Catena 76 (2008) 27–35

Contents lists available at ScienceDirect

Catena

j ourna l homepage: www.e lsev ie r.com/ locate /catena

Spatial and temporal variability of sediment and dissolved loads from two alpinewatersheds of the Lesser Himalayas

Omvir Singh a,⁎, Milap C. Sharma b, A. Sarangi c, Pratap Singh d

a Division of Environmental Sciences, IARI, Pusa Campus, New Delhi 110012, Indiab Centre for the Study of Regional Development, Jawaharlal Nehru University, New Delhi 110067, Indiac Water Technology Centre, IARI, Pusa Campus, New Delhi 110012, Indiad Hydro Tasmania Consulting, 12th Floor, Eros Corporate Tower, Nehru place, New Delhi 110019, India

⁎ Corresponding author. Tel.: +91 11 25841490; fax: +E-mail addresses: ovshome@yahoo.com, ovshome@g

asarangi@iari.res.in (A. Sarangi).

0341-8162/$ – see front matter © 2008 Elsevier B.V. Aldoi:10.1016/j.catena.2008.08.003

a b s t r a c t

a r t i c l e i n f o

Article history:

Estimation of sediment lo Received 17 January 2008Received in revised form 7 August 2008Accepted 21 August 2008

Keywords:Sediment transportDenudation rateWeatheringHimalayan Watersheds

ad from Himalayan basins is of considerable importance for the planning,designing, installation and operation of hydro-power projects, including management of reservoirs. In thepresent study, an assessment of physical and chemical load, sediment yield and erosion rate has beenundertaken at eight different locations in the Sainj and Tirthan watersheds. The analysis revealed that themaximum load was transferred during the monsoon season. Moreover, the estimated average chemicalerosion rate of the Sainj (83 t km−2 yr−1) and Tirthan (80 t km−2 yr−1) watersheds were higher than that ofthe Indian average (69 t km−2 yr−1) representing all the rivers. Both watersheds were eroding physically andchemically at a faster rate than that of the world global average erosion rate (185 t km−2 yr−1). The flattishnature of the channels in some segments of these watersheds showed a lower transport of sediments, whereas the constricted segments having steep bed slopes increased the velocity of flow and the sedimenttransport rate. These findings have important implications for water resource management in the context ofsediments mobilization, erosion, channel management, ecological functions and operation of the hydro-power projects in the Lesser Himalayan region.

© 2008 Elsevier B.V. All rights reserved.

1. Introduction

Pollution from non-point sources is being recognized as a majorsource of surface water quality deterioration. Sediment transportationfrom land surfaces to oceans through rivers is one of the mostimportant processes that affect riverbank stabilization, soil formation,upliftment rates, biogeochemical cycling of elements, crust evolutionand many other earth related processes (Chakrapani, 2005). Thetransport dynamics of sediments in fluvial systems have been studiedby engineers, hydrologists, soil scientists, geologists and processgeomorphologists in different parts of the world (Holeman, 1968;Subramanian, 1979, 1993; Milliman and Meade, 1983; Jha et al., 1988;Chakrapani and Subramanian, 1990, 1993; Lu and Siew, 2006; Soleret al., 2007; Vanacker et al., 2007). Presently, the rivers alonecontribute to about 95% of sediments entering the oceans in a globalscale (Syvitski, 2003). It is estimated that 65% of water and 80% of thesediments are being transported to oceans each year from SouthernAsian, Oceania and north-eastern South American rivers (Syvitskiet al., 2003). It is also reported that the Himalayan Rivers are themajor

91 11 25848037.mail.com (O. Singh),

l rights reserved.

contributors, which transport about 50% of the global sediment flux(Stoddart, 1969).

Therefore, research interests pertaining to fluvial sediment trans-port in the Himalayan river systems have attracted much attention inrecent times (Kumar, 1987; Rawat and Rai 1997; Singh et al., 2003,2005; Sharma and Rai, 2004; Haritashya et al., 2006; Singh, 2007;Bhattacharya et al., 2008). The estimation of the sediment load and thetransport rate governs the geomorphological, hydrological, sedimen-tological and ecological processes of river basins. Further, itsestimation from basins is of much significance for planning, designing,installation and operation of hydro-power projects, including man-agement of reservoirs. However, there exist hardly any studies relatedto the transportation and quantification of sediment fluxes from theLesser Himalayan watersheds, which are currently being developedfor various hydro-power projects. Hence, in view of this research gaprelated to the study of sediment fluxes in the Lesser Himalayanwatersheds, the present work was undertaken with an objective tocollect and analyze the primary hydrologic andwater quality data. Theacquired data pertaining to quantity and quality of the sediment(physical) and dissolved (chemical) loads transported by two alpinewatersheds were analyzed to understand the watershed hydrologyand sedimentation processes. Such type of investigations assumesimportance in preparation of the water and sediment load budget at

Fig. 1. Delineated map of the Sainj and Tirthan watersheds with natural drainage network and sampling locations.

28 O. Singh et al. / Catena 76 (2008) 27–35

regional scale, which will assist in the development of a global sedi-ment flow budget.

