terenowych oraz metody modelowania. Wynikibadan wskazuja na to, ze w latach 1938-2009 zmalały roczne przepływy maksymalneoraz liczba wezbran, podczas gdy w tym samym czasie wzrosły roczne przepływy min-imalne. Powyzsze zmiany sa zgodne z oczekiwanym wpływem rozmarzania wiecznejzmarzliny na ustrój przepływów rzecznych. Analiza danych pomiarowych orazwyników modelowania wskazuja na to, ze obecne zmiany dotyczace liczby orazwielkosci wezbran moga znacznie wpłynac na transport osadów dennych w korytachrzek. Dodatkowo, w ciagu ostatnich 20 lat (1995-2014), srednie roczne przepływyznacznie spadły ze wzgledu na przedłuzajacy sie okres suszy na terenie zlewni. Anal-iza danych terenowych pochodzacych z obszarów górniczych wykazała, ze podczas ob-nizonych przepływów, w zanieczyszczonych znaczna iloscia metali osadach rzecznych,dominuje materiał pochodzacy z działalnosci człowieka (około 80% transportowanychosadów). Nalezy zatem przewidywac, ze jesli obecne zmiany w ustroju przepływów wdorzeczu Selengi (lub w innych podobnych dorzeczach na swiecie) beda postepowac, toich nastepstwem moze byc dalszy wzrost koncentracji zanieczyszczen (metali pochodza-cych z obszarów górniczych) rzek. Ponadto, w obecnym okresie obnizonych przepły-wów, na terenach zalewowych i jeziorach delty rzeki Selengi zatrzymuje sie praw-dopodobnie mniej osadów. Wyniki badan wskazuja jednak na to, ze osady rzecznemoga byc wciaz transportowane do brzegów i obszarów wodnych znajdujacych sie wstrefie, w której stany wodne cieków delty sa pod wpływem stanów wodnych jezioraBajkału. Akumulacja materiału w tych czesciach delty Selengi moze pozytywnie wpły-wac na stabilizacje naturalnych wałów oraz mokradeł i zwiekszac sedymentacje netto.

Dissertation content

This doctoral thesis consists of a summary text and four papers The paperswill be referred to by their Roman numbers (Papers I-IV) in the summary text.The published papers are reprinted by permission from the respective copyrightholders.

I Törnqvist R., Jarsjö J., Pietron J., Bring A., Rogberg P., Asokan S.M.and Destouni, G. 2014. Evolution of the hydro-climate systemin the Lake Baikal basin. Journal of Hydrology 519, 1953-1962,doi:10.1016/j.jhydrol.2014.09.074.

II Pietron J., Chalov S.R., Chalova A.S., Alekseenko A.V. and Jarsjö J. 2017.Extreme spatial variability in riverine sediment load inputs due to soilloss in surface mining areas of the Lake Baikal basin. Catena 152, 82-93,doi:10.1016/j.catena.2017.01.008

III Pietron J., Jarsjö J., Romanchenko A.O. and Chalov S.R. 2015. Modelanalyses of the contribution of in-channel processes to sedimentconcentration hysteresis loops. Journal of Hydrology 527, 576-589,doi:10.1016/j.jhydrol.2015.05.009

IV Pietron J., Nittrouer J.A., Chalov S.R., Dong T.Y., Shinkareva G. and Jarsjö J.2017. Sedimentation processes in the Selenga River delta: implication forsequestering sediment-bound metals. Hydrological Processes, in review

Jan Pietron

Author contributions

The contributions from listed authors are divided as follows for each article:

I TR led the writing with the help of all co-authors. JP acquired historicalhydroclimatic data and spatial data. RS acquired projected climatic data.All co-authors contributed to the data analysis and the design of the study.

II JP led the writing, complied the datasets and did the data analysis. Thewriting was assisted by all co-authors. The study and related methods weredesigned by JP with help of JJ. The field data were collected by: JP, JJ, SRCand AVA. JP acquired and analyzed spatial data. Long term hydrologicaldata were provided by SRC.

III JP led the writing, complied the datasets and did the data analysis. Thewriting was assisted by all co-authors. The study and related methods weredesigned by JP with help of JJ and SRC. All co-authors contributed to thefield data collection. JP acquired and analyzed spatial data. Long term hy-drological data were provided by SRC.

IV JP led the writing, complied the datasets and did the data analysis. The writ-ing was assisted by JJ, JAN, SRC and TD. The study and related methodswere designed by JP with help of JJ and JAN. All co-authors contributedto the field data collection. Long term hydrological data were provided bySRC.

Sediment transport from source to sink in the Lake Baikal basin

List of relevant co-authored papers, not included as part of thisthesis

1. Bring A., Asokan S.M., Jaramillo F., Jarsjö J., Levi L., Pietron J., Prieto C.,Rogberg P. and Destouni G. 2015. Implications of freshwater flux datafrom the CMIP5 multimodel output across a set of Northern Hemispheredrainage basins. Earth’s Future 3(6), 206-217, doi:10.1002/2014EF000296.

2. Chalov S.R., Jarsjö J., Kasimov N.S., Romanchenko A.O., Pietron J.,Thorslund J. and Promakhova E. V. 2015. Spatio-temporal variation ofsediment transport in the Selenga River Basin, Mongolia and Russia. En-vironmental Earth Sciences 73, 663–680, doi:10.1007/s12665-014-3106-z.

3. Chalov S.R., Thorslund J., Kasimov N., Aybullatov D., Ilyicheva E., Karthe D.,Kositsky A., Lychagin M., Nittrouer J., Pavlov M., Pietron J., ShinkarevaG., Tarasov M., Garmaev E., Akhtman Y. and Jarsjö, J. 2016. The SelengaRiver delta: a geochemical barrier protecting Lake Baikal waters. RegionalEnvironmental Change, doi:10.1007/s10113-016-0996-1.

4. Fischer S., Pietron J., Bring A., Thorslund J. and Jarsjö, J. 2017. Present tofuture sediment transport of the Brahmaputra River: reducing uncertaintyin predictions and management. Regional Environmental Change 17(2), 515-526, doi:10.1007/s10113-016-1039-7.

5. Chalov S.R., Tsyplenkov A.S., Pietron J., Chalova A.S., Shkolnyi D.I., Jarsjö J.and Maerker M. 2017. Sediment transport in headwaters of a volcaniccatchment—Kamchatka Peninsula case study. Frontiers of Earth Science11(3), 565-578, doi:10.1007/s11707-016-0632-x.

6. Jarsjö J., Chalov S.R., Pietron J., Alekseenko A. and Thorslund J. 2017. Pat-terns of soil contamination, erosion, and river loading of metals in agold mining region of Northern Mongolia. Regional Environmental Change,doi:10.1007/s10113-017-1169-6.

7. Thorslund J., Jarsjö J., Jaramillo F., Jawitz J.W., Manzoni S., Basu N.B., ChalovS.R., Cohen M.J., Creed I.F., Goldenberg R., Hylin A., Kalantari Z., KoussisA.D., Lyon S.W., Mazi K., Mård J., Persson K., Pietron J., Prieto C., QuinA., Van Meter K. and Destouni G. 2017. Wetlands as large-scale nature-based solutions: Status and challenges for research, engineering and man-agement. Ecological Engineering, doi:10.1016/j.ecoleng.2017.07.012.

Contents

1 Introduction 11.1 Introduction and problem description . . . . . . . . . . . . . . . . 11.2 Aims and scopes of the thesis . . . . . . . . . . . . . . . . . . . . . 3

2 Studied Area 5

3 Methods 93.1 Analysis of long-term hydroclimatic data . . . . . . . . . . . . . . 93.2 Analysis of the impact of mining activities on riverine sediment

load input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93.3 Analysis of in-channel sediment transport and storage dynamics 113.4 Analysis of grain size distribution of the deltaic sediment and sed-

imentation processes . . . . . . . . . . . . . . . . . . . . . . . . . . 11

4 Results 134.1 Hydroclimatic trends in the Selenga River basin . . . . . . . . . . 134.2 Impact of surface mining on riverine sediment load inputs . . . . 164.3 Transfer of sediment and role of in-channel storage . . . . . . . . 184.4 Sediment deposition on the Selenga River delta . . . . . . . . . . . 20

5 Discussion 235.1 Impact of changing hydroclimate . . . . . . . . . . . . . . . . . . . 235.2 Impact of human activities on transport patterns of sediment and

contaminants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

6 Conclusions 27

7 Future Prospects 29

Acknowledgments 31

References 33

Variables and Abbreviations

ΔS [L T−1] basin scale change in water storage

ΔSSL [L M−1] difference in suspended sediment load

Am, An, AT [L2] mining, natural and total basin area

CV [-] coefficient of variation

CVQ [-] coefficient of variation of waterdischarge

D10 , D50, D90 [L] sediment size coarser than 10%, 50%and 90% of the sediment grain sizedistribution

DB10 , DB50, DB90 [L] D10, D50, D90 of bed material (sedi-ment > 63 μm)

ET [L T−1] evapotranspiration

P [L T−1] precipitation

Peff [L T−1] effective precipitation

Pn [-] Rouse number

Q [L3 T−1] water discharge

Qmax [L3 T−1] maximal water discharge

Qmin [L3 T−1] minimal water discharge

R [L T−1] runoff

SDm, SDn, SD [M T−1] sediment load contribution from Am,An, AT

SLOm , SLOn , SLO [M T−1 L−2] soil loss from Am, An, AT

SSL [M T−1] suspended sediment load

T [ºC] air temperature

TSL [M T−1] total sediment load

U [L T−1] depth-average flow velocity

u∗ [L T−1] shear stress velocity

τb [P] boundary shear stress

CMIP5 the Coupled Model IntercompariosonProject, Phase 5

1 Introduction

1.1 Introduction and problem description

The hydrological cycle drives sediment supply and transport in fluvial systems.For example, different magnitude, intensity and type of precipitation (P, mm)can impact runoff (R, mm), hillslope erosion (Planchon et al., 2000), and con-trol transport of sediment along river channels (Peizhen et al., 2001; Berezh-nykh et al., 2012; Li and Gao, 2015). Flood events can be exceptionally impor-tant in sediment transport, due to high water discharge (Q, m3 s−1) and largemasses of conveyed material (Leenaers, 1989; Coppus and Imeson, 2002; Cáno-vas et al., 2008; Chalov et al., 2015a). In contrast, evapotranspiration (ET, mm)decreases discharge by removing water from the fluvial systems back to the at-mosphere. Hence, depending on regional hydroclimatic conditions, sedimenttransport regimes may differ considerably, which for instance impacts the totalsediment export from different river basins (Dedkov and Moszherin, 1992).