2. Materials and methods

2.1. Description of the study area

The study area is delineated into two distinct watersheds named asSainj (741 km2) and Tirthan (687 km2). Both watersheds fall under theleft bank of Upper Beas river system in the Lesser Himalayan alpinezone. The area extends between latitudes 31°30′28″ and 31°55′02″north and longitudes 77°13′02″ and 77°45′57″ east as shown in Fig. 1.The watersheds present a typical mosaic of moderate to high ruggedtopography with numerous mountain peaks over 4000 m abovemean

Table 1Relationship between different land use classes and erosion rates in the Sainj andTirthan watersheds

Station Erosionrates(t km−2

yr−1)

Different land use

Agriculturalland

Degradedforest

Forest Glaciers Hilly openscrub/rockyoutcrop

Snowbound

(%)

Sainj atNiharni 608.00 0.38 1.25 27.30 12.48 29.04 29.55Suind 164.47 2.02 1.59 36.20 10.35 25.28 24.56Sainj 446.32 2.07 1.55 36.46 10.61 23.97 25.35Talara 733.80 4.25 1.57 38.11 9.61 23.52 22.95Larji 266.86 4.49 1.54 38.21 9.44 23.77 22.55Jiwa Nal 258.26 2.23 1.43 37.25 11.40 19.97 27.74

Tirthan atKhrongcha 574.54 2.38 0.60 42.83 3.50 28.93 21.76Larji 196.56 16.88 1.19 56.16 1.14 17.57 7.07

sea level (amsl). The average slope of the Sainj and Tirthanwatershedsare 38.12° and 40.04° along with the mean elevation of 3510 and2826 m amsl, respectively. The rock types were mainly of colluvium,alluvium, glacial deposits, phyllite, slate, quartzites, dolomites, sand-stone, schist and granites. The soil texture varies from sandy loamto loam with average organic matter content of around 70%. Soilsvary from very shallow to moderately deep in depth and pale yellow,yellowish brown and dark brown in colour.

The climate of the watersheds is mostly warm temperate andreceive an average annual rainfall of 1000 mm and more than 50% ofwhich is received during the south-west monsoon (June–September).Average annual snowfall in the region is about 345 mm, which isconfined to upper reaches and occurs only during the winter season

Table 2Distribution and subsidiary information of water sampling points in the Sainj andTirthan watersheds

Majorwatershed

Name oflocation

IDno.

Latitude Longitude Samplingsiteelevation(m)

Area(km2)

Dominantvegetationtype

Sainj Niharni 1 31° 46′ 29″ 77° 22′ 90″ 1586 410.65 Deodarand pinemixed

Suind 2 31° 47′ 08″ 77° 19′ 92″ 1419 495.01 PineSainj 3 31° 46′ 19″ 77° 18′ 56″ 1251 658.41 PineTalara 4 31° 45′ 79″ 77° 15′ 95″ 1149 727.26 Scattered

pineLarji 5 31° 43′ 64″ 77° 13′ 60″ 997 740.02 PineBhatkanda(Jiwa Nal)

6 31° 47′ 24″ 77° 19′ 81″ 1388 163.40 Scatteredpine

Tirthan Khrongcha 7 31° 39′ 55″ 77° 27′ 33″ 1786 191.13 Mixedvegetation

Larji 8 31° 43′ 48″ 77° 13′ 36″ 989 686.36 Scatteredpine

Table 3Maximum and minimum daily discharge rates at different gauging sites during 2004

Watershed Station Maximumdischargerate (m3 s−1)

Date Minimumdischargerate (m3 s−1)

Date

Sainj Niharni 71.9 June 29 6.3 February 4Suind 63.8 August 8 8.4 January 14Sainj 41.1 August 8 3.6 January 13Talara 101.3 August 8 12.3 January 13Bhatkanda(Jiwa Nal)

106.4 August 8 14.5 December 29

Larji 117.8 August 8 15.3 December 29Tirthan Khrongcha 38.5 September 11 2.1 January 28

Larji 141.6 August 8 7.4 January 16

29O. Singh et al. / Catena 76 (2008) 27–35

(October–March). Themeanmonthly temperatures at Larji (the out letof watersheds) ranged from aminimum of 8.7 °C during January to themaximum of 26.3 °C during June. The minimum and maximumrelative humidity is recorded in themonths of May (63.3%) and August(78.7%) respectively. Evaporation was observed to be minimum in themonths of December (36.1 mm) and January (38.7 mm), the coldestmonths of the year and maximum during June (165.0 mm), thewarmest month of the year.

2.2. Watersheds delineation and preparation of land use map

The study watersheds (Sainj and Tirthan) were delineated from thetopological information of the watersheds using the WatershedMorphology Estimation Tool (WMET) interface (Sarangi et al., 2004).The land use maps of these watersheds were generated using thetopographical maps and False Colour Composite (FCC) images ofDecember 5,1999 captured by IndianRemote Sensing (IRS-1C) satellite.While generating the land usemaps, the ground truth information andvisual interpretation techniqueswere also considered. Further, the FCCgenerated land usemap of thewatersheds were scanned, digitized andprojected to realworld co-ordinate systems using the Auto Cad andArcInfo GIS softwares and their areal extent under different land useconditions were estimated at different sampling points.

It was observed from the land use maps of the Sainj and Tirthanwatersheds that the forested area was maximum with 38.2(282.8 km2) and 56.2% (385.5 km2), respectively. The second majorland use type was identified as rocky outcrops, which covered an areaof 23.8 (175.9 km2) in the Sainj and 17.6% (120.6 km2) in the Tirthanwatershed. Snow bound areawas 22.6 (166.9 km2) and 7.1% (48.5 km2)in the Sainj and Tirthan watershed, respectively. The agricultural landwas predominant in the Tirthan watershed with 16.9% (115.8 km2) ascompared to 4.5% (33.2 km2) in the Sainj watershed. The area coveredby glaciers were 9.4 (69.9 km2) and 1.1% (7.8 km2) for the Sainj and

Fig. 2. Gravel bedded nature of channels in the study watersheds.

Tirthan watersheds, respectively. Moreover, the land use informationat different sampling sites corresponding to the delineated watershedarea between two consecutive gauging sites is presented in Table 1.Besides, there was a growth of 27.2% in human and 47.4% in livestockpopulation envisaged during a period from 1981 to 2001 in thesewatersheds. Also, 25–30% encroachment of agricultural activitiesin the forest land coupled with enhanced deforestation led to theremoval of 119,438.4 m3 of timber during a period from 1987 to 2005.The construction of 55 km of roads during 1994 to 2002 in thewatersheds resulted in excavationworks amounting to 2.2 to 4.4 Mm3

of debris accentuating the sediment loads in the channels (Singh,2007).

2.3. Collection and analysis of water samples

The water samples from the drainage channels of the watershedswere collected form both the main channel and its tributaries joiningthe main stream at the gauging site thrice a day during morning, noonand evening. This sequence was followed once in each season duringsummer (April–May), monsoon (June–September) and winter (Octo-ber–March) from eight different gauging sites (Fig. 1). The sampleswere collected seasonally (May1–8, for summer season; August 15–22,during monsoon season and December 20–27, 2004 inwinter season).This resulted in acquisition of 24 samples for the year 2004. The depthintegrating sampling technique was used for collection of 1000 mlsample in plastic bottle from reasonable turbulent river flow (Ostrem,1964; Bhutiyani, 2000). Further, a global positioning system (GPS) wasused to record the latitude, longitude and the altitudinal location ofthese water sampling points (Table 2).

The collected samples were analyzed in the laboratory following astandard procedure to estimate the suspended sediment concentra-tion (SSC). Suspended sediments were separated from the watersamples in the laboratory by using 0.45 µmmilliporemembrane filtersof 47 mm diameter. To facilitate faster filtration, a vacuum pump wasattached with the filtration assembly. The weight of the sediment wasmeasured by subtracting the weight of the filter paper from theweight of the dried filter paper along with the filtered sediment afterfiltration. Further, before taking the weight of the sediments on thefilter paper, it was kept in a desiccator to remove moisture from thesediment. The SSC was calculated for 1 l of the collected water sample.Further, the pH and EC of the collected samples were estimated usingthe Jackson (1967) and Richards (1954) methods. These parametersweremeasured in the laboratory by using the collected water samplesbefore the filtration process.

Further, information pertaining to human interventions (humanand livestock growth, road construction, deforestation and encroach-ments) were collected through personal interaction with the local

Table 4Suspended sediment concentration, TDS and alkalinity at various stations in the Sainjand Tirthan watershed (2004)

Location Summer (Maya) Monsoon (Augusta) Winter (Decembera)

TDS SSC pH TDS SSC pH TDS SSC pH

(mg l−1) (mg l−1) (mg l−1) (mg l−1) (mg l−1) (mg l−1)

Sainj atNiharni 30.1 346.85 8.52 72.1 466.83 7.87 58.1 18.55 8.73Suind 43.4 268.70 8.32 68.6 89.46 7.51 59.5 6.79 8.53Sainj 27.3 540.13 7.44 68.6 350.86 7.88 44.8 5.21 8.31Talara 34.3 435.74 8.03 71.4 749.33 8.00 55.3 3.40 8.27Larji 34.3 108.70 8.52 72.1 201.83 8.10 57.4 7.62 8.11Bhatkanda(Jiwa Nal)

34.3 149.00 8.20 63.7 85.44 8.04 58.8 13.81 8.05

Tirthan atKhrongcha 27.3 75.42 7.50 60.2 369.74 7.82 51.8 5.16 8.54Larji 42.0 59.80 8.26 78.4 156.00 8.05 70.7 4.74 8.26

a Month of the season during which the samples were collected and analyzed.

Table 5Comparison of the alkalinity levels in the Indian rivers and the rivers of study watershed(Subramanian, 1979; Kumar, 1987)

Rivers pH value

Major Indian riversCauvery 7.8Gandak 7.3Godavari 7.1Krishna 7.5Mahanadi 8.0Narmada 7.2Tungabadra 7.5

Himalayan riversBrahmaputra 7.1Chenab 7.7Dhauli Ganga 8.0Ganga 7.6Jhelum 7.5

Study watershedsSainj 8.2Tirthan 7.6

30 O. Singh et al. / Catena 76 (2008) 27–35

population of the study watersheds and from various governmentalorganization in the state of Himachal Pradesh, India.

2.4. Computation of load

2.4.1. Suspended sediment loadThe suspended sediment concentration (SSC) for all days of the

season were derived by multiplying the daily discharge rate at thegauging station with the ratio of suspended sediment concentration(SSC) and discharge rate recorded on the day of sample collection. Thisprocedure of proportional changes in the SSC and the discharge ratewas considered to account for the changes in SSC with the varyingdischarge rates at the gauging siteswith an assumption that the fluvialenergy, sediment availability for transport, rainfall intensity–dura-tion–frequency, droplet size etc. does not change significantly duringthe data acquisition periods. Moreover, the daily discharge rate valueswere obtained from the hydrologic monitoring stations maintained byBhakra Beas Management Board (BBMB), in which the flow depthsacquired from stage level recorders were converted to the dischargerate using the developed stage discharge rating curves for therespective gauging sites (BBMB, 1997 ). The daily discharge ratesacquired from the gauging sites were analyzed to obtain themaximumand minimum daily flows and are shown in Table 3. The abovediscussed proportionality method was used to estimate the dailysediment flow rate for all the seasons using the data of eight gaugingsites. Further the seasonal and annual river flow volumes and the totalsuspended sediment load was estimated by using the standardaccounting procedures spanning through the days, months, seasonsand years.