Recent changes in climate are associated with global alterations of air tem-peratures and water balance components (P, ET and R), which can impact sed-iment transport in different ways (Poesen et al., 2003; Walling and Fang, 2003;Bring et al., 2015; Fischer et al., 2017). For example, changing air temperaturescan impact evapotranspiration and vegetation growth, which in turn may influ-ence sediment supply through changed soil stability and soil erosion (Pruski andNearing, 2002). In cold environments, changed temperatures impact snow meltand related hillslope erosion (Chalov et al., 2017b) as well as permafrost distri-bution, which affects runoff and water discharge patterns in streams (Framptonet al., 2011).

Various human activities can also impact sediment transport considerably.Dam and reservoir constructions act as traps that reduce sediment delivery todownstream river sections causing channel incision (Kondolf, 1997; Walling andFang, 2003). Water use, such as irrigation, can influence sediment transport (e.g.,Gebrehiwot et al., 2015) via reduced river water discharge (Törnqvist and Jarsjö,2012). Soil erosion and associated sediment delivery to stream networks can beimpacted by changes in land-use or land-cover related to agricultural expansion,deforestation and surface mining activities (Walling, 2006; Jaramillo et al., 2015).Mining operations may also involve erratic discharge of turbid wastewater intonearby streams and altering of the natural river channels (Chalov, 2014). Moreeffective mining methods (Hooke, 2000) and increasing demands from a grow-ing population (Paull et al., 2006) have generally increased the mining impactson sediment transport in the last century. However, many mining areas are lo-cated in remote and unmonitored regions, which complicates comprehensiveassessments of environmental risks (Paull et al., 2006; Thorslund et al., 2012).Such risks may for instance be related to erosion of soils enriched with metals

1

Sediment transport from source to sink in the Lake Baikal basin

Concerns about the water quality of Lake Baikal and status of its ecosystemshave been raised (Henry and Nikanarov, 1988; Lindström-Seppä et al., 1998;Thorslund et al., 2012; Chalov et al., 2015a; Batsaikhan et al., 2017), for instance,due to increasing development of mining activities, and related sediment andmetal contamination within the lake’s largest tributary, the transboundary Se-lenga River. The development of mining significantly contributes to the eco-nomic growth in the region (Malsy et al., 2016) and its impacts on water re-sources and landscape disturbances are therefore likely going to continue inthe future (Malsy et al., 2013). Therefore, understating of the hydrological flowregime, soil erosion and sediment transport processes of the mining affected,free-flowing (undammed) Selenga River, as well as their sensitivity to climatechange, can aid in studies regarding the sediment transfer processes and its dis-tribution among sinks (e.g. river channels and floodplains; Ciszewski, 2001).This includes large sinks, such as the Selenga River delta, named a “geochemicalbarrier” (Chalov et al., 2016) between the Selenga River basin and Lake Baikal.

Effects of changing climate on hydrological flow and sediment transportregimes can be challenging to recognize and predict, for instance because riverbasins of the world are commonly dammed, such that long-term cause-and-effect relations can be masked by direct human flow regulations (see Nilssonet al., 2005). Hence, case studies of large to medium free-flowing rivers such asthe Selenga River basin, which are exceedingly rare, can aid in the general un-derstanding of the climate changes effects and related consequences. Addition-ally, studying mining affected areas in such rivers can increase our knowledge ofspreading and storage of sediment-bound contaminants under combined effectsof climate change and mining disturbances.

1.2 Aims and scopes of the thesis

This thesis considers the unregulated Selenga River system of the Lake Baikalbasin with an overall objective to determine relative contributions of mining ac-tivities and hydroclimatic change on sediment transport processes, from sourcezones in headwater areas and hillslope areas, via transport through riverchannels to net deposition areas, i.e. lakes and wetlands of the river delta. Spe-cific aims of the thesis are to:

A. Identify historical hydroclimatic trends and their possible causes.

B. Investigate the spatial variability of riverine sediment loading in the min-ing affected areas.

C. Investigate the sediment transfer and storage processes within riverinechannels as well as deltaic systems.

These aims are addressed in Papers I-IV, which focus on different areas (Figure1), variables and processes of the Lake Baikal basin, according to the following:

I. Paper I considers the entire Selenga River basin (within the grey frame inFigure 1) to address aim A. It also considers projected (2010-2039 and 2070-2099) trends in air temperature and water balance components of the Se-lenga River basin, based on CMIP5 (the Coupled Model IntercompariosonProject, Phase 5) ensemble mean of 22 individual climate models (Stockeret al., 2013).

3

Jan Pietron

II. Paper II considers the Tuul River basin (headwaters of Selenga River basin;black frame in Figure 1) to address aim B, using an observation basednested catchment approach combined with soil erosion modelling.

III. Paper III considers the downstream part of the Tuul River (green framein Figure 1) to address aim C, using a field data based sediment transportmodel.

IV. Paper IV considers the Selenga River delta (red frame in Figure 1), to ad-dress aim C, using observational data on grain size distribution of de-posited and suspended sediment combined with analytical methods toevaluate flow and transport dynamics in the delta.

4

Jan Pietron

The Selenga River basin accounts for over 82% of the area of Lake Baikaland its basin (576,500 km2). Continental climate characterizes the Selenga Riverbasin as well as the entire Lake Baikal basin. The average monthly tempera-tures of the Selenga River basin span from around -24 ºC in January to 17 ºC inJuly (CRU TS3.10/CRU TS3.10.01 climate data; Harris et al., 2014). The annualaverage precipitation within the Selenga River basin spans from 250 mm in thelower parts, up to over 600 mm in the mountainous parts. Around 80-90% ofthe annual precipitation occurs during May to September (Tulokhonov et al.,2015; Potemkina and Potemkin, 2015). Most of the high rainfall-runoff eventsnaturally occur around this period (Chalov et al., 2015a).

The landscape of Selenga River basin comprises mainly of mountainous taigaand steppe types. The basin relief ranges from 456 m a.s.l. at the mouth of theSelenga River to 3,539 m a.s.l. in the Khangal Mountains (southeast part of thebasin). Most of the basin area is underlain by permafrost that varies from iso-lated patches in the central parts to continuous in the highland parts of the basin(Brown et al., 1997; Figure 1 in Paper I). Studies have shown that the permafrostlayers are thermally unstable and vulnerable to impacts of climate change (Zhaoet al., 2010). A rapid growth of mining industries since the early 1990s (copper,molybdenum, gold and coal) has been a major cause of landscape disturbanceand metal contamination of river waters within the Selenga River basin (Robin-son et al., 2004; Thorslund et al., 2012; Batsaikhan et al., 2017). One of the largestplacer (alluvial) mining area is the Zaamar Goldfield located in the Tuul Riverbasin (headwaters of the Selenga River basin, Figure 3). The mining activitieslast approximately eight months a year, from mid-April to mid-December (Kar-poff and Roscoe, 2005).

Figure 3. Location of the Zaamar Goldfield mining area along the Tuul River (headwaters of theSelenga River basin). Locations related to Paper II: A – the Tuul River basin with the studied basinarea (red border), the location of Ulaanbaatar/Ulan Bator gauging station; B – lower part of thestudied basin (green area in A) (Figure 1 in Paper II: Pietron et al., 2017).

6

Jan Pietron

8

3 Methods

3.1 Analysis of long-term hydroclimatic data

To analyze historical changes in the hydroclimatic variables of the basin (Pa-per I), records of runoff (R, mm year−1), precipitation (P, mm year−1), and airtemperature (T, ºC) were considered. Area-average monthly P and T for period1938-2009 were acquired from the CRU TS3.10/CRU TS3.10.01 datasets (Har-ris et al., 2014). The data covered the extent of the Selenga River basin, whichwas defined using a digital elevation model (Section 2.2 in Paper I). Later area-average annual P and T for the basin were estimated. The R data were calculatedas average annual discharges (Q, m3 s−1) divided by the area of the SelengaRiver basin (with an outlet at Mostovoy station, Russian Federation; Figure 1).Additionally, evapotranspiration (ET, mm year−1) data were estimated by clos-ing the water balance ET = P − R −ΔS, where the long-term average changein water storage ΔS (mm year−1) is assumed to be zero. Additionally, T, P, Rand ET results (historical and projected) of 22 individual CMIP5 climate models(Stocker et al., 2013) over the extent of the Selenga River basin were extractedto investigate their consistency with observed historical data and examine thecharacteristics of projected future trends of the basin’s hydroclimate (2010-2039and 2070-2099; Section 2.2 in Paper I).

To identify major hydrological trends in the Selenga River basin, variousanalyses of daily discharge (Q, m3 s−1) data were performed. First, analysesof intra-annual changes of the annual maximum and minimum Q at the down-stream Mostovoy gauging station of Selenga River (Paper I) were carried outfor the period 1938-2009. Additionally, data on the relative frequency of dailyQ were compared between two periods: (1) before the recent change of aver-age discharge at the Selenga River outlet (1975-1994; Chalov et al., 2015a) and(2) after this change (1995-2014). These analyses were also carried out for theTuul River at Ulaanbaatar gauging station (Mongolia; Figure 1; data source: Pa-per II). Lastly, to investigate long-term intra-annual changes in daily dischargedata variations within the Selenga River basin, annual coefficients of variationof daily discharges (CVQ) were estimated for the Selenga River (Mostovoy, 1938-2014) and the Tuul River (Ulaanbaatar, 1945-2014).