Table 6Sediment and dissolved load transfer in the Sainj and Tirthan watersheds

Station/river Sediment/physical (P) load (103 t yr−1) D

Summer Monsoon Winter Total sediment load S

Sainj atNiharni 37.59 173.50 2.92 214.02 2Suind 23.30 25.34 1.30 49.94 3Sainj 74.68 168.97 1.57 245.23 3Talara 66.44 408.57 1.13 476.14 4Larji 17.50 116.15 2.62 136.29 5Bhatkanda (Jiwa Nal) 7.38 16.43 1.44 25.26 1

Tirthan atKhrongcha 4.39 87.38 0.36 92.13 1Larji 6.93 71.68 1.31 79.93 4

2.4.2. Bed loadThe bed load, which is primarily composed of coarser particles,

plays a major role in fluvial process, even though its quantity in mostcases remains less than that of the suspended load (Xiaoping, 2003 ).In the present study, the bed load is being estimated because it wasnot possible for authors to measure the bed load in the river section.Some studies have suggested the crude empirical relationshipbetween suspended load and bed load (Colby and Hembree, 1955;Fergusson, 1984; Bhutiyani, 2000; Sitaula et al., 2007). It has beenreported that the ratio of bed load to the suspended sediment loadbased on annual loads amounts to about 0.05–0.15 (5–15%) with apossible estimation error of ±5%. However, for rivers with relativelystable boundary and inflow conditions, the range of variation may notbe so large (Zhang and Long, 1998). Further, this ratio also varies withthe compositions of channel bed and concentrations of suspendedsediment load (Vanoni, 1975). In general, the channels of the studywatersheds were observed to be gravel bedded due to mass wastingprocesses (Fig. 2). Therefore, the bed load was heuristically consideredas 10% of the suspended sediment load in the present study.

2.4.3. Dissolved loadThe estimation of total dissolved solids (TDS) was carried out by

following the methodology as proposed by Gregory and Walling(1973). In this method, the TDS was estimated by using the equation:

TDS ¼ AK ð1Þ

Where, K=electrical Conductivity (EC) in dS m−1; A=conversionfactor.

In general, the conversion factor varied from 0.55 to 0.75 accordingto the types of solutes (Gregory and Walling, 1973). The conversionfactor of 0.64 pertaining to Indian river systems was used for esti-mation of TDS in this study (Singh et al., 1999). The total dissolved load(TDL) was calculated in the same manner as that of total suspendedsediment load.

Subsequently, the total sediment load (TSL) was estimated byadding the total suspended sediment load and the bed load. Finally,the total load (TL) was calculated by adding the estimated TSL and TDL.

2.5. Estimation of erosion rates

The rates of denudation were estimated by using the relationshipgiven by (Gregory and Walling, 1973):

Rate of denudation ¼ Total load ðtonnesÞArea of watershed� specific gravity

ð2Þ

Denudation rates are expressed in mm/1000 yr or in m3/km2/yr. Inthis study, the specific gravity of the rock material was considered tobe 2.65 (Valdiya, 1987).

issolved/chemical (C) load (103t yr−1) C/P ratio Total load(103t yr−1)ummer Monsoon Winter Total dissolved load

.96 24.36 8.31 35.64 0.17 249.67

.42 17.66 10.38 31.46 0.63 81.41

.43 30.03 15.16 48.62 0.20 293.86

.75 35.39 17.36 57.51 0.12 533.66

.02 37.72 18.44 61.18 0.45 197.47.54 11.13 4.25 16.93 0.67 42.19

.44 12.93 3.29 17.67 0.19 109.81.42 32.75 17.79 54.97 0.69 134.90

Fig. 4. Alluviation and widening of channel at Suind in Sainj watershed resulting in thereduced sediment transport.

31O. Singh et al. / Catena 76 (2008) 27–35

3. Results and discussion

3.1. Suspended sediment concentration, TDS and alkalinity at differentstations

Suspended sediment concentration (SSC), total dissolved solids(TDS) and pH values at various locations during summer, monsoonand winter seasons are summarized in Table 4. The spatial and tem-poral variations in SSC and TDS had a significant effect on the annualsediment transport behaviour from both the watershed systems.Majority of the sediment load was in the form of suspended sedimentand was transported during the monsoon months at all locations. Theamount of TDS increased towards the downstream during all theseasons in Tirthan watershed due to joining of various spring fedstreams to the main channel. Conversely, changes in TDS values werenot adequately pronounced over the Sainj watershed in differentseasons. Moreover, the winter season observed higher TDS amountsthan the summer due to prolonged dry season spanning from Octoberto December in both the watersheds. Because, during this period,there was occurrence of base flow in which the solute contentincreased due to low discharge contributed mainly by the waterstored in the aquifer. Further, TDS content was comparatively lowerduring summer season than the other seasons and it was presumablydue to the contribution of increased snowmelt runoff in the water-sheds, which diluted the total flow. The TDS content was on the higherside during the monsoon, which could be attributed to high seasonalflows. Overall, the dry season could be important for chemicalprocesses and the wet season may be critical for sediment masstransfer due to high intensity of rainfall and mass wasting phenom-enon in thewatersheds. It was observed that, like other major rivers ofthe world, the Sainj and Tirthan rivers were also alkaline in nature.The comparison of the alkalinity levels of the river system in the studywatersheds and some other Indian rivers are also presented in Table 5.It was observed from the Table 5 that the alkalinity level of the Sainjwatershed was slightly higher than the majority of Indian rivers.Where as, the alkalinity of the Tirthan watershed was almost in linewith other Indian rivers.

3.2. Sediment and dissolved load transport behaviour of watersheds

Based on the values of SSC, bed load, TDS and the discharge atvarious stations, the annual sediment transport as a function of spaceand time was estimated and are summarized in Table 6. The totalsediment and total dissolved load from the Lesser Himalayan alpinewatersheds indicated the pronounced effect of the seasonality on

Fig. 3. Development of Hydro-power projects generates huge quantity of muck andincreases the sediment load in Sainj watersheds.

sediment transport. The total load (TL) from the Sainj and Tirthanwatersheds was 197.47 and 134.90 t3 during 2004. Moreover, about78% of the TL was transported during the monsoon season from bothwatersheds. The TL of these watersheds showed alternatively risingand falling trends during summer and monsoon seasons respectively.Moreover, the TL not only revealed the temporal variation in differentseasons but also showed the spatial variation during the same season.The dissolved load contributed about 31 and 40% of the TL in the Sainjand Tirthan watershed, respectively. The elevated transport of totalsediment load (TSL) during the monsoon season was due to higherdischarge rates coupled with non-point sources of SSC in surfacerunoff. The TSL from the Sainj and Tirthan watersheds constitutedabout 60–90%. Also, the total dissolved load (TDL) constituted about10–40% of the TL depending upon the site characteristics. Highaltitude of both the watersheds resulted in generation of more TSLthan the TDL.

The highest TL for all seasons were observed to be at Talarabridge (533.66 t3) falling in the Sainj watershed. The reason for thisexceptionally high load could be attributed to the widening of roads,hydro-power construction activities, mining and agricultural activ-ities as shown in Fig. 3. Further, very high load in Sainj watershedwas observed throughout the year and could be attributed tothe joining of glacial fed tributaries, steep sidewalls and frequentmass wasting phenomenon. However, the Sainj river carried lowestamount of TSL at Suind (81.41 t3), located between Niharni and Sainj,because of more flattish nature of river course, low gradient, maxi-mum channel alluviation and higher channel widening activities(Fig. 4).

Fig. 5. Encroachment of agricultural activity on marginal sloping lands of the studywatersheds.

Fig. 6. Monthly variations of discharge rate, dissolved, suspended and total load at different locations in study watersheds.

32 O. Singh et al. / Catena 76 (2008) 27–35

33O. Singh et al. / Catena 76 (2008) 27–35

Similarly, in the Tirthanwatershed, Khrongcha site yielded very highTSL (92.13 t3). This could be attributed to the presence of glacial fedstreams in upper course, steep sidewalls and more frequent landslidesalong the channel in upstream section. Moreover, the enhanced TSLwas due to the presence of cultivated land and encroachments ofagricultural activities in the forestlands and pastures (Fig. 5). Theamount of TSL decreased at Larji site due to the flattish nature, channelwidening and sedimentation at the downstream site of Khrongcha.Additionally, the tributaries joining downstream site of Khrongchahave smaller contributions of SSC due to various spring fed streams.

Analysis of data form the gauging stations in the Sainj and Tirthanwatersheds also indicated monthly variations in discharge rate, TSL,TDL and TL (Fig. 6). TSL and TDL were maximum during the month ofAugust for all stations, which may be attributed to the weatheringof materials during the dry season and maximum discharge duringthe month. Further, TDL was observed to be uniform from October toMarch, which may be attributable to the ground water contribution inthe watershed systems.

3.3. Comparison of physical and chemical erosion

The ratio of chemical (C) to physical (P) loads (C/P) for all thegauging sites in the Sainj and Tirthan watersheds was observed tobe less than one, indicating overall dominance of TSL over the TDL(Table 6). However, higher C/P ratio of individual location on themainstream and tributary suggested that they are chemically moreactive. The differences in C/P ratio at different locations were mainlydue to the differences in lithological pathways through which thewater flows. Generally, the C/P ratio decreases with the increase ofwatershed area. However, in the present study, the C/P ratio of theSainj watershed does not reflect a specific trend but an increasingtrend from Khrongcha to Larji in the Tirthanwatershed was observed.Such variations may be attributed to the changes in human activity,

Fig. 7. Relationship between chemical and physica

vegetation cover and spatial and temporal variability in the intensityof precipitation throughout and within the sub-watersheds.

Further, the scattered diagram indicated an increasing trendbetween physical and chemical erosion rates for the selected locationsin the Sainj and Tirthan watersheds. However, similar relationshipbetween these two processes has been observed in other Indian,Himalayan and world rivers (Fig. 7).

3.4. Erosion rates and lowering of the watersheds

Rates of erosion have been estimated based on the total sedimentand total dissolved load discharged by the watersheds at variouslocations. The physical, chemical and total denudation rates werecalculated at various points in Sainj and Tirthan watersheds and theirtributary sub-basins (Table 7). The total rates of erosion in Sainjwatershed varied from a minimum value of 164.47 t km−2 yr−1 for theSuind site to a high value of 734 t km−2 yr−1 at Talara site. The erosionrates for entire Sainj and Tirthan watersheds during 2004 samplingperiod were calculated to be 267 t km−2 yr−1 and 197 t km−2 yr−1

respectively. These erosion rates of the Sainj and Tirthan watershedswere very high as compared to other Indian, Himalayan and somemajor world rivers (Table 8).