3.2 Analysis of the impact of mining activities on riverine sed-iment load input

To investigate possible relations between riverine sediment load increases andthe areal extent of mining regions of the Tuul River basin (Paper II), the basinarea (AT , km2; red border, Figure 3A) was divided into mining areas (Am, km2)

9

Jan Pietron

and remaining natural areas (An = AT − Am; km2; Section 2.2 in Paper II). Twodifferent approaches were used to study soil loss including impacts of humanactivities within both area classes: (1) an area-weighted nested catchment ap-proach and (2) a spatially distributed soil erosion model. The former methodwas based on snapshot measurements of total sediment load (TSL, t day−1) atup to four locations along the Tuul River (green points in Figure 3) during threedifferent sampling campaigns. In contrast to method (1) that captures impactsof anthropogenic disturbances, method (2) only accounts for rainfall-runoff ero-sion of soil. A comparison of the results between both methods is expected toprovide a quantitative measure of the contribution of direct human activities tosoil losses, such as the direct discharge of turbid wastewater to the streams.

The nested catchment approach is used to interpret snapshot TSL observa-tions acquired during three field campaigns (June 2012, September 2013 andAugust 2014; Paper II), during different hydrometeorological conditions (Fig-ure 4 in Paper II). The method divides the studied part of the Tuul River basin(red border, Figure 3A) into one upper, “reference” basin (including the river’ssource, e.g. yellow basin area, Figure 5) and one lower, incremental “nested”basin (e.g. purple basin area, Figure 5). The areas of the basins were constrainedby TSL measurement points (i.e. T5b, T6, T6a, T6b; Figure 3 and Figure 5). Adetailed description of the methods can be found in Section 2.4 of Paper II. Thenested catchment approach was used to estimate soil loss (SLO, t day−1 km−2)from Am and An for each studied period of observations. Additionally, sedimentload contributions (SD, t day−1) from Am and An were approximated accordingto: SDn = AnSLOn and SDm = AmSLOm, respectively.

The soil loss and sediment load contributions from rainfall-runoff soil ero-sion were also estimated using the empirical and spatially distributed modelWATEM-SEDEM (Van Rompaey et al., 2001; Verstraeten et al., 2006; Paper II).The following data were used in the model development: soil types from fieldmeasurements, digital elevation model, satellite images and climate data (pre-cipitation). The model was developed to simulate hydrometeorological condi-tions similar to those prevailing in the August 2014-field campaign. An effectiverainfall contribution (Peff , mm month−1) corresponding to the observed dis-charge at the time of the campaign was therefore used in the model instead ofannual average precipitation values (Sections 2.2.2 in Paper II). The model was

Figure 5. Two examples (A and B) of pairs of the upper “reference” and lower, incremental“nested” catchments (Figure 3 in Paper II: Pietron et al., 2017).

10

Sediment transport from source to sink in the Lake Baikal basin

coded using the programing language R at 25 m pixel resolution. All the usedequations and the modelling routine are described in detail in Section 2.5 of Pa-per II.

3.3 Analysis of in-channel sediment transport and storage dy-namics

The natural, event-based dynamics of in-channel storage of the bed materialload and its contribution to the evolution of hysteresis in sediment load concen-tration data were investigated (Paper III). The focus was on a 14 km river reachin a part of the Tuul River located just downstream the Zaamar placer miningregion. For the analyses, a one-dimensional dynamic sediment transport modelwas developed in HEC-RAS 4.1 (Section 2.2 in Paper III). The modelled riverreach extend up to 245 km from the Tuul River mouth in the upstream direc-tion (see blue river reach in Figure 6b). A 14 km long focus reach (Figure 6c) islocated approximately in the middle of the entire model extent. This allowedthe model to simulate the sediment input to the focus reach based on the trans-port of bed material (that is also incorporated into the model) upstream it. Theriver and floodplain geometry was based on interpolation of measured cross-sectional data and the digital elevation model. To obtain reliable depth-averagestream velocities in the model, the flow model was calibrated by adjusting Man-ning’s n-values for the channel using a set of measured flow profiles acquiredin 2011 and 2012. The bed representation in the model was based on measuredbed material during the 2011 and 2012 field campaigns. Two additional scenar-ios of the model were created to test model’s sensitivity to different bed-materialinitial boundary conditions. Daily discharge data of year 2011 (Q, m3 s−1) fromthe Ulaanbaatar station were used as flow input conditions. The flow eventswere separated from the base flow using local-minima method (Sloto and Couse,1996). A more detailed description of the model development is given in Sec-tion 2.2 of Paper III. Furthermore, discharge data as well as SSL data from theMostovoy and Kabansk gauging stations at the Selenga River were comparedto analyze differences between these locations (data for period: 1981-2005). TheKabansk station is located about 20 km upstream the Selenga River delta’s apexand 80 km downstream the Mostovoy station (see Figure 1). The data used inthe analysis were acquired as a part of a collaborative project (see Section 2.2 ofPaper I).

3.4 Analysis of grain size distribution of the deltaic sedimentand sedimentation processes

The grain size distribution of sediment found at various locations within the Se-lenga River delta (Figure 4) was studied in Paper IV. These considered locationswere: marginal areas of confined channels or areas of subaqueous levees for-mation, for simplicity denoted submerged banks (18 samples); subaerial sandbars (4 samples) as well as various wetlands and water bodies characterized bydifferent level of connectivity with the main channels (7 samples in total). Thelatter group involve floodplain lakes that receive water and sediment duringhigh discharge conditions (3 samples), marshlands that are well connected withthe main channels also under conditions of lower discharges (2 samples) and the

11

Jan Pietron

Figure 6. (a) Location of the area from Paper III in the Selenga River basin (b) the downstreamTuul River and the modelled reach extents with the location of the focus reach, (c) the exact extentof the focus reach (Figure 1 in Paper III: Pietron et al., 2015).

newly formed subdelta front (2 samples). Additionally, the study in Paper IV in-vestigated which flow conditions are needed in the delta channels to transportsediment outside of the channel margins. It was tested if bed material sediment(> 63 μm) can be conveyed outside of the channel in suspended transport mode(see Section 3.2 of Paper IV) based on estimates of the dimensionless Rouse num-ber (Pn). If the estimated Pn-values for the considered sediment particle sizeare lower than the critical Rouse number Pn* = 2.5, it means that the sedimentstart to be transported with the suspended load (Middleton and Southard, 1984;Huston, 2014). If estimated Pn-values are higher than Pn*, it indicates that thesediment is most likely transported with the bed load (Lynds et al., 2014). Thetested bed material ranges were based on grain size statistics of the sampled lo-cations. The considered flow characteristics were approximated from indepen-dently reported minimum, median and maximum boundary shear stress values(τb, Pa) for bankfull flow conditions (Dong et al., 2016) and from measured dis-charges (Q) and depth-average flow velocities (U , m s−1) at different channelcross-sections within the delta (see Table 2 in Paper IV).

12

4 Results

4.1 Hydroclimatic trends in the Selenga River basin

The results from the analysis of long-term hydroclimatic data show that the an-nual average T of the Selenga River basin increased by 1.6 ºC (0.022 ºC year−1

on average) over the studied period of 1938-2009 (Figure 7a). The temperaturemostly increased gradually, however in the late 1980s to the mid-1990s T in-creased more rapidly. The largest and smallest changes of seasonal T betweenthe periods 1961-1985 and 1986-2009 occur in winter (+1.32 ºC) and autumn(+0.85 ºC), respectively. The annual mean values of the water balance compo-nents (P, ET and R) show relatively small temporal variability during the stud-ied period (Figure 7b). For example, a comparison of the data between the two20-year periods (1961-1980 and 1986-2005) shows that the annual mean P in-crease by 7.8 mm and the annual mean R decrease by 1.4 mm. The biggest de-crease in seasonal R occur during summers. An increased annual average ET by9.2 mm can have contributed to the above-described decrease in R. Accordingto the water balance assessments, on average 85% of P is lost through ET duringthe entire studied period (1938-2009).

Additionally, a comparison with the historical observed and modelled Tdata (Paper I; based on CMIP5 ensemble mean of 22 individual climate mod-els; Stocker et al., 2013) show reasonable consistency. The ensemble mean ofthe projected future T in the Selenga River basin shows continued warming ofthe climate. However, a comparison of the historical modelled and measureddata of the water balance components (P, R and ET) between the modelled andmeasured data show large differences. This decreases the reliability of the cor-responding projections that show long-term increasing values of all the waterbalance components (see Figure 4 in Paper I).

The average observed Q in the Selenga River basin (Mostovoy station) forthe period 1938-2009 is 855 m3 s−1. A comparison of mean inter-annual dailydischarges for three periods (1938-1961, 1962-1985 and 1986-2009) shows a re-cent decrease in discharges for the period between April and October (Figure7c). The annual minimum daily discharge has increased, whereas the maximumdaily discharge has decreased since 1938 (Figure 7d and 7e, respectively). Acomparison of the daily Q frequency between the periods 1975-1994 and 1995-2014 shows major changes in the Selenga River’s hydrological flow regime (Fig-ure 8a). The average daily discharge, for instance decreased from 893 m3 s−1

to 725 m3 s−1. Moreover, during the last 20 years the frequency of moderatedischarges (Q = 750 - 1250 m3 s−1) increased and high discharges (Q > 1350m3 s−1) decreased, both by about 10 percentage points (Paper IV). Long-termdaily discharge data show quite unchanged variability of flow during the pe-riod 1938-1994, with 10-year running average annual coefficients of variation of

13

Jan Pietron

P

R

Tem

pera

ture

(C°)

0

-1

-2

-3

-4

-5

-6

P, R

, ET

(mm

yea

r-1)

600550500450400350300 ET

T

0

(a)

(b)

(c)

(d)

(e)

25020015010050

0

Qm

ax (m

3 s-1

)Q

min

(m3

s-1)

Q (m

3 s-1

)

600

500

400

300

200

100

0

0

2

4

6

8

10

12

1500

500

2000

2500

1000

y = -0.74 + 267R2 = 0.03

y = 0.03 + 4.3R2 = 0.15

Jan

Feb

Mar Apr

May Ju

n

Jul

Aug

Sep

Oct

Nov

Dec

1938-19611962-19851986-2009

1938

1943

1948

1953

1958

1963

1968

1973

1978

1983

1988

1993

1998

2003

2008

1938

1943

1948

1953

1958

1963

1968

1973

1978

1983

1988

1993

1998

2003

2008

1938

1943

1948

1953

1958

1963

1968

1973

1978

1983

1988

1993

1998

2003

2008

Figure 7. Long term patterns in mean annual (a) temperature (T), (b) precipitation (P), evapo-transpiration (ET) and runoff (R), (d) minimum (Qmin) and (e) maximum (Qmax) daily dischargewithin a year (Mostovoy station, see Figure 1) for the period 1938-2009 in the Selenga River basin.Dashed or dotted lines indicate annual values and solid lines indicate 10-year running average.Graph (c) presents changes in daily discharge (Q) patterns over the year for the periods 1938-1961, 1962-1985 and 1986-2009 (Mostovoy station) (modified from Figure 2 and Figure 3 in PaperI: Törnqvist et al., 2014).

daily discharges (CVQ) ranging between 0.88 and 0.95 (Figure 9a). After 1995,the CVQ drops and the 10-year annual average CVQ is 0.81 for the most recentdecade (2005-2014).