Moreover, based on the TL at different sites, the lowering rates ofthe watersheds were estimated and are given in Table 7. Physicalweathering seems to be an effective mechanism of erosion leading tothe lowering of the watersheds. The lowering rates for the entire Sainjand Tirthan watersheds based on the observations during 2004 wereestimated to be 0.10 mm yr−1 (0.07 mm yr−1 due to physical erosionand 0.3 mmyr−1 due to chemical erosion) and 0.07 mmyr−1 (0.04 mmyr−1 due to physical erosion and 0.03 mm yr−1 due to chemicalerosion), respectively. These findings were in agreement with thedenudation rate of the watersheds in the Central Himalayan region ofIndia (Rawat and Rawat, 1994).

l erosion rates at different watershed scales.

Table 7Average erosion rates in the Sainj and Tirthan watersheds

Station/river

Drainagearea (km2)

Physicalerosion

Chemicalerosion(t km−2 yr−1)

Totalerosion

Lowering rate of the basin

Physical Chemical Total

(mm yr−1)

Sainj atNiharni 410.65 521.20 86.81 608.00 0.20 0.03 0.23Suind 495.01 100.90 63.57 164.47 0.04 0.02 0.06Sainj 658.41 372.46 73.86 446.32 0.14 0.03 0.17Talara 727.26 654.72 79.08 733.80 0.25 0.03 0.28Larji 740.02 184.17 82.69 266.86 0.07 0.03 0.10Bhatkanda(Jiwa Nal)

163.40 154.61 103.65 258.26 0.06 0.04 0.10

Tirthan atKhrongcha 191.13 482.07 92.47 574.54 0.18 0.03 0.22Larji 686.36 116.46 80.09 196.56 0.04 0.03 0.07

34 O. Singh et al. / Catena 76 (2008) 27–35

3.5. Effect of watershed elevation and land use on erosion rate

The results of the present study did not show any clear relationshipbetween rates of erosion and watershed elevation. The Sainjwatershed having higher mean elevation of 3510 m amsl and basinarea of 741 km2 transported sediments at the rate of 267 t km−2 yr−1,whereas Tirthan discharged 197 t km−2 yr−1 with a mean elevation of2826 m amsl and an area of 687 km2. Further, Jiwa Nal sub-basin withmean elevation of 3684 m amsl and basin area of 164 km2 dischargeda total sediment load of 258 t km−2 yr−1, whereas Tirthan in its upper

Table 8Comparative erosion rates for the Sainj and Tirthan rivers and some major world rivers

River Country Erosion rate (t km−2 yr−1)

Physical Chemical To

Indian riversGanges (Calcutta) India 438 111 5Bhramputra India 865 88 9Yamuna (Allhabad) India 186 122 3Chambal India 129 154 2Betwa India 211 1817 20Ken India 219 230 4Gomti India 225 40 2Ghaghra India 978 130 11Son India 709 96 8Gandak India 383 140 5Krishna India 16 41Godavari India 555 55 6Cauvery India 0.5 40Mahanadi India 13.3 67.6India average 327 69 3

Himalayan riversGanga (Hardwar India 152 26 1Yamuna (Tajewala) India 1889 289 21Ramganga India 337 87 4Dhauliganga India 392 278 6Nana Kosi India 96 100 1Sainj India 184 83 2Jiwa Nal India 155 104 2Tirthan India 116 80 1

World riversHuang Ho China 1402 30 14Yangtze China 246 116 3Irrawady Burma 662 211 8Amazon Brazil 146 46 1Magdalena Colombia 916 117 10Mississippi U.S.A 64 40 1Orinoco Venezuela 212 52 2Mekong Vietnam 200 74 2Congo Zaire 13 10World average 150 35 1

course transported about 575 t km−2 yr−1 sediments with a meanelevation of 3585 m amsl and basin area of 191 km2.

Land use and agricultural practices in a watershed system havedrastic effects on erosion and sediment yield. In general, the landcovered with natural vegetation retards the rate of erosion (Cox et al.,2006). However, in this study the data of different locations in Sainjwatershed revealed that the erosion rates are not directly related withthe area under the forest cover. Moreover, there exists a direct rela-tionship between forest cover and erosion rates in Tirthan watershed(Table 1).

Nonetheless, it was observed during periodic field visits thatvarious anthropogenic factors were responsible for high level ofsediment transport besides the natural factors. The anthropogenicinterventions include large-scale encroachment in forest lands foragriculture, deforestation, cultivation of steep slopes and indiscrimi-nate grazing in the watershed by local and migratory animals. Also,road construction, mining, forest fires and other developmentalprojects have accelerated the process of soil erosion. Amongst allthese, the natural factors which could have accentuated the erosionprocess are the steep topographic gradient, weak geology, glacialerosion, erosive storms, erosion by streams, less infiltrability of soilsand multitude of mass wasting phenomena.

4. Limitations of study

Hydrological investigations in inaccessible mountainous terrainsand subsequent data analysis are based on some specific assumptions.The present study carried out in the Lesser Himalayan watersheds is

Physical/chemical ratio Source

tal

49 3.90 Abbas and Subramanian (1984)53 9.80 Subramanian 198308 1.52 Jha et al. (1988)83 0.83 Jha et al. (1988)88 0.12 Jha et al. (1988)49 0.95 Jha et al. (1988)65 5.62 Abbas and Subramanian (1984)08 7.52 Abbas and Subramanian (1984)05 7.38 Abbas and Subramanian (1984)23 2.73 Abbas and Subramanian (1984)57 0.39 Ramesh and Subramanian (1988)10 10.00 Bikshamaiah and Subramanian (1988)40.5 0.01 Subramanian et al. (1985)80.9 0.19 Chakrapani and Subramanian (1990)96 4.74 Jha et al. (1988)

78 5.84 Abbas and Subramanian (1984)78 6.54 Jha et al. (1988)24 3.87 Abbas and Subramanian (1984)70 1.41 Kumar (1987)96 0.96 Rawat and Rawat (1994)67 2.21 Present study58 1.49 Present study97 1.45 Present study

32 46.70 Milliman and Meade (1983)62 2.10 Milliman and Meade (1983)73 3.10 Milliman and Meade (1983)92 3.10 Milliman and Meade (1983)33 7.80 Milliman and Meade (1983)04 1.60 Milliman and Meade (1983)64 4.00 Milliman and Meade (1983)74 2.70 Milliman and Meade (1983)23 1.30 Milliman and Meade (1983)85 4.28 Jha et al. (1988)

35O. Singh et al. / Catena 76 (2008) 27–35

not an exception to that. Uncertainty in data collection and analysis ofpresent investigations are highlighted in this section. In this study, thesampling of water from channel was done using very simple method.Hence, uncertainties pertain to the measurement of width, depth,velocity of water in channel and the efficiency of the suspendedsediment sampler. However, the laboratory analysis of SSC, EC, pH etc.was carried out as per the international standards. Authors feel thatthe sample collection procedure did not represent the distribution ofSSC in total channel width, depth and velocity. Moreover, it is under-stood that a number of sections along the width and at differentdepths (near the bed, mid depth and at water surface) and an increasein the frequency of observations could have provided a better re-presentation of the sediment load. It is expected that because of errorsin sample collection and assumptionsmade in the analysis might haveresulted in an error of about 10% in the results.

5. Conclusions

Thewatersheds exhibited a dominance of total sediment load (TSL)over the total dissolved load (TDL) reflecting that the physicalweathering (mass wasting, slumping of heavily sediment laden portalice in the streams, occurrence of rainstorms etc.) were more pro-nounced in the watersheds. The bulk of sediment load transported byglaciers in the upper reaches gets transformed into temporary storagewithin the alpine basins and is removed during the periods of highdischarges. The uniform values of TDS during the lean runoff period(October–March) reflected the contributions of base flow in thewatersheds. Further, human interventions in the form of develop-mental activities played a significant role in the spatial and temporalvariability of sediment load. Estimated sediment yield in thesewatersheds ranged between 164 to 734 t km−2 yr−1 at differentgauging locations. These estimates were very high as compared to theestimates available for the drainage basins around the globe. Above all,the periodic monitoring of water quality in the streams within thewatershed systems would assist in estimating its spatial and temporalvariability. In addition, this would also help in understanding thehydrologic responses of the Lesser Himalayan alpine watershedsystems for taking up appropriate soil and water conservation mea-sures leading to integrated watershed management.

Acknowledgements

We are grateful to the BBMB authorities for providing dailydischarge rate values for respective gauging sites. Additionally, theauthors would also like to thank the editor of the Catena journal andthe two anonymous reviewers for their critical, but highly constructivecomments that have considerably improved the quality of the paper.

References

Abbas, N., Subramanian, V., 1984. Erosional and sediment transport in the Ganges RiverBasin, India. Journal of Hydrology 69, 173–182.

BBMB (Bhakra Beas Management Board), 1997. Sedimentation Survey Report. BBMB,Bhakra Dam Circle, Nangal, India.

Bhattacharya, P., Bhatt, V.K., Mandal, D., 2008. Soil loss tolerance limits for planning ofsoil conservation measures in Shivalik–Himalayan region of India. Catena 73,117–124.

Bhutiyani, M.R., 2000. Sediment load characteristics of a proglacial stream of SiachenGlacier and the erosion rate in Nubra valley in the Karakoram Himalayas, India.Journal of Hydrology 227, 84–92.

Bikshamaiah, G., Subramanian, V., 1988. Sediment transport of the Godavari river basinand its controlling factors. Journal of Hydrology 101, 275–290.

Chakrapani, G.J., 2005. Factors controlling variations in river sediment loads. CurrentScience 88, 569–575.

Chakrapani, G.J., Subramanian, V., 1990. Prelimnary studies on the geochemistry of theMahanadi River basin, India. Chemical Geology 81, 241–253.

Chakrapani, G.J., Subramanian, V., 1993. Rates of erosion and sedimentation in theMahanadi river basin, India. Journal of Hydrology 149, 39–48.

Colby, B.R., Hembree, C.H., 1955. Computations of total sediment discharge of NiobraraRiver near Cody, Nebraska. US Geological Survey Water Supply Paper No. 1357.USGS, Washington, DC.

Cox, C.A., Sarangi, A., Madramootoo, C.A., 2006. Effect of land management on runoffand soil losses from two small watersheds in St. Lucia. Land Degradation andDevelopment 17, 55–72.

Fergusson, R.I., 1984. Sediment load of Hunza river. In: Miller, K.J. (Ed.), InternationalKarakoram Project 2. Cambridge University Press, Cambridge, pp. 581–597.

Gregory, K.J., Walling, D.E., 1973. Drainage Basin Form and Processes — A Geomorpho-logical Approach. Edward Arnold, London, UK.

Haritashya, U.K., Singh, P., Kumar, N., Gupta, R.P., 2006. Suspended sediment from theGangotri Glacier: quantification, variability and associations with discharge and airtemperature. Journal of Hydrology 321, 116–130.

Holeman, J.N., 1968. Sediment yield of major rivers of the world. Water ResourcesResearch 4, 737–747.