Daily discharge frequencies in the Tuul River (Ulaanbaatar station) for thetwo different periods (1975-1994 and 1995-2014) are presented in Figure 8b. Theaverage daily discharges decreased from 34 m3 s−1 to 16 m3 s−1 during themost recent 20 years. Low discharges (0.0-5.0 m3 s−1) are most frequent, oc-curring 49% of the time (1995-2014) which corresponds to 177 days per year.The relative frequency of discharges between 0.0-35 m3 s−1 increased from 76%to 86% during the period of 1995-2014. The frequency of discharges between35-55 m3 s−1 did not change between the periods, and are around 7.0%. Thefrequency of discharges greater than 55 m3 s−1 decreased from 18% to 6.6%.Long-term daily discharge data show changing variability of flow during theperiod 1938-1988, with 10-year running average annual coefficients of variationof daily discharges (CVQ) ranging between 1.4 and 2.0 (Figure 9b). After 1989,the 10-years running average CVQ drops from to 1.4 and slightly increases toalmost 1.5 in the most recent decade (2005-2014).

14

Sediment transport from source to sink in the Lake Baikal basin

2014

2013

10.7% 34.6% 42.8%52.3%

0

100

200

300

400

500

600

700

800

upstream T5b upstream T6 upstream T6a upstream T6b

33.3% 68.9% 75.8%82.1%

0

100

200

300

400

500

600

700

800

upstream T5b upstream T6 upstream T6a upstream T6b

natural sources

mining sources

observed

2012 from mining areas

from natural areas

3.4% 13.3% 17.9% 24.1%

0

100

200

300

400

500

600

700

800

upstream T5b upstream T6 upstream T6a upstream T6b

SD

or T

SL

(t da

y-1)

SD

or T

SL

(t da

y-1)

SD

or T

SL

(t da

y-1)

Figure 10. Estimated approximate contribution from natural and mining areas to the total sedi-ment load (TSL, t day−1) upstream of the measurement locations; red dots – observed TSL; Per-centages indicate the relative contribution from mining areas to the estimated TSL (Figure 5 inPaper II: Pietron et al., 2017).

the Tuul River basin is 699 t day−1 for higher flow (August 2014) down to 255t day−1 for the low flow conditions (June 2012). The contribution from miningareas (SDm) to those values range from 24% to 82%, respectively (Figure 10).

The results of the soil erosion model for August 2014 considering the lowerpart of the studied basin (green area, Figure 3A) show that on average the min-ing areas are characterized by larger net erosion than the natural areas (Figure11). This result is mainly due to poor vegetation cover in the mining areas (seeFigure 6 in Paper II). On the other hand, net deposition pools that are locatedclose to the mining erosion hot spots limit transfer of a considerable part of theeroded sediment to the streams (Figure 11). According to the balance betweennet erosion and deposition, 18.2 t day−1 and 1.30 t day−1 of the sediment origi-

17

Jan Pietron

Figure 11. Modelled net erosion and deposition patterns in the lower part of the studied basin(see Figure 3B) and its selected parts in large scales (August 2014). Light grey – border of themining areas (Figure 7 in Paper II: Pietron et al., 2017).

nate from natural and mining areas, respectively. A comparison of these resultswith the results of the observation-based approach (August 2014) shows that thesediment contributions from the natural areas of the lower part of the studiedbasin are similar (15.2 t day−1) to the modelled ones. However, the observedcontribution from the mining areas was 116 times higher (151 t day−1) than thecorresponding model result (see Figure 8 in Paper II). This suggests that a con-siderable part of the mining impact in the Zaamar Goldfield comes from directhuman activities (that are not included in the soil erosion model), such as dis-charge of turbid waste waters, extensive channelization and destabilization ofrivers’ banks (Paper II).

4.3 Transfer of sediment and role of in-channel storage

Five major flow events (the ones that doubled current base flow) and eleven mi-nor events can be distinguished in the daily discharge data from 2011 (Figure12). The results of the sediment transport model show that the overall changein the net storage of sediment within the 14 km long channel of the focus reach(Figure 6c) is -14 kt for the studied period. Additional scenarios show that theresults of the model are sensitive to the initial bed material conditions. In par-ticular, in case of initial abundance or scarcity of the finest bed material thesame hydrograph (of the considered hydrological period) can yield negative

18

Jan Pietron

0.025 0.035

0.1

0.2

0.5

1.0

2.0

5.0

10.0

Order 3

u* (m s 1)

Rou

se N

umbe

r,Pn

(-) Pn*

_

___

_BQ-408

0.000 0.015 0.030

0.1

0.2

0.5

1.0

2.0

5.0

10.0

Order 4

u* (m s 1)

Pn*

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

BQ-402

BQ-404

BQ-409

0.010 0.020 0.030

0.1

0.2

0.5

1.0

2.0

5.0

10.0

Order 5

u* (m s 1)

Pn*

_

_

_

_

_

_

_

_

_

_

_

_

BQ-401

BQ-403

0.005 0.015

0.1

0.2

0.5

1.0

2.0

5.0

10.0

Order 7

u* (m s 1)

Pn*

K1 BQ-405

BQ-410

D10 D50 D90

0.065 0.080

0.1

0.2

0.5

1.0

2.0

5.0

10.0

Order 2

u* (m s 1)

Rou

se N

umbe

r,Pn

(-) Pn*

bed

load

susp

ende

d lo

adbe

d lo

adsu

spen

ded

load

0.02 0.06

0.1

0.2

0.5

1.0

2.0

5.0

10.0

Order 3

u* (m s 1)

Pn*

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

0.02 0.06

0.1

0.2

0.5

1.0

2.0

5.0

10.0

Order 4

u* (m s 1)

Pn*

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

0.02 0.06

0.1

0.2

0.5

1.0

2.0

5.0

10.0

Order 5

u* (m s 1)

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

Pn*

0.010 0.025

0.1

0.2

0.5

1.0

2.0

5.0

10.0

Order 7

u* (m s 1)

Pn*

a

b

Figure 15. Rouse numbers (Pn) for different channel orders in the Selenga River delta for shearvelocity values u∗ (m s−1) representing: (a) bankfull flow conditions (based on τb data from Donget al., 2016) and (b) in situ measurements of depth-average flow velocities (U , m s−1). The pre-sented Pn-values are for the average and extreme (whiskers) D10, D50 and D90 of the sand-gravelmaterial (DB10, DB50 and DB90) of submerged bank samples within different channel orders(modified from Figure 7 in Paper IV: Pietron et al., in review).

22

5 Discussion

5.1 Impact of changing hydroclimate

Results of the long-term discharge (Q) data analysis (Paper I) indicate that thereare considerable trends in the hydrological regime of the Selenga River. In par-ticular, during the period 1938-2009 the annual maximum daily flow decreased,whereas the annual minimum daily flow increased. This result is consistent withexpected impacts of large-scale permafrost thaw (Frampton et al., 2011) withinthe basin. The thawing permafrost in the Lake Baikal basin is likely a result ofchanging climate that is depicted by the rate of increasing temperature 0.022 ºCyear−1, which is faster than the global average 0.012 ºC year−1 (Stocker et al.,2013). The analysis of daily Q presented in this thesis shows that the frequencyof high Q decreased and the frequency of moderate as well as low Q increasedboth in the upstream and downstream parts of the Selenga River basin (Figure8). The variation of daily Q decreased in the most recent 20 years (1995-2014)too, as depicted by CVQ results in Figure 9. Therefore, despite the discussedeffect of permafrost thaw on the stream flow patterns, additional factors mighthave contributed to these relatively large changes in the hydrological regime ofthe Selenga River basin. These include a drought that started in mid-1990s in theupper parts of the basin. Climate analyses of the last 400 years showed that suchdry periods are usual for the semi-arid climate of the northern Mongolia (Daviet al., 2013). However, the ongoing drought is the warmest observed (tree-ringsanalysis) since over the last millennium (Pederson et al., 2014). Additionally,the effects of the dry conditions can be amplified by rapidly increasing waterdemands (for mining and irrigation, Priess et al., 2011; Davi et al., 2013; Malsyet al., 2016). Land-use changes in the Russian part of the Selenga River basin,such as afforestation (Bazhenova and Kobylkin, 2013) could also contribute tothe changes in observed Q (Hundecha and Bárdossy, 2004). Overall, except forthis recent drought, there are no clear trends in the annual precipitation (P) overthe last century in the Selenga River basin (Figure 7b).

The above-mentioned, recent P decrease over the Selenga River basin couldbe linked with shifts in the summertime atmospheric circulation over East Asia(Berezhnykh et al., 2012). Parallel studies show that there are recent shifts in therainfall patterns of northern Mongolia too. In particular, the number of long, lowintensity rainfalls (up to 3 days of ≤ 2.5 mm hr−1) has decreased, whereas thefrequency of short high intensity and patchy rain events has increased (coupleof minutes of > 7.6 mm hr−1; Marin, 2010; Goulden et al., 2016). Such changescan impact the generation of runoff and soil erosion patterns in semi-arid areas(Mohamadi and Kavian, 2015). High intensity patchy rainfalls can lead to rapidsurface runoff generation and flash floods in small catchments (Cudennec et al.,2007). However, such precipitation patterns may not have the same impact on

23

Jan Pietron

generation of peak flow events in larger river basins due to the scattered andlocal character of these intense rainfall events (Bracken and Croke, 2007).

The results of this thesis show that the number and magnitude of high-flowevents are important for the transfer and exchange patterns of a river’s fine bedsediment (0.062-0.25 mm; Paper III). Within a year, more than 80% of the net bederosion happened during about 12% of the time, during the five highest flowevents (E1-E5) of the year. The three first high flow events of the year (E1-E3;Figure 12) were important for remobilisation of the in-channel sediment. Thelatter two events, which were the highest in the studied period (E4-E5; Figure12), did not remobilize similar amounts of sediment, partly because supply lim-ited conditions emerged in the channel. Similar patterns have been observedin other rivers, e.g. in Mexico (Hudson, 2003) or upper Mississippi (Magilli-gan et al., 1998) were the first events in hydrological periods exhausted thein-channel sediment storage. Changes in magnitude and number of high flowevents can theoretically alter the dynamics of the bed sediment transfer. In theSelenga River case, broader annual hydrographs with relatively lower peaks canmake the inter-event character of sediment erosion more pronounced, and de-position less pronounced. Such replenishment of the bed sediment storage canoccur even during time of lower discharges or flow events that according to thehere discussed model account for more than 30% of the total net deposition inthe studied period (Figure 13d).

A comparison of rating curves for the average annual suspended load (SSL)between the Mostovoy and the Kabansk gauging stations (approximately 80 kmlong river reach between the stations, Figure 1) in the most downstream SelengaRiver reach, shows significantly increased SSL in Kabansk (situated closer tothe Selenga River delta) at conditions of greater annual average discharge (Fig-ure 14b). A relatively clear relation between the annual maximum discharges(Qmax) and the difference in SSL between the stations (ΔSSL; Figure 14c) indi-cates that the increased suspended load at the Kabansk station is linked to themagnitude of Qmax. Extreme events can also resuspend the sediment usuallytransported as bed load, and thus contribute to the observed differences in theSSL. An important role of flow events for the transfer of sediment and bed stor-age dynamics in the Selenga River system was previously seen also in its head-water parts through the model study of the Tuul River (Paper III). The reachbetween the Mostovoy and Kabansk stations can have similar storage functionsand dynamics as the focus reach studied in Paper III.

The hydroclimatic changes discussed above are important for understand-ing (potential) changes in the sedimentation processes of the Selenga River delta.The delta is characterized by a considerably decreased stream power due to flowpartitioning between up to eight orders of channels (Dong et al., 2016). Thiscauses a non-linear deposition of sediment within the channel network, startingwith the coarsest sediment load (gravel) in channels of low orders, eventuallygiving way to sand and silt in the terminal channels at the interface betweenthe delta and Lake Baikal (Ilyicheva et al., 2015; Dong et al., 2016). The resultsof this thesis suggest that part of the bed material load can also be transportedoutside of the channels with suspended load even during moderate flow condi-tions (Figure 15; Paper IV), characterized by relatively low average shear veloc-ities (u∗) in the distributary channels (Figure 8 in Paper IV). Eventually, the bedmaterial (mainly between 70-148 μm; Paper IV) together with the wash load (siltand clay, < 63 μm) can be trapped in submerged banks and marshlands within

24

Sediment transport from source to sink in the Lake Baikal basin

the backwater zone (∼ 9 km from the delta’s outlet, Dong et al., 2016). Such stor-age process in the wetlands and in bank areas can partly explain the observeddecreases in sediment concentrations along channels of the Selenga River delta(Chalov et al., 2016, 2017b). On the other hand, the changed hydrological regimewith lower median flows of the Selenga River (Figure 8a) is currently suppress-ing the connectivity between the main channel and floodplain water bodies thattrap mainly silt and clay fractions.

5.2 Impact of human activities on transport patterns of sedi-ment and contaminants

Results of this study (Paper II) suggest that mining impacts on sediment trans-port in the Tuul River originate mostly from direct inputs, such as input ofsediment-loaded wastewaters to the streams or input from poorly maintainedsettling ponds, which lead to discharge of fine-grained material into the TuulRiver (Farrington, 2000; Stubblefield et al., 2005). Destabilization of banks, min-ing too close to the river (Farrington, 2000) as well as dredging of the channels(Chalov et al., 2015a) can also contribute to the mining impact. The latter canfurthermore lead to spreading of bed incision and channel degradation outsideof the mined river reaches (Kondolf, 1997; Simon and Rinaldi, 2006). Identifica-tion of such key processes and assessment of their effects is important for estab-lishing better management strategies (Rinaldi et al., 2005) to mitigate adverseimpacts on downstream water systems (MEGD, 2012; Chalov et al., 2015a,b;Jarsjö et al., 2017). Due to the here identified high impact of wastewaters andponds, it can be expected that better management solutions of the mining areas(e.g. Farrington, 2000) can aid in mitigation of their adverse impact on sedimenttransport in the Tuul River.

Present results also showed that the sediment input from mining areas are es-sentially independent of changing hydrometeorological conditions and rainfall-runoff soil erosion processes. Therefore, during dry conditions ( Q ∼ 12 m3 s−1)the mining activities at the Zaamar Goldfield can contribute with as much as80% of the TSL at the site. The highest considered discharges in the study wereup to bankfull (∼ 60 m3 s−1), which were close to the August average at theUlaanbaatar station (Figure 2B in Paper II; Q ≤ 60 m3 s−1 occurred on averagefor about 94% in the recent decades in the Tuul River, see Figure 8b). Duringsuch conditions around 24% of the TSL in the Tuul River can originate from thehuman activities (mostly the direct ones) at the mining areas (Figure 10). Miningareas frequently exhibit elevated concentrations of many metals in soils com-pared to non-mining areas (Jarsjö et al., 2017). These metals can hence enter theriver through the mining activities and be transported with the sediment load asshown above. Thorslund et al. (2016) observed that concentrations of mining-related metals in suspended sediment increased during lower flow conditionsin the Zaamar Goldifeld area. The present results suggest that this may be aneffect of increased relative sediment load contribution from mining sources dur-ing dry conditions (Figure 10). In this case, the metals are likely transportedwith fine-grained fractions (silt and clay, < 63 μm), which can be transported insuspension even during low discharge conditions (see Einstein et al., 1940). Onthe other hand, the concentration of some metals (i.e. Cu and Zn) were higherduring periods of increased discharge (Thorslund et al., 2016). Since these met-als were found to bond with coarser sediment fractions than silt and clay in the

25

Jan Pietron

Selenga River system (e.g. fine sands, ∼ 63-250 μm; Lychagin et al., 2017), theylikely are transported with bed material load. In that case, high flow eventswere shown to be important in mobilization and transient storage of associatedsediment and sediment-bound metals (Paper III). Observations in rivers world-wide show that high discharge events are important in “flushing” of river chan-nels from contaminated sediment in the human impacted areas (Davide et al.,2003). The sediment transported during such flows is likely redistributed be-tween downstream sinks (floodplains and channel bed).

Paper IV of this study shows that sediment deposits and sediment loads inthe Selenga River delta, the final sink of the sediment in the basin, are character-ized by high cohesion. The median diameter of sediment in submerged banksand of suspended sediment were D50 = 37 μm and D50 = 32 μm, respectively(Paper IV). Potentially, the fine particles that reach the delta can be trapped inthe wetlands and banks within the backwater zone (Paper IV). This is consistentwith studies concluding that the Selenga River delta plays a prominent role inthe storage of sediment and associated metals (Khazheeva et al., 2008; Chalovet al., 2016, 2017a). Furthermore, the gradual deposition of sediment on thechannel beds is another process that can cause decreasing metal concentrationsin the delta (Chalov et al., 2016). An effect of such a process would be that thegrain size of the bed sediment should change with distance and degree of flowpartitioning in the delta, which has been observed in the case of the Selengadelta (Dong et al., 2016). Although these sedimentation processes in the deltawill decrease the amount of the potential metal pollution entering Lake Baikalwith suspended material, they may impact biota of the delta itself (Plyusnin andZhambalova, 2014) and may eventually also be further transported as dissolvedconstituents into Lake Baikal. Taken together, to quantify the impact of the min-ing on the large scale sediment transport, the distribution and delivery to thedelta’s wetlands, future studies in the Selenga River basin should focus on iden-tification and tracking of the mining sediment (e.g. via fingerprinting analyses).Such efforts can more generally aid in understanding the sediment associatedmetal contamination of the environment by mining (see Thorslund et al., 2012,2016).

26

6 Conclusions

In this thesis, data analysis and modelling approaches are used to investigatehow a changing climate and human activities alter transport of sediment withinthe Lake Baikal basin. The main conclusions are:

• During the period 1938-2009, the annual maximum daily flow in the Se-lenga River basin has decreased, as well as the annual number of highflow events, whereas the annual minimum daily flow has increased. Thesechanges in discharge characteristics are consistent with expected impactsof basin-scale permafrost thaw.

• In the past 20 years, drought conditions that may be linked with shiftsin the summertime atmospheric circulation over East Asia have causedconsiderable decreases in water and sediment discharges in the basin.

• Under high (bankfull) discharge conditions in mining-impacted hillslopeareas of the Selenga River basin headwaters (the Tuul River basin, Mongo-lia), about 24% of the sediment input to the river originated from the placermining areas that occupied roughly 0.12% of the considered hillslope areas(57× 103 km2).

• The sediment input from the placer mining areas were observed to be con-stant regardless of hydrometeorological conditions, which under low dis-charge conditions lead to a dominance (82%) of sediment originating fromthe placer mining areas, most probably entering the river through wastew-ater discharge or malfunctioning settling ponds.

• Metal-enriched sediment from mining areas can hence dominate river wa-ters under low flow conditions, which implies that if discharge will con-tinue to decrease in the Tuul and Selenga rivers (and in other mining re-gions of the world) increased riverine metal concentrations may be one ofthe consequences.

• Regarding in-channel storage and transfer of the river bed-material (sedi-ment with grain size > 63 μm), recorded data and modelling results showedthat changes in magnitude and number of high-flow events can consider-ably influence the dynamics of bed sediment transfer. Given also the con-clusive results regarding on-going changes in peak flow characteristics, itis likely that in-channel sediment transport characteristics may now be un-der conditions of considerable change.