Jackson, M.L., 1967. Soil chemical analysis. Prentice Hall of India Pvt. Ltd., New Delhi,India.

Jha, P.K., Subramanian, V., Sitasawad, R., 1988. Chemical and sediment mass transfer inthe Yamuna River — a tributary of the Ganges system. Journal of Hydrology 104,237–246.

Kumar, P., 1987. Geomorphological evaluation of environmental degradation andmanagement in Dhauliganga Basin (Central Himalayas). Ph.D. thesis, JawaharlalNehru University, New Delhi, India.

Lu, X.X., Siew, R.Y., 2006. Water discharge and sediment flux changes over the pastdecades in the lower Mekong River: possible impacts of the Chinese dams.Hydrology and Earth System Sciences 10, 181–195.

Milliman, J.D., Meade, R.H., 1983. Worldwide delivery of river sediments to the oceans.Journal of Geology 91, 1–21.

Ostrem, G.C., 1964. A method of measuring water discharge in turbulent streams.Geographical Bulletin 21, 21–43.

Ramesh, R., Subramanian, V., 1988. Temporal, spatial and size variation in the sedimenttransport in the Krishna river basin, India. Journal of Hydrology 98, 53–65.

Rawat, J.S., Rai, S.P., 1997. Pattern and intensity of erosion in the environmentallystressed Khulgad watershed, Kumaun Himalaya. Journal of Geological Society ofIndia 50, 331–338.

Rawat, J.S., Rawat, M.S., 1994. Accelerated erosion and denudation in the Nana Kosiwatershed, Central Himalayas, India, Part I: sediment load. Mountain Research andDevelopment 14, 25–38.

Richards, L.A., 1954. Diagnoses and Improvement of Saline-Alkali Soils. US Departmentof Agriculture, Agricultural Handbook No. 60. USDA, Washington, DC.

Sarangi, A., Madramootoo, C.A., Singh, D.K., 2004. Development of ArcGIS assisted userinterfaces for estimation of watershed morphologic parameters. Journal of Soil andWater Conservation 3, 139–149.

Sharma, P., Rai, S.C., 2004. Streamflow, sediment and carbon transport from aHimalayan watershed. Journal of Hydrology 289, 190–203.

Singh, D., Chhonkar, P.K., Pandey, R.N., 1999. Soil Plant Water Analysis — A MethodsManual. Indian Agricultural Research Institute, New Delhi, India.

Singh, O., 2007. Geomorphological Evaluation, Soil Erosion Assessment and Manage-ment in Himalayan Watersheds, Himachal Pradesh. Ph.D. thesis, Jawaharlal NehruUniversity, New Delhi, India.

Singh, P., Haritashya, K.S., Ramasastri, K.S., Kumar, N., 2005. Diurnal variations indischarge and suspended sediment concentration, including runoff-delayingcharacteristics, of the Gangotri glacier in the Garhwal Himalayas. HydrologicalProcesses 19, 1445–1457.

Singh, P., Ramasastri, K.S., Kumar, N., Bhatnagar, N.K., 2003. Suspended sedimenttransport from the Dokriani glacier in the Garhwal Himalayas. Nordic Hydrology 34,221–244.

Sitaula, B.P., Garde, M., Burbank, D.W., Oskin, M., Heimsath, A., Gabet, E., 2007. Bed loadto suspended load ratio and rapid bedrock incision from Himalayan landslide-damlake record. Quaternary Research 68, 111–120.

Soler, M., Regüés, D., Latron, J., Gallart, F., 2007. Frequency–magnitude relationships forprecipitation, stream flow and sediment load events in a small Mediterranean basin(Vallcebre basin, Eastern Pyrenees). Catena 71, 164–171.

Stoddart, D.R., 1969. World erosion and sedimentation in water. In: Chorley, R.J. (Ed.),Water, Earth and Man. Methuen, London, UK.

Subramanian, V., 1979. Chemical and suspended sediment transport characteristics ofrivers of India. Journal of Hydrology 44, 37–55.

Subramanian, V., 1983. Factors controlling the chemical composition of river waters ofIndia. Proceedings of the Hamburg Symposium, IAHS Publication No. 141, pp.145–151.

Subramanian, V., 1993. Sediment load of Indian rivers. Current Science 64, 928–930.Subramanian, V., Van't Dack, L., Van Grieken, R., 1985. Chemical composition of river

sediments from the Indian sub-continent. Chemical Geology 48, 271–279.Syvitski, J.P.M., 2003. Supply and flux of sediment along hydrological pathways:

research for the 21st century. Global and Planetary Change 39, 1–11.Syvitski, J.P.M., Peckham, S.D., Hilberman, R., Mulder, T., 2003. Predicting the terrestrial

flux of sediment to the global ocean: a planetary perspective. Sedimentary Geology162, 5–24.

Valdiya, K.S., 1987. Environmental Geology — Indian Context. Tata Mc-Graw Hill Ltd.,New Delhi.

Vanacker, V., Molina, A., Govers, G., Poesen, J., Deckers, J., 2007. Spatial variation ofsuspended sediment concentration in a tropical Andean river system: the PauteRiver, southern Ecuador. Geomorphology 87, 53–67.

Vanoni, V.A., 1975. Sedimentation Engineering. ASCE, New York.Xiaoping, Y., 2003. Manual of sediment management and measurement. World

Meteorological Organization Operational Hydrology Report No. 47. Secretariat ofWorld Meteorological Organization, Geneva, Switzerland.

Zhang, Y., Long, Y., 1998. Improvement of Einstein's bed load function. Journal ofSediment Research 4.