• Under present conditions of extended drought, less sediment may havebeen distributed over the floodplain wetlands in the Selenga River delta.However, our field-data based modelling shows that, despite these changes

27

Jan Pietron

in ambient conditions, sediment can still be transported to, and trappedwithin, the banks and water bodies located in the backwater zone of theSelenga River delta. This can aid banks and levees stabilization, supportthe development of wetlands and foster net sedimentation.

28

7 Future Prospects

As shown in this thesis, ongoing climatic changes have likely contributed toshifts in hydrological characteristics and related sediment transport characteris-tics in the Selenga River system in recent decades (Sections 4.1 and 5.1). In thefuture, it is conceivable that the flow regime of the Selenga River will be affectedby dam and reservoir construction, for instance due to plans of the Mongoliangovernment to construct a cascade of hydropower dams in the region (see, e.g.Chu et al., 2010). The construction of dams and related flow changes can affectthe sediment transport processes and the ecology of river systems in differentways and to various degrees (Ligon et al., 1995; Kondolf, 1997; Walling and Fang,2003). Hence, an essential first step in taking well-informed management deci-sions regarding such large-scale flow regulation projects in the Selenga Riverwould be to systematically investigate the potential impacts on sediment andcontaminant transport along the river, all the way to the delta and the deltaicwetlands. Present results, including the here developed models and analysismethods, may provide a useful basis for some of the needed impact assessments.

The results showing that mining can have large impacts on the compositionof transported sediment (in particular under low flow conditions; Section 4.2)call for more detailed investigations of the actual origin of sediment found alongtransport pathways in the Selenga River system. This should include refined as-sessment of contributions from multiple mining areas to downstream sedimentloads and metal pollution. Methodologically, this can for instance be achievedby conditioning transport models such as HEC-RAS with information from fin-gerprinting analysis methods, including isotope analyses (Viparelli et al., 2013),geochemical assessments and composite approaches (Collins et al., 1997; Milleret al., 2015; Tiecher et al., 2016). Such efforts would also increase our knowledgeregarding the role of the Selenga River delta in accumulation and storage of themining-originating sediment and contaminants.

This thesis provided an example of how particle size can control whereparticle-bound metals can be sequestered in river deltas (Section 4.4). Althoughfine sediment fractions generally contain higher concentrations of most metalscompared to coarse sediment fractions (de Groot et al., 1971), recent studies haveshown that some metals in the Selenga River waters consistently bond prefer-entially with coarse sediment fractions (Lychagin et al., 2017). This may be forvarious reasons such as the specific mineral composition of the coarser sediment,or the close proximity of a sampled site to a source of metal contamination withparticular characteristics (Whitney, 1975; Tessier et al., 1982; Singh et al., 1999).To better understand the ultimate fate of different sediment-bound metals, de-tailed studies are needed regarding the generality and predictive capabilities ofmetal-specific relations between particle size and metal concentrations.

29

Jan Pietron

30

Acknowledgments

First of all, I would like to thank my main supervisor Jerker Jarsjö for his kindguidance and selfless support during my studies. Thank you for your openears to my inquiries and ideas. I am very grateful that I could always count onyou. Despite losing your watch during our first field work expedition, you havealways found the time for me. You are an amazing teacher and scientist!

During my PhD studies, I have met many fascinating people, some ofwho have largely contributed to this work. First, I would like to thank my co-supervisor Gia Destouni for her advice and for sharing her vast knowledge inhydrology. Furthermore, I would like to thank Sergey Chalov for his terrific ad-vice, time and great hospitality. Working in the field by his side was always aninspiring and valuable experience. I hope for more "czajniki" and "herbatniki"in the future. I would also like to thank Jeff Nittrouer for sharing his in-depthknowledge of sediment transport processes, his wise advice and generous hos-pitality. Lastly, I would like to thank all the others whom I have met during mystudies and field campaigns, especially: Galya, Tian, Katya, Rebecka, Annya,Alexandra, Alexey, Elena and Max.

My time at the department would have been much more difficult withoutSusanna Blåndman. I would like to thank her for helping me in all my issuesrelated to the PhD studies and adapting in Sweden. I would also like to thankHelle Skånes, Peter Jansson and Stefano Manzoni who helped to organize mythesis work. Special thanks go to Małgosia (my mum) and Paweł Milewski whohelped me writing the Polish version of the abstract.

Big thanks go to all my colleagues and friends who I met during my ad-ventures at the Department of Physical Geography, in particular: Simon, Lea,Josefin, Juri, Lucile, Ida, Fernando, Sandra, Niels, Alex, Norris, Stefan, Marika,Benoît, Beny, Steve, Arvid, Ylva, Amelie, Valentin, Natacha, Moa, Karin, René,Antonis, Emelie, Carmen, Linus, Romain and Guillaume. You made my paththrough the PhD work and studies much easier, in both scientific and socio-emotio-economical way!

I would like to thank all my friends outside the department walls, who havestood by me and supported me when needed, especially: Nico, Marta, Hoon,Alex, Justine, Przemek, Josefin K., Matilda, Agatka, Jas and Linn. Special thanksgo to the Tellström family, without whom I would not be here in Sweden. Thankyou for your great help and hospitality!

Lastly, I want to thank my parents Małgosia Litak-Pietron and Krzys Pietronas well as my dear sisters Gusia and Hania. You have always supported me, andI know that I can always count on you despite the distance in space and time.You have been the best, and you will always be! Kocham Was!

31

Jan Pietron

Financial support

This work was funded through the Swedish Research Council Formas (project2012-790). Travel and field work were additionally granted from the EU 7thframework programme Marie Curie Action – International Research Staff Ex-change “FLUMEN” (Grant Agreement Number: PIRSES-GA-2012-318969), theK & A Wallenberg foundation, the Climate Research School at the Bolin Centrefor Climate Research and the Faculty of Science at Stockholm University.

32

References

Axtmann EV, Luoma SN. 1991. Large-scale distribution of metal contamination in thefine-grained sediments of the Clark Fork River, Montana, U.S.A. Applied Geochem-istry 6:75–88. doi:10.1016/0883-2927(91)90064-V.

Batsaikhan B, Kwon JS, Kim KH, Lee YJ, Lee JH, Badarch M, Yun ST. 2017. Hydro-chemical evaluation of the influences of mining activities on river water chemistry incentral northern Mongolia. Environmental Science and Pollution Research 24:2019–2034. doi:10.1007/s11356-016-7895-3.

Bazhenova OI, Kobylkin DV. 2013. The dynamics of soil degradation processes withinthe Selenga basin at the agricultural period. Geography and Natural Resources34:221–227. doi:10.1134/S1875372813030050.

Berezhnykh TV, Abasov NV, Mordvinov VI. 2012. Changes in the Summertime Atmo-spheric Circulation Over East Asia and Formation of Long-Lasting Low-Water Peri-ods Within the Selenga River Basin. Geography and Natural Resources 33:223–229.doi:10.1134/S1875372812030079.

Bracken LJ, Croke J. 2007. The concept of hydrological connectivity and its contributionto understanding runoff-dominated geomorphic systems. Hydrological Processes21:1749–1763. doi:10.1002/hyp.6313.

Bring A, Asokan SM, Jaramillo F, Jarsjö J, Levi L, Pietron J, Prieto C, Rogberg P, DestouniG. 2015. Implications of freshwater flux data from the CMIP5 multimodel outputacross a set of Northern Hemisphere drainage basins. Earth’s Future 3:206–217. doi:10.1002/2014EF000296.

Brown J, Ferriansm O, Heginbottom J, Melnikov E. 1997. Circum-Arctic Map of Per-mafrost and Ground Ice Conditions. Boulder: Natlional Snow and Ice Data Center,World Data Centert for Glacioliology.

Brunello AJ, Molotov VC, Dugherkuu B, Goldman C, Khagamanova E, Strijhova T, Sig-man R. 2006. Lake Baikal. Technical report. Tahoe-Baikal Institute. South Lake Tahoe.

Cánovas CR, Hubbard CG, Olías M, Nieto JM, Black S, Coleman ML. 2008. Hydrochem-ical variations and contaminant load in the Río Tinto (Spain) during flood events.Journal of Hydrology 350:25–40. doi:10.1016/j.jhydrol.2007.11.022.

Chalov SR. 2014. Effects of placer mining on suspended sediment budget: case study ofnorth of Russia’s Kamchatka Peninsula. Hydrological Sciences Journal 59:1081–1094.doi:10.1080/02626667.2014.903330.

Chalov SR, Bazilova VO, Tarasov MK. 2017a. Modelling suspended sediment distribu-tion in the Selenga River Delta using LandSat data. Proceedings of the InternationalAssociation of Hydrological Sciences 375:19–22. doi:10.5194/piahs-375-19-2017.

Chalov SR, Jarsjö J, Kasimov NS, O Romanchenko A, Pietron J, Thorslund J, Pro-makhova EV. 2015a. Spatio-temporal variation of sediment transport in the SelengaRiver Basin, Mongolia and Russia. Environmental Earth Sciences 73:663–680. doi:10.1007/s12665-014-3106-z.

Chalov SR, Shkol’nyi DI, Promakhova EV, Leman VN, Romanchenko AO. 2015b. For-mation of the sediment yield in areas of mining of placer deposits. Geography andNatural Resources 36:124–131. doi:10.1134/S1875372815020031.

Chalov SR, Thorslund J, Kasimov N, Aybullatov D, Ilyicheva E, Karthe D, Kositsky A,Lychagin M, Nittrouer J, Pavlov M, Pietron J, Shinkareva G, Tarasov M, Garmaev E,Akhtman Y, Jarsjö J. 2016. The Selenga River delta: a geochemical barrier protectingLake Baikal waters. doi:10.1007/s10113-016-0996-1.

33

Jan Pietron

Chalov SR, Tsyplenkov AS, Pietron J, Chalova AS, Shkolnyi DI, Jarsjö J, Maerker M.2017b. Sediment transport in headwaters of a volcanic catchment - Kamchatka Penin-sula case study. doi:10.1007/s11707-016-0632-x.

Chu J, Lee C, Janchivdorj L, Oktyabrevich GB, Park S, Mun H. 2010. Integrated WaterManagement Model on the Selenge River Basin - Development and Evaluation of theIWMM on the SRB (Phase 3) (Phase I). Technical report.

Ciszewski D. 2001. Flood-related changes in heavy metal concentrations within sed-iments of the Biała Przemsza River. Geomorphology 40:205–218. doi:10.1016/S0169-555X(01)00044-7.

Collins AL, Walling DE, Leeks GJL. 1997. Source type ascription for fluvial suspendedsediment based on a quantitative composite fingerprinting technique. Catena 29:1–27. doi:10.1016/S0341-8162(96)00064-1.

Colman SM. 1998. Water-level changes in Lake Baikal, Siberia: tectonism versus climate.Geology 26:531–534. doi:10.1130/0091-7613(1998)026<0531:WLCILB>2.3.CO;2.

Coppus R, Imeson AC. 2002. Extreme events controlling erosion and sediment transportin a semi-arid sub-Andean valley. Earth Surface Processes and Landforms 27:1365–1375. doi:10.1002/esp.435.

Cudennec C, Leduc C, Koutsoyiannis D. 2007. Dryland hydrology in Mediterraneanregions—a review. Hydrological Sciences Journal 52:1077–1087. doi:10.1623/hysj.52.6.1077.

Davi NK, Pederson N, Leland C, Nachin B, Suran B, Jacoby GC. 2013. Is eastern Mon-golia drying? A long-term perspective of a multidecadal trend. Water ResourcesResearch 49:151–158. doi:10.1029/2012WR011834.

Davide V, Pardos M, Diserens J, Ugazio G, Thomas R, Dominik J. 2003. Characterisationof bed sediments and suspension of the river Po (Italy) during normal and high flowconditions. Water Research 37:2847–2864. doi:10.1016/S0043-1354(03)00133-7.

de Groot AJ, Goeij JJM, Zegers C. 1971. Contents and behaviour of mercury as comparedwith other heavy metals in sediments from the rivers Rhine and Ems. Geologie EnMijnbouw 50:393–398.

Dedkov AP, Moszherin VI. 1992. Erosion and sediment yield in mountain regions of theworld. Erosion Debris Flows and Environment in Mountain Regions :29–36.

Dong TY, Nittrouer JA, Il’icheva E, Pavlov M, McElroy B, Czapiga MJ, Ma H, Parker G.2016. Controls on gravel termination in seven distributary channels of the SelengaRiver Delta, Baikal Rift basin, Russia. Bulletin of the Geological Society of America128:1297–1312. doi:10.1130/B31427.1.

Edmonds DA, Slingerland RL. 2010. Significant effect of sediment cohesion on deltamorphology. Nature Geoscience 3:105–109. doi:10.1038/ngeo730.

Einstein HA, Anderson AG, Johnson JW. 1940. A distinction between bed-load and sus-pended load in natural streams. Transactions American Geophysical Union 21:628–633. doi:10.1029/TR021i002p00628.

Farrington J. 2000. Environmental problems of placer gold mining in the Zaamar Gold-field, Mongolia. World Placer Journal 1:107–128.

Fischer S, Pietron J, Bring A, Thorslund J, Jarsjö J. 2017. Present to future sediment trans-port of the Brahmaputra River: reducing uncertainty in predictions and management.Regional Environmental Change 17:515–526. doi:10.1007/s10113-016-1039-7.

Frampton A, Painter S, Lyon SW, Destouni G. 2011. Non-isothermal, three-phase simu-lations of near-surface flows in a model permafrost system under seasonal variabilityand climate change. Journal of Hydrology 403:352–359. doi:10.1016/j.jhydrol.2011.04.010.

Gebrehiwot KA, Haile AM, de Fraiture CMS, Chukalla AD, Embaye TaG. 2014. Opti-mizing Flood and Sediment Management of Spate Irrigation in Aba’ala Plains. WaterResources Management 29:833–847. doi:10.1007/s11269-014-0846-1.

Goulden CE, Mead J, Horwitz R, Goulden M, Nandintsetseg B, McCormick S, Boldgiv B,Petraitis PS. 2016. Interviews of Mongolian herders and high resolution precipitationdata reveal an increase in short heavy rains and thunderstorm activity in semi-aridMongolia. Climatic Change 136:281–295. doi:10.1007/s10584-016-1614-4.

Harris I, Jones PD, Osborn TJ, Lister DH. 2014. Updated high-resolution grids of

34

Sediment transport from source to sink in the Lake Baikal basin

monthly climatic observations - the CRU TS3.10 Dataset. International Journal ofClimatology 34:623–642. doi:10.1002/joc.3711.

Henry MG, Nikanarov AM. 1988. Degradation of Lake Baikal: Fate and Effects of Con-taminants Present. In: NW Schmidtke, editor. Toxic Contamination in Large Lakes.Lewis Publishers.

Hooke RL. 2000. On the efficacy of humans as geomorphic agents. Geology 28:843–846.doi:10.1130/0091-7613(2000)28<843:OTHOHA>2.0.CO;2.

Hudson PF. 2003. Event sequence and sediment exhaustion in the lower Panuco Basin,Mexico. Catena 52:57–76. doi:10.1016/S0341-8162(02)00145-5.

Hundecha Y, Bárdossy A. 2004. Modeling of the effect of land use changes on the runoffgeneration of a river basin through parameter regionalization of a watershed model.Journal of Hydrology 292:281–295. doi:10.1016/j.jhydrol.2004.01.002.

Huston D. 2014. Particulate organic carbon trapping efficiency in the stream bed forcobbles and gravels: a laboratory test. Kaleidoscope 48:73.

Ilyicheva EI, Gagarinova OV, Pavlov MV. 2015. Hydrologo-geomorphological analy-sis of landscape formation within the Selenga River delta. Geography and NaturalResources 36:263–270. doi:10.1134/S1875372815030063.

Jaramillo F, Baccard M, Narinesingh P, Gaskin S, Cooper V. 2015. Assessing the Role ofa Limestone Quarry as Sediment Source in a Developing Tropical Catchment. LandDegradation and Development 27:1064–1074. doi:10.1002/ldr.2347.

Jarsjö J, Chalov SR, Pietron J, Alekseenko A, Thorslund J. 2017. Patterns of soil con-tamination, erosion, and river loading of metals in a gold mining region of NorthernMongolia. Regional Environmental Change doi:10.1007/s10113-017-1169-6.

Karpoff B, Roscoe W. 2005. Report on Placer Gold Properties in the Tuul Valley. ZaamarGoldfield, Mongolia. Technical report. Roscoe Postle Associates INC. Toronto.

Khazheeva ZI, Tulokhonov AK, Yao R, Hu W. 2008. Seasonal and spatial distribution ofheavy metals in the selenga river delta. Journal of Geographical Sciences 18:319–327.doi:10.1007/s11442-008-0319-7.

Kondolf GM. 1997. Hungry water: Effects of dams and gravel mining on river channels.Environmental Management 21:533–551. doi:10.1007/s002679900048.

Lane CR, Anenkhonov O, Liu H, Autrey BC, Chepinoga V. 2015. Classification andinventory of freshwater wetlands and aquatic habitats in the Selenga River Delta ofLake Baikal, Russia, using high-resolution satellite imagery. Wetlands Ecology andManagement 23:195–214. doi:10.1007/s11273-014-9369-z.

Leenaers H. 1989. The transport of heavy metals during flood events in the pollutedriver Geul (the Netherlands). Hydrological Processes 3:325–338. doi:10.1002/hyp.3360030404.

Li T, Gao Y. 2015. Runoff and sediment yield variations in response to precipitationchanges: A case study of Xichuan Watershed in the Loess Plateau, China. Water7:5638–5656. doi:10.3390/w7105638.

Ligon FK, Dietrich WE, Trush WJ. 1995. Downstream Ecological Effects of Dams. Bio-Science 45:183–192. doi:10.2307/1312557.

Lindström-Seppä P, Huuskonen S, Kotelevtsev S, Mikkelson P, Räänen T, Stepanova L,Hänninen O. 1998. Toxicity and mutagenicity of waste waters from Baikalsk Pulpand Paper Mill: Evaluation of pollutant contamination in Lake Baikal. In: MarineEnvironmental Research. volume 46. p. 273–277. doi:10.1016/S0141-1136(97)00074-3.

Lychagin M, Chalov S, Kasimov N, Shinkareva G, Jarsjö J, Thorslund J. 2017. Surfacewater pathways and fluxes of metals under changing environmental conditions andhuman interventions in the Selenga River system. Environmental Earth Sciences 76:1.doi:10.1007/s12665-016-6304-z.

Lynds RM, Mohrig D, Hajek EA, Heller PL. 2014. Paleoslope reconstruction in sandysuspended-load-dominant rivers. Journal of Sedimentary Research 84:825–836. doi:10.2110/jsr.2014.60.

Magilligan FJ, Phillips JD, James LA, Gomez B. 1998. Geomorphic and SedimentologicalControls on the Effectiveness of an Extreme Flood. The Journal of Geology 106:87–96.doi:10.1086/516009.

Malsy M, Flörke M, Borchardt D. 2016. What drives the water quality changes in the

35

Jan Pietron

Selenga Basin: climate change or socio-economic development? Regional Environ-mental Change :1–13doi:10.1007/s10113-016-1005-4.

Malsy M, Heinen M, aus der Beek T, Flörke M. 2013. Water resources and socio-economic development in a water scarce region on the example of Mongolia. Geo-Öko 34:27–49.

Marin A. 2010. Riders under storms: Contributions of nomadic herders’ observationsto analysing climate change in Mongolia. Global Environmental Change 20:162–176.doi:10.1016/j.gloenvcha.2009.10.004.

MEGD. 2012. Tuul River Basin Integrated Water Resource Management Plan. Techni-cal report. Mongolian Ministry for Environment and Green Development (MEGD).Ulaanbaatar.

Middleton GV, Southard JB. 1984. Mechanics of Sediment Movement.2nd edition. volume 3. SEPM (Society for Sedimentary Geology). doi:10.2110/scn.84.03.0001\r10.2110/scn.84.03.0002\r10.2110/scn.84.03.0003\r10.2110/scn.84.03.0004\r10.2110/scn.84.03.0005\r10.2110/scn.84.03.0006\r10.2110/scn.84.03.0007\r10.2110/scn.84.03.0008. arXiv:1011.1669v3.

Miller JR, Mackin G, Orbock Miller SM. 2015. Geochemical Fingerprinting. In:Application of geochemical tracers to fluvial sediment. Springer. p. 11–43. doi:10.1007/978-3-319-13221-1.

Mohamadi MA, Kavian A. 2015. Effects of rainfall patterns on runoff and soil erosionin field plots. International Soil and Water Conservation Research 3:273–281. doi:10.1016/j.iswcr.2015.10.001.

Mun Y, Ko I, Janchivorj L, Gomboev B, Kang S, Lee C. 2008. Integrated Water Manage-ment Model on the Selenge River Basin – Statys Syrvey and Investigation (Phase I).Technical report. Korea Environment Institute. Seoul.

Nilsson C, Reidy CA, Dynesius M, Revenga C. 2005. Fragmentation and Flow Regula-tion of the World ’s Large River Systems. Science 308:405–408. doi:10.1126/science.1107887.

Orton GJ, Reading HG. 1993. Variability of deltaic processes in terms of sedimentsupply, with particular emphasis on grain size. Sedimentology 40:475–512. doi:10.1111/j.1365-3091.1993.tb01347.x.

Paull D, Banks G, Ballard C, Gillieson D. 2006. Monitoring the Environmental Impact ofMining in Remote Locations through Remotely Sensed Data. Geocarto International21:33–42. doi:10.1080/10106040608542372.

Pederson N, Hessl AE, Baatarbileg N, Anchukaitis KJ, Di Cosmo N. 2014. Pluvials,droughts, the Mongol Empire, and modern Mongolia. Proceedings of the NationalAcademy of Sciences 111:4375–4379. doi:10.1073/pnas.1318677111.

Peizhen Z, Molnar P, Downs WR. 2001. Increased sedimentation rates and grain sizes2–4 Myr ago due to the influence of climate change on erosion rates. Nature 410:891–897. doi:10.1038/35073504.

Pietron J, Chalov SR, Chalova AS, Alekseenko AV, Jarsjö J. 2017a. Extreme spatial vari-ability in riverine sediment load inputs due to soil loss in surface mining areas of thelake baikal basin. CATENA 152:82 – 93. doi:doi.org/10.1016/j.catena.2017.01.008.

Pietron J, Jarsjö J, Romanchenko AO, Chalov SR. 2015. Model analyses of the contribu-tion of in-channel processes to sediment concentration hysteresis loops. Journal ofHydrology 527:576 – 589. doi:doi.org/10.1016/j.jhydrol.2015.05.009.

Pietron J, Nittrouer JA, Chalov SR, Dong TY, Shinkareva G, Jarsjö J. 2017b. Sedimenta-tion processes in the selenga river delta: implication for sequestering sediment-boundmetals. Hydrological Processes. In review.

Planchon O, Esteves M, Silvera N, Lapetite JM. 2000. Raindrop erosion of tillage inducedmicrorelief: Possible use of the diffusion equation. Soil and Tillage Research 56:131–144. doi:10.1016/S0167-1987(00)00134-3.

Plyusnin AM, Zhambalova DI. 2014. Effect of land reclamation on the ecological stateof surface and subsurface water in Ust’-Selenga depression. Water Resources 41:839–843. doi:10.1134/S0097807814070100.

Poesen J, Nachtergaele J, Verstraeten G, Valentin C. 2003. Gully erosion and environ-mental change: Importance and research needs. Catena 50:91–133. doi:10.1016/

36

Sediment transport from source to sink in the Lake Baikal basin

S0341-8162(02)00143-1.Potemkina T, Potemkin V. 2015. Sediment load of the main rivers of lake baikal in a

changing environment (east siberia, russia). Quaternary International 380:342–349.doi:10.1016/j.quaint.2014.08.029.

Priess JA, Schweitzer C, Wimmer F, Batkhishig O, Mimler M. 2011. The consequences ofland-use change and water demands in Central Mongolia. Land Use Policy 28:4–10.doi:10.1016/j.landusepol.2010.03.002.

Pruski FF, Nearing MA. 2002. Climate-induced changes in erosion during the 21stcentury for eight U.S. locations. Water Resources Research 38:34–1–34–11. doi:10.1029/2001WR000493.

Rinaldi M, Wyzga B, Surian N. 2005. Sediment mining in alluvial channels: Physicaleffects and management perspectives. doi:10.1002/rra.884.

Robinson WP, Anosova GB, Center B. 2004. Mining and Mineral Development Manage-ment Policy in the Selenga River Watershed. In: Science for Watershed Conservation:Multidisciplinary Approaches for Natural Resource Management Conference.

Simon A, Rinaldi M. 2006. Disturbance, stream incision, and channel evolution: Theroles of excess transport capacity and boundary materials in controlling channel re-sponse. Geomorphology 79:361–383. doi:10.1016/j.geomorph.2006.06.037.

Singh AK, Hasnain SI, Banerjee DK. 1999. Grain size and geochemical partitioning ofheavy metals in sediments of the Damodar River - a tributary of the lower Ganga,India. Environmental Geology 39:90–98. doi:10.1007/s002540050439.

Stocker T, Qin D, Plattner GK, Tignor M, Allen S, Boschung J, Nauels A, Xia Y, Bex V,Midgley P. 2013. IPCC, 2013: Climate Change 2013: The Physical Science Basis. Con-tribution of Working Group I to the Fifth Assessment Report of the IntergovernmentalPanel on Climate Change. Technical report.

Stubblefield A, Chandra S, Eagan S, Tuvshinjargal D, Davaadorzh G, Gilroy D, SampsonJ, Thorne J, Allen B, Hogan Z. 2005. Impacts of gold mining and land use alterationson the water quality of central Mongolian rivers. Integrated environmental assess-ment and management 1:365–73.

Syvitski JPM, Kettner AJ, Overeem I, Hutton EWH, Hannon MT, Brakenridge GR, DayJ, Vörösmarty C, Saito Y, Giosan L, Nicholls RJ. 2009. Sinking deltas due to humanactivities. Nature Geoscience 2:681–686. doi:10.1038/ngeo629.

Tessier A, Campbell P, Bisson M. 1982. Particulate trace metal speciation in stream sed-iments and relationships with grain size: Implications for geochemical exploration.Journal of Geochemical Exploration 16:77–104. doi:10.1016/0375-6742(82)90022-X.

Thorslund J, Jarsjö J, Chalov SR, Belozerova EV. 2012. Gold mining impact on riverineheavy metal transport in a sparsely monitored region: the upper Lake Baikal Basincase. Journal of Environmental Monitoring 14:2780. doi:10.1039/c2em30643c.

Thorslund J, Jarsjö J, Wällstedt T, Mörth CM, Lychagin MY, Chalov SR. 2016. Speciationand hydrological transport of metals in non-acidic river systems of the Lake Baikalbasin: Field data and model predictions. Regional Environmental Change doi:10.1007/s10113-016-0982-7.

Tiecher T, Caner L, Minella JPG, Bender MA, dos Santos DR. 2016. Tracing sedimentsources in a subtropical rural catchment of southern Brazil by using geochemicaltracers and near-infrared spectroscopy. Soil and Tillage Research 155:478–491. doi:10.1016/j.still.2015.03.001.

Tulokhonov A, Plyusini V, Kudelya S, Beshentsev A. 2015. The Ecological Atlas of theBaikal Basin. Irkutsk: UNDP-GEF.

Törnqvist R, Jarsjö J. 2012. Techniques : Basin-Scale Quantification in Semi-Arid Envi-ronments. Water Resources Management 26:949–962. doi:10.1007/s11269-011-9819-9.

Törnqvist R, Jarsjö J, Pietron J, Bring A, Rogberg P, Asokan SM, Destouni G. 2014. Evo-lution of the hydro-climate system in the lake baikal basin. Journal of Hydrology519:1953 – 1962. doi:doi.org/10.1016/j.jhydrol.2014.09.074.

UNESCO. 1996. Convention concerning the protection of the world cultural and naturalheritage: Report, 2-7/12/1996. Technical report. Merida.

Van Rompaey AJJ, Verstraeten G, Van Oost K, Govers G, Poesen J. 2001. Modellingmean annual sediment yield using a distributed approach. Earth Surface Processes

37

Jan Pietron

and Landforms 26:1221–1236. doi:10.1002/esp.275.Verstraeten G, Oost K, Rompaey A, Poesen J, Govers G. 2006. Evaluating an inte-

grated approach to catchment management to reduce soil loss and sediment pollutionthrough modelling. Soil Use and Management 18:386–394. doi:10.1111/j.1475-2743.2002.tb00257.x.

Viparelli E, Wesley Lauer J, Belmont P, Parker G. 2013. A numerical model to developlong-term sediment budgets using isotopic sediment fingerprints. Computers andGeosciences 53:114–122. doi:10.1016/j.cageo.2011.10.003.

Walling DE. 1988. Erosion and sediment yield research - Some recent perspectives. Jour-nal of Hydrology 100:113–141. doi:10.1016/0022-1694(88)90183-7.

Walling DE. 2006. Human impact on land-ocean sediment transfer by the world’s rivers.Geomorphology 79:192–216. doi:10.1016/j.geomorph.2006.06.019.

Walling DE, Fang D. 2003. Recent trends in the suspended sediment loads of the world’srivers. Global and Planetary Change 39:111–126. doi:10.1016/S0921-8181(03)00020-1.

Whitney PR. 1975. Relationship of manganese-iron oxides and associated heavy metalsto grain size in stream sediments. Journal of Geochemical Exploration 4:251–263.doi:10.1016/0375-6742(75)90005-9.

Zhao L, Wu Q, Marchenko SS, Sharkhuu N. 2010. Thermal state of permafrost and activelayer in central Asia during the international polar year. Permafrost and PeriglacialProcesses 21:198–207. doi:10.1002/ppp.688.

